FATIGUE

Neural and Muscular Mechanisms

ADVANCES IN EXPERIMENTAL MEDICINE AND BIOLOGY
Editorial Board:
NATHAN BACK, State University of New York at Buffalo
IRUN R. COHEN, The Weizmann Institute of Science
DA VID KRITCHEVSKY, Wistar Institute
ABEL LAlTHA, N. S. Kline Institute for Psychiatric Research
RODOLFO PAOLETTI, University of Milan

Recent Volumes in this Series

Volume 377
TISSUE RENIN-ANGIOTENSIN SYSTEMS: Current Concepts of Local Regulators in
Reproductive and Endocrine Organs
Edited by Amal K. Mukhopadhyay and Mohan K. Raizada

Volume 378
DENDRITIC CELLS IN FUNDAMENTAL AND CLINICAL IMMUNOLOGY, Volume 2
Edited by Jacques Banchereau and Daniel Schmitt

Volume 379
SUBTILISIN ENZYMES: Practical Protein Engineering
Edited by Richard Bott and Christian Betzel

Volume 380
CORONA- AND RELATED VIRUSES: Current Concepts in Molecular Biology
and Pathogenesis
Edited by Pierre J. Talbot and Gary A. Levy

Volume 381
CONTROL OF THE CARDIOVASCULAR AND RESPIRATORY SYSTEMS IN
HEALTH AND DISEASE
Edited by C. Tissa Kappagoda and Marc P. Kaufman

Volume 382
MOLECULAR AND SUBCELLULAR CARDIOLOGY: Effects of Structure and Function
Edited by Samuel Sideman and Rafael Beyar

Volume 383
IMMUNOBIOLOGY OF PROTEINS AND PEPTIDES VIII: Manipulation or Modulation
of the Immune Response
Edited by M. Zouhair Atassi and Garvin S. Bixler, Jr.

Volume 384
FATIGUE: Neural and Muscular Mechanisms
Edited by Simon C. Gandevia, Roger M. Enoka, Alan J. McComas, Douglas G. Stuart,
and Christine K. Thomas

Volume 385
MUSCLE, MATRIX, AND BLADDER FUNCTION
Edited by Stephen A. Zderic

A Continuation Order Plan is available for this series. A continuation order will bring delivery of each new volume
immediately upon publication. Volumes are billed only upon actual shipment. For further information please contact
the publisher.
FATIGUE
Neural and Muscular Mechanisms

Edited by
Simon C. Gandevia
Prince of Wales Medical Research Institute
Sydney, New South Wales, Australia

Roger M. Enoka
The Cleveland Clinic Foundation
Cleveland, Ohio

Alan J. McComas
McMaster University
Hamilton, Ontario, Canada

Douglas G. Stuart
The University of Arizona
Tucson, Arizona

and
Christine K. Thomas
University of Miami
Miami, Florida

Technical Editor
Patricia A. Pierce
The University of Arizona
Tucson, Arizona

SPRINGER SCIENCE+BUSINESS MEDIA, LLC

 Library of C o n g r e s s C a t a l o g i n g - i n - P u b l i c a t i o n Data

On file

Proceedings based in part on the symposium on Neural and Neuromuscular Aspects of Muscular Fatigue,
held November 10-13, 1994, in Miami, Florida

ISBN 978-1-4899-1018-9 ISBN 978-1-4899-1016-5 (eBook)

DOI 10.1007/978-1-4899-1016-5

© Springer Science+Business Media N e w York 1995

Originally published by Plenum Press, N e w York in 1995
Softcover reprint of the hardcover 1st edition 1995

1098 76 54 3 2 1

All rights reserved

No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any
means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written
permission from the Publisher
 The challenge is to understand fatigue from the level of the single muscle fiber to the
whole organism. [Photograph from Abbott, Bigland, and Ritchie, Journal of Physiology
(London) 117:382,1952 . Reprinted with permission.]
PREFACE

This volume describes the current state of our knowledge on the neurobiology of
muscle fatigue, with consideration also given to selected integrative cardiorespiratory
mechanisms. Our charge to the authors of the various chapters was twofold: to provide a
systematic review of the topic that could serve as a balanced reference text for practicing
health-care professionals, teaching faculty, and pre- and postdoctoral trainees in the biomedi-
cal sciences; and to stimulate further experimental and theoretical work on neurobiology.
Key issues are addressed in nine interrelated areas: fatigue of single muscle fibers,
fatigue at the neuromuscular junction, fatigue of single motor units, metabolic fatigue studied
with nuclear magnetic resonance, fatigue of the segmental motor system, fatigue involving
suprasegmental mechanisms, the task dependency of fatigue mechanisms, integrative
(largely cardiorespiratory) systems issues, and fatigue of adapted systems (due to aging,
under- and overuse, and pathophysiology). The product is a volume that provides compre-
hensive coverage of processes that operate from the forebrain to the contractile proteins.
The neurobiology of muscle fatigue received minimal attention throughout the first
six decades of the 20th century. Then, in the mid- to late 1970s, a series of reports appeared
on neuromuscular aspects of fatigue that were featured in influential international symposia
held in 1980 (London, UK), 1990 (Paris, France), 1992 (Amsterdam, The Netherlands), and
1994 (Miami, Florida). This sudden acceleration of interest in the study of neural and
muscular aspects of fatigue is attributable to at least three factors: the practical importance
and relevance of the topic; its long-neglected fundamental and clinical significance; and the
type of integrated biological thought and experimentation needed to advance the field,
extending from the cellular/molecular level to the behaving organism. This latter integrative
emphasis is a feature of the present volume in all nine subsections. The volume also stresses
the need for neuroscientists to expand our understanding of the role of the eNS in fatigue
processes to the same extent as our current understanding of peripheral neuromuscular
mechanisms. To this end, several chapters in the present volume emphasize how the full
experimental armamentarium of the field of neuroscience can be applied to the neural aspects
of fatigue.
While the volume forms a cohesive whole, we have respected the individuality of
our authors, including innumerable nuances in their definitions of muscle fatigue. Similarly,
in selected areas, there is a degree of overlap between chapters such that the reader can
contemplate how the same issue is approached by different laboratories. These features of
the volume should prove helpful when reviewed in pre- and postdoctoral training programs.
To facilitate this possibility, the nine subsections of the volume are provided with a brief
introduction prepared by the editors.
Since the mid-1970s, Brenda Bigland-Ritchie has been in the forefront of the study
of the neuromuscular aspects of muscle fatigue. Using human subjects, she has shown to

vii
viii Preface

what extent the fatigue findings from animals or isolated tissues apply to real-life situations.
More than this, she has stimulated inquiry into the search for reflex mechanisms which may
serve to match the activity of the spinal cord with that of the fatiguing muscles. The present
generation of scientists who study muscle fatigue owe much to her contributions and
enthusiasm for the field. The chapters that follow document her impact. They also provide
the rationale for why the editors and contributing authors have enthusiastically dedicated
this volume to Dr. Bigland-Ritchie. Her efforts have been exemplary.

Simon C. Gandevia
Roger M. Enoka
Alan J. McComas
Douglas G. Stuart
Christine K. Thomas
ACKNOWLEDGMENTS

We greatly appreciate the dedication, effort, and talent that our technical editor,
Patricia Pierce, has displayed in bringing this research monograph to successful fruition. We
would also like to thank several University of Arizona colleagues: Dr. Edwin Gilliam, Dr.
Jennifer McDonagh, Robert Gorman, and George Hornby for their help with the review and
executive processing of the chapters in this volume.
Much of the work in this volume was presented by the authors of each chapter at an
International Symposium, Neural and Neuromuscular Aspects of Muscle Fatigue, held in
Miami, FL, November 10-13, 1994. Designated as a Satellite Symposium of the Annual
Meeting of the Society for Neuroscience, the symposium honored Professor Brenda Bigland-
Ritchie and was organized by Drs. Roger Enoka, Simon Gandevia, Alan McComas, Douglas
Stuart, and Christine Thomas. The presentations in Miami by 33 invited scientists from
Canada and the United States were supported largely by their own research funds, with some
assistance from the University of Miami (Department of Neurological Surgery, Miami
Project to Cure Paralysis) and the University of Arizona (Research Office of the College of
Medicine, and Regents' Professor funds of Douglas Stuart). Travel awards were provided to
37 scientists from other countries (Australia, Belgium, Denmark, Finland, France, Germany,
Israel, Norway, Slovenia, Switzerland, The Netherlands, UK, Yugoslavia) by the American
Physical Therapy Association (Research and Analysis Division, and Section on Research),
the Muscular Dystrophy Association (Conference grant no. 46168), the National Institutes
of Health (National Center for Medical Rehabilitation Research, NICHD) and the Universi-
ties of Miami and Arizona. Ten trainee awards were provided by The Miami Project to Cure
Paralysis. We would also like to thank Bette Mas (Miami Project to Cure Paralysis,
University <;>f Miami) and Lela Aldrich, Lura Hanekamp and Joan Lavoie (Department of
Physiology, University of Arizona) for their considerable efforts to ensure the success of the
symposium. The abstracts and discussion summaries from the symposium will be published
by Muscle & Nerve in the Fall of 1995, as sponsored by the Henry L. and Kathryn Mills
Charity Foundation and the Cleveland Clinic Foundation.

The Editors

ix
CONTENTS

Looking Back ......................................................... .

Brenda R. Bigland-Ritchie

The Scientific Contributions of Brenda Bigland-Ritchie ........................ 11

C. K. Thomas, R. M. Enoka, S. C. Gandevia, A. J. McComas, and
D. G. Stuart

Section I: Fatigue of Single Muscle Fibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 27

1. Myofibrillar Fatigue versus Failure of Activation. . . . . . . . . . . . . . . . . . . . . . . . .. 29

K. A. P. Edman

2. Mechanisms of Excitation-Contraction Coupling Relevant to Skeletal Muscle

Fatigue ....................................................... 45
D. G. Stephenson, G. D. Lamb, G. M. M. Stephenson, and M. W. Fryer

3. The Role ofIntracellular Acidosis in Muscle Fatigue. . . . . . . . . . . . . . . . . . . . . .. 57

D. G. Allen, H. Westerblad, and J. Liinnergren

4. Role ofInterstitial Potassium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 69

G. Sj0gaard and A. J. McComas

Section II: Fatigue at the Neuromuscular Junction .......................... 81

5. Fatigue at the Neuromuscular Junction: Branch Point vs. Presynaptic vs.

Postsynaptic Mechanisms .... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 83
G. C. Sieck and Y. S. Prakash

6. The Role of the Sarcolemma Action Potential in Fatigue .................... 101

A. J. Fuglevand

7. Single Fiber Electromyography in Studies of Neuromuscular Function ........ 109

J. V. Trontelj and E. StAlberg

xi
xii Contents

Section III: Fatigue of Single Motor Units .................................. 121

8. Intrinsic Properties of Motoneurons: Implications for Muscle Fatigue ......... 123

A. Sawczuk, R. K. Powers, and M. D. Binder

9. Neuromuscular Frequency Coding and Fatigue ........................... 135

D. Kernell

10. Human Motor Units Studied by Spike-Triggered Averaging and Intraneural

Motor Axon Stimulation ......................................... 147
C. K. Thomas

11. Human Motor Units Studied by Intramuscular Microstimulation ............. 161

J. M. Elek and R. Dengler

Section IV: Fatigue Studied with NMR Techniques .......................... 173

12. Bioenergetics and Muscle Cell Types ................................... 175

M. 1. Kushmerick

13. Metabolic Correlates of Fatigue from Different Types of Exercise in Man ...... 185
N. K. V011estad

14. Mechanisms of Human Muscle Fatigue: Quantitating the Contribution of

Metabolic Factors and Activation Impairment ........................ 195
R. G. Miller, 1. A. Kent-Braun, K. R. Sharma, and M. W. Weiner

15. Emerging Opportunities with NMR ..................................... 211

L. A. Bertocci

Section V: The Case for Segmental Motor Mechanisms ....................... 225

16. Variable-Frequency Stimulation Patterns for the Optimization of Force during

Muscle Fatigue: Muscle Wisdom and the Catch-like Property ........... 227
S. A. Binder-Macleod

17. Overview: Potential Role of Segmental Motor Circuitry in Muscle Fatigue ..... 241
U. Windhorst and G. Boorman

18. The Fusimotor System: Its Role in Fatigue ............................... 259

K.-E. Hagbarth and V. G. Macefield

19. Role of Muscle Afferents in the Inhibition of Motoneurons during Fatigue ..... 271
S. J. Garland and M. P. Kaufman

Section VI: The Case for Central Fatigue .................................. 279

20. Central Fatigue: Critical Issues, Quantification and Practical Implications ...... 281
S. C. Gandevia, G. M. Allen, and D. K. McKenzie
Contents xiH

21. Single-Trial Readiness Potentials and Fatigue ............................ 295

D. Popivanov, A. Mineva, and J. Dushanova

22. The Senses of Effort and Force during Fatiguing Contractions ............... 305
L. A. Jones

23. Tryptophan, 5-Hydroxytryptamine and a Possible Explanation for Central

Fatigue ....................................................... 315
E. A. Newsholme and E. Blomstrand

Section VII: Task Dependency of Fatigue Mechanisms ....................... 321

24. The Significance of Motor Unit Variability in Sustaining Mechanical Output of

Muscle ....................................................... 323
A. J. Sargeant and D. A. Jones

25. Intramuscular Pressures for Monitoring Different Tasks and Muscle Conditions 339
O. M. Sejersted and A. R. Hargens

26. Task-Dependent Nature of Fatigue in Single Motor Units ................... 351

B. R. Botterman

27. Task-Dependent Factors in Fatigue of Human Voluntary Contractions ......... 361

B. Bigland-Ritchie, C. L. Rice, S. J. Garland, and M. L. Walsh

Section VIII: Integrative Systems Issues ................................... 381

28. Fatigue and the Design of the Respiratory System ......................... 383
S. L. Lindstedt and H. Hoppeler

29. An Integrative View of Limitations to Muscular Performance ................ 393

J. A. Dempsey and M. A. Babcock

30. Respiratory Muscle Fatigue ........................................... 401

D. K. McKenzie and F. Bellemare

31. Fatigue of Jaw Muscles and Speech Mechanisms .......................... 415

T. S. Miles and M. A. Nordstrom

Section IX: Fatigue of Adapted Systems: Overuse, Underuse, and

Pathophysiology ........................................................ 427

32. Fatigue in Adapted Systems: Overuse and Underuse Paradigms .............. 429
T. Gordon

33. Associations between Muscle Soreness, Damage, and Fatigue ............... 457
P. M. Clarkson and D. J. Newham

34. Muscle Fatigue in Old Animals: Unique Aspects of Fatigue in Elderly Humans 471
J. A. Faulkner and S. V. Brooks
xiv Contents

35. Historical Perspective: A Framework for Interpreting Pathobiological Ideas on

Human Muscle Fatigue .......................................... 481
R. H. T. Edwards, V. Toescu, and H. Gibson

36. Fatigue Brought on by Malfunction of the Central and Peripheral Nervous

Systems ...................................................... 495
A. J. McComas, R. G. Miller, and S. C. Gandevia

Epilogue

37. Neurobiology of Muscle Fatigue: Advances and Issues ..................... 515

S. C. Gandevia, R. M. Enoka, A. J. McComas, D. G. Stuart, and
C. K. Thomas

Contributors ............................................................ 527

Index ................................................................. 533

LOOKING BACK

Brenda R. Bigland-Ritchie

Department of Pediatrics
Yale University School of Medicine
New Haven, Connecticut 06510

I do not know why I was selected for this honor; a book and a symposium both
dedicated to me. Why me when so many others seem more worthy? But having been selected,
I want first to express my deep gratitude and appreciation to those who decided to pay this
quite extraordinary tribute to my work - the organizers, Roger Enoka, Simon Gandevia, Alan
McComas, Douglas Stuart, and Christine Thomas. I do not know how they first got this idea,
but I thank them for it more than I can say.
When asked to write this chapter, I was encouraged to look back over my career, and
to describe the early events that influenced my path. It was all an adventure; mainly, of
course, the adventure we all share when we keep nibbling away at some scientific problem
until one of nature's well-kept secrets is at least partially revealed. I also found great
satisfaction in exploring the limits of my own abilities at each stage in my career. In looking
back, I realize now just how greatly I was influenced by the standards set by my early
teachers, and also by chance encounters with people who just happened to be in the right
place at the right time. We seldom know where fate may lead us. However, many incidents
are recalled simply because they were amusing and came from an era now long gone.
From time to time, over the years, I have been told that I have played some kind of
role model for other, younger women who faced problems similar to mine. If that is so, then
it may also be of interest to hear how it felt to be a woman during a period in which public
attitudes about the role of women in the workplace have changed so much, and how this has
affected me. And what was the impact on my research of having chosen to interrupt my
career for many years to raise my children, of changing countries; and what it took to get up
and running again in physiology, a task I could never have achieved without the constant
support, generosity and insistence of others. And finally, why was the effort all so very well
worth while?

EARLY YEARS

I was an only child who spent her most formative adolescent years in British war-time
(1939-45), mostly secluded in a country boarding school safely away from London bombing.
The experience of those hard and anxious times was probably important for all of us who

1
2 B. R. Bigland-Ritchie

later engaged in research; for the shortages of almost everything fostered a certain ingenuity
in improvising. By luck, I was ready to start my undergraduate years at University College
London in 1945, when the war in Europe was just over. What could be more exciting for a
raw, overly protected, and inexperienced school girl than to find that the majority of her
classmates were older ex-service men and women returning from a wide range of exciting
and highly responsible war-time experiences, ranging from sea captains to prisoners of war
assigned to such jobs as building the Burma Road. They opened our eyes as no ordinary
freshman class could do to an awareness of what real life, responsibility and suffering could
be. In those early years, I learned far more from them than from any of my formal classes
(which I took fairly lightly at that time).
In 1945, the College had just returned from 4 years evacuation to Bangor, Wales,
where large numbers of Welshman had enrolled. One of my fondest memories is of glorious
male choruses which frequently erupted to fill the cloisters with harmony. This could happen
at any time, as when sitting in the cafeteria someone would start humming, then the refrain
was quickly taken up from different sides until it engulfed everyone.
I entered college not knowing which branch of science most attracted me. At one
stage, I applied to take a degree in chemistry. Thank goodness they turned me down! Then
someone said that a new degree course called Physiology was to be started. None of us had
heard of that subject or what it would include; but we all rushed off to register because it
meant taking courses with medical students whom every one knew were the most fun! So
started a career which has brought such reward throughout my life, but which I might so
easily have missed. I cannot now imagine being equally intrigued by any other topic. Most
of the directions my life has taken since then have also been shaped by just such unpredictable
chance happenings.
By going to University when I did, I was in time to catch the coat-tails of many of
the outstanding scientists of those times. Sir Henry Dale, Gaddum, and other equally famous
figures were all invited frequently to speak at our Medical Student Society evening meetings.
Neither they nor we ever considered that this should entail an honorarium. How times have
changed! I was taught by such people as Lovatt-Evans (colleague of Starling), Leonard
Bayliss, Bernard Katz, to name a few. But the most profound early influence, then and later,
came from A.V. Hill. My introduction to him was quite amusing. In those days, an important
feature of science education, unfortunately lost today, was to emphasize the concept that all
scientists stand on the shoulders of their predecessors, and that one can learn a lot from both
the achievements and mistakes of those who went before. Thus, each lecture started with a
brief history of how it had developed and by whom. They might start with such ancients as
Socrates, or Aristotle, or more recent figures including Galvani, Cajal, Claude Bernard,
Helmholz, Sherrington, etc. - all revered figures, but usually long dead. When the lectures
on muscle began, reference was made at once to A.V. Hill, with the lecturer usually glancing
upwards. I thus got the image of the great man playing a harp on some cloud in heaven. It
was only much later I discovered that he was quite alive and well, but just happened to work
on the floor above.

[Link]
My first actual meeting with A. v., as everyone called him, came through Murdoch
Ritchie (later my husband). Murdoch was one of three young scientists (Bernard Abbott and
Eric Denton were the others) whom A. V. recruited when he returned to science after spending
the war years as a leading science advisor to the British and American governments. A.V.
decided then that for modem biology to progress satisfactorily, it must, henceforth, engage
the skills of engineers and physicists. So he proceeded to found what became the first
Looking Back 3

officially named department of Biophysics. Murdoch, like the others, already had a degree
in physics and mathematics, and had spent the final year of the war assigned to work on
radar research as a boffin boy. They joined our classes when A.V. told them to go learn some
biology.
Eventually I graduated and started thinking about a job. Things were so different in
those days, and so much better than for young people today. I never worried as to whether
there would be a job for me, or what that job would be, or whether there would be funds to
let me do it. Something would tum up, and it did. I was lucky enough to be given a fellowship
to stay on working in the same Physiology Department. I was assigned to work with another
new graduate, Barbara Jehring, and we were given a large empty room to work in. After what
seemed like many weeks, the professor (G. L. Brown) walked in one day and asked us
whether we had yet decided what we were going to work on, how we would do it, and what
equipment we would need. In those days, we were all expected to make our equipment,
which we constructed in the excellent workshops provided by each department. Our first
simple experiments were designed to answering a question posed by the biochemists as to
whether hypertrophy from administration of growth hormone did make muscles stronger-
the first time, as far as I know, that this question had been addressed. What frustrating
experiments! We were lent a string galvanometer from which to measure force from the
deflection of a beam oflight focused onto a small mirror, with the magnitude ofthe deflection
registered at the other side of the room. Hence the room had to be totally dark. Our final
conclusion was that growth hormone did not affect muscle strength, but at that time I had
little faith in our data. Curiously, this conclusion seems to have survived (although I have
seldom seen our paper referred to). By that time, Barbara and I were both about to marry,
and we consoled ourselves with the thought that, if necessary, we could use our new names
in future to dissociate ourselves from those first bumbling efforts.
My next adventure, suggested directly by A. V. Hill, was clearly planned and executed
much more seriously, and also much more fun. At that time, AV. was measuring the heat
produced in isolated frog muscles when they were allowed to shorten or were stretched. He
suggested that Bud Abbott, Murdoch and I make similar comparisons of the relative
metabolic costs in man by measuring oxygen consumption in each case. Despite his
reputation for mathematical exactitude, and the esoteric and highly theoretical appearance
of his work, he maintained a strong interest throughout his life in applying his conclusions
to explain whole animal or human performance and athletic achievements measured in the
field. One essential requirement for our proposed experiments was to be sure that the
shortening and lengthening contractions were both made by identical muscle groups, moving
at the same velocity, while exerting equal tensions. Only the direction of movement should
be reversed. To do this, we borrowed a primitive bicycle adapted as an ergometer and hitched
it back-to-back with the one the department already had (Fig I). One cyclist pedaled forward
(his quadriceps muscles shortening) while the muscles of the other cyclist were stretched by
resisting the motion with enough force to prevent any change in speed. This assured that the
forces in both directions were identical. Expired air was collected from each subject and its
oxygen content measured. Enormous differences were revealed between the energy require-
ments in each case. Despite his obvious awareness as a physicist, AV. insisted that the two
movements be called positive and negative work. He had no problem with these terms since,
when a muscles is stretched, work is done on it.
At AY. 's urging, we twice demonstrated these experiments at the Royal Society
where they aroused enormous interest and amusement. I have fond memories of the Lord
Mayor of London, complete with all his ceremonial regalia, asking to have a go. Our fame
quickly spread through London where the equipment came to be named the push-me,
pull-you, after Dr. Doolittle's two-headed animal that never knew which way it was going.
Many years later, a well-known scientist wrote about how a svelte young woman could resist
4 B. R. Bigland-Ritchie

Figure 1. Apparatus used in 1952 to compare the metabolic costs of positive and negative work (from Abbott
Be, Bigland B & Ritchie JM (1952). The Physiological cost of negative work. Journal ofPhYSiology (London)
117,380-390; their Fig. I).

and quickly exhaust the efforts of a husky young man, but had forgotten that these two were
Murdoch and I! In later studies, we replaced the forward positive cyclist by a motor!
A.Y. was a painfully shy man with no small talk. One of the things we all dreaded
was to find ourselves with him at some departmental social function. But all was well if one
could get him on to one of his favorite scientific theories, such as why dimensionally similar
animals - whales and dolphins, or race-horses and greyhounds - can run or swim at the same
top speed; or why the heart rates of humming birds and elephants must be so different. He
would work it all out mathematically from their corresponding masses, the relative stresses
on their bones and tendons, and their metabolism. In 1927, at a muscle meeting at Rockefeller
University, someone asked how long it would take to run up the Empire State building. The
next day the answers were to be compared with the best performance of two top athletes.
Others made wild guesses which all turned out to be far too long. A.v., however, made a
quick calculation, based on force-velocity, length-tension, and power-velocity relations to
come up with an answer that was correct to within a fraction of a minute. For him, the role
played by art, music or other forms of culture was best satisfied by the beauty he saw in
mathematics applied to every aspect of life. His detachment from popular culture was well
illustrated when I jokingly suggested to him one day that, with all the amusement we had
caused, I should resign from our bicycle experiments so that they could be authored by Abbott
& Costello. (Jose del Castillo occupied the next lab.) He just said "Why?". Clearly he had
never heard of these comedians!
However, these experiments and their notoriety were important at a more serious
level, in that they spawned a series of future studies by Asmussen, ourselves and many others
who came to recognize that muscles resist stretch as often as they shorten, and that to
coordinate their actions accurately in each mode requires quite different control strategies
by the CNS. Our experiments also helped to promote A. Y. 's famous challenge to biochemists
in which he asked them, among other things, to show whether the reactions of the Krebs
cycle may be reversible and if work done on a muscle (negative work) might result in the
re-synthesis of ATP. I remember many years later being told by a now famous scientist that
Looking Back 5

when he first read our paper as a graduate student, he went to the gym to see if he could
exhale oxygen while climbing down a rope!

OLOF LIPPOLD
The next person to have a major impact on formulating my research interests was
Olof Lippold. What a creative mind and imaginative teacher! It was during the next two
years which I spent in his laboratory that I became convinced that the questions I found most
fascinating concern how the CNS controls muscular performance, rather than about the
properties of the muscles themselves. He has always been a strong advocate of human
physiology, and working with him made me realize that many questions I wanted to address
could only be answered by work on human volunteers; a view I have maintained ever since.
When I joined him, Olof had demonstrated a linear relation between integrated EMG and
force during isometric contractions of human gastrocnemious. Together we showed a similar
linear relation during movement, with differences in slope between those for isotonic
shortening and lengthening similar to those predicted from the oxygen measurements during
cycling. We also obtained quite respectable force-velocity curves based on EMG.
Obviously these experiments raised questions as to what integrated EMG measures
mean in terms of motor drive. Did different levels reflect changes in motor unit recruitment,
or was rate coding the dominant feature - questions that continue to engage many of us today.
While there were already a few reports on this topic, Olof and I were probably the first to
make a systematic study relating the frequency of action potentials recorded from single
motor units to the force exerted during voluntary contractions of human muscles. We
recorded spike-trains on miles of bromide paper which we subsequently unrolled down the
length of the main corridor so that we could measure inter-spike intervals using a meter ruler
on our hands and knees.

FIRST RETIREMENT

I gave up my job in 1953 when my first child was born, as most women did then. I
could have continued to work but I decided, somewhat piously, that having children was my
choice and that they, therefore, deserved my full attention. I was so sure, but now I believe
this decision may have been a mistake. I might have been a better mother if I had retained
some intellectual interests with which to balance the more tedious aspects of domestic life
and the inevitable sense of isolation. I certainly wanted to spend a lot of time with the children
in the first few years; but had it not been for other unanticipated events, I believe I would
have worked again much sooner than I did. At that stage in our lives we were relatively poor,
always wondering whether we could make it to the end of each month.

MOVE TO AMERICA

In 1956 we were invited to spend a year in the United States, just as visitors. That
was one of the best years ever. We traveled madly every weekend, thinking we might never
have another chance again. But again fate stepped in. My husband was offered a permanent
position at The Albert Einstein College of Medicine, then brand new, and so we were faced
with the hardest decision of our life. It is not easy to abandon a land that has been good to
you, or to abandon family and friends; but the job and prospects in the U. S. were so much
better than we were likely to find in Britain then. Besides, we liked it in the U.S. So we
stayed.
6 B. R. Bigland-Ritchie

RETURN TO PHYSIOLOGY

For several years I tried to fill my intellectual void by volunteer work (the League of
Women Voters, attempts to desegregate the local schools, low-income housing, etc.). But I
always became disenchanted when the people who appeared to be most dedicated failed to
act on their convictions if, in practice, this might threaten the quality of their own lives. By
this time, my children were in high school, and my husband was so deeply engrossed in his
own science that he had little time for me. At aged 40, with perhaps another 40 years to go,
I felt useless and lonely. I had to find an absorbing occupation of my own. But I was also
convinced I could never catch up with the science I had missed. How could I compete after
all these years?
At this stage my husband, who always had more faith in my abilities than I did,
persuaded me to work part-time in his laboratory, just to see how it felt. I remember
reluctantly agreeing to try this on the conditions that I was regarded as only a useful pair of
hands, not required to think! Of course this venture was only partially successful. He was
always so much smarter and quicker at each job than 1. Nevertheless, I am proud of the three
papers we published on the mode of action of local anesthetics. Part of our problem was that
we were unable to leave the lab behind when we went home. I remember several times one
of us waking in the night to say "perhaps if we tried it this way it would work." Indeed, I
am filled with admiration for any couple who work well professionally together.
I still did not know what I wanted to do more permanently. I kept telling myself that
it would not be possible for someone of my age to compete with today's students. Then fate
played its hand again. A friend told me that Hunter College was seeking someone to teach
just one section of a physiology course. Although I had always been quite certain that I did
not want to teach, I realized that this would be one way to find out what I was capable of
doing. I would no longer have excuses to put that off. If one had a class to teach at 8 a.m.,
one better have something relevant to say. To my amazement, I discovered several striking
things about myself. First, I could indeed learn much faster and more comfortably than the
students. Certainly I could no longer memorize as they could - but who needs to memorize?
Maturity had taught me to discriminate between what was important and what was not.
Without my noticing it, my powers to reason had become so much better. Second, I found
that teaching at that level was enormously rewarding. Few things gave me a greater thrill
than to ignite a gleam of excitement and a flash of understanding in the dull eye of a student
who has previously asked only What do we need to know for the exam? I also found that
learning the techniques of effective teaching, particularly to less able students, could be as
challenging as designing the right protocol for a good experiment. Most difficult of all was
how to keep challenging the really able students, while not swamping those with more limited
abilities. Finally, it taught me once again how wrong I can be about those things that I am
most certain I know about already. I had found a most satisfying role for myself where I least
expected.

COMPLETING MY Ph.D

However, I soon realized that I must face a major hurdle if I hoped that university
level teaching might provide a reasonable basis for a late-breaking, alternative career. At that
time, I did not yet have a Ph.D.! I had been registered to take one, and could easily have
done so during the four years I worked at University College London, where I had published
more than enough papers to meet all Ph.D. requirements. Indeed, my supervisors strongly
encouraged me to complete one when they heard I would be leaving to care for children.
Looking Back 7

They even offered to give me the oral exam in the maternity hospital! But I refused. In those
days a Ph.D. was not considered a necessary requirement for a successful academic career
in Britain; and I was so sure that I would be much too old to need one later.
At that point, I was all for giving up, but I had to live with a husband who kept nagging
me to write to London University to see whether they would, in fact, still allow me to
complete a thesis, even fifteen years after the original deadline had expired! Eventually I
agreed to write just to shut him up, although it seemed utterly impossible that permission
could be granted. Imagine my chagrin when they finally replied some three months later that
my application to complete my Ph.D. had been accepted. By that time, my supervisor had
left the University and the external examiner they had originally appointed had died, so I
had no one to advise me. But eventually these details were settled. Now I was left with no
alternative but to prove myself by writing a real thesis, in contrast to just mugging up facts
from some textbook before a class. The next summer I sent one child away to camp and the
other home to grandparents in England and got to work.
My job was to re-write my original experiments as though they had been done only
about six months ago instead of sixteen years; a daunting task when the field had changed
so much. At the time when I was a student, ideas about the nature of muscular contraction
were so different. Then we talked about dash-pots and visco-elastic elements, heat measure-
ments, and mathematical equations to describe different types of muscle behavior, but we
had absolutely no knowledge of sliding filaments or cross-bridge structures. These new
concepts had all burst upon us during the years when I had been scientifically dormant, so
I was faced with lots of reading. Once again, I was lucky in that, during that time, rather
little detailed attention had been given to my particular comers of interest, a comparison of
the cost of shortening and lengthening contractions executed at different velocities, inte-
grated EMG as a quantitative measure of CNS muscle activation, and the motor unit firing
patterns that underlie them.
Initially, my aim in writing a thesis was simply to acquire the paper qualification to
apply for a faculty position - a piece of paper with three letters on it which I viewed as a
bureaucratic chore. I did not think it would make me one iota better at my job. But, in fact,
it turned out to be a task I enjoyed and one which largely restored my confidence in my
abilities. If! could master one small section of physiology, I could surely do it in other areas
should the need arise. More importantly, I found the niche for myself in science which I had
been seeking. I saw myoid experiments with new eyes, finding in them new concepts that
had not occurred to us initially. I became excited once again about how these ideas might be
developed. I was eager to try again to add some new pieces to the puzzle; and I came to
believe that I had found a different way to look at things. And these ideas were my own; no
one else had put them in my head. Clearly the most difficult step for mature women who
wish to re-enter the work place after raising children is to overcome their lack of confidence
in their ability to compete. Thus, in my case, not writing my thesis at the proper time turned
out to be blessing in disguise.
Although I got my Ph.D. in 1969, it was still some years before I did any more
experiments to explore those new ideas. First, I had to choose once more between family
needs and my own. By that time, I had spent three years teaching happily at Marymount
College in Tarrytown, N.Y. Then my husband was offered the chairmanship of Pharmacology
at Yale. Of course we discussed the impact of this move on my fledgling new career; but in
reality, I knew that my female pre-conditioning gave me no choice. Whatever he said, I could
never be happy unless I put his needs first. Nowadays, major academic institutions often
overcome this problem for professional couples by offeringjobs to both partners if they want
someone badly enough. But this was unheard of then. Nor would I have considered my own
status at that time worthy of making that demand. However, after some searching, I was
lucky enough to find a new home at Quinnipiac College and eventually I got tenure there.
8 B. R. Bigland-Ritchie

QUINNIPIAC COLLEGE

Quinnipiac College is a liberal arts college with teaching as its primary function.
Nevertheless, in order to advance in rank, we had to prove we had contributed successfully
to basic research. Of course, at that time, such colleges provided neither space nor funds for
these activities. Clearly, research in that context meant something one could keep in a drawer
and pull out between classes. Furthermore, the teaching load was such that anything more
than token research was quite impossible. Nevertheless, with the help of another new faculty
member (Joe Woods), I decided to try my hand at writing a full NIH research grant
application. I was not sure that we could put something together that would justify submis-
sion, but I certainly enjoyed trying. As I remember, my main aim was to get some impartial
assessment of the ideas I had expressed in my thesis. I feared that the encouragement I had
received previously might be mainly out of kindness. I certainly did not expect our
application to be funded. Nor do I believe that it would have been, except for another piece
of extraordinary good luck, the presence on the study section of Doug Stuart, Milic-Emili
and others who happened to remember our early papers and decided to give us another
chance.

GETTING STARTED AGAIN

What could we do! We had no place to work. The only space the college could find
for us to work was a large storage closet in the gym, next to the men's weight room and right
over the commercial dryers. It was not surprising that we spent most of our time battling
mechanical and electrical interference problems. Despite this, however, Joe Woods and I
eventually produced three papers designed to validate some of the ideas I had expressed in
my thesis. In doing this, we were helped enormously by friends at the John B. Pierce
Laboratory. They built our motorized bicycle ergometer in their workshop and Hans
Graichen often rescued us from electronic problems. (Quinnipiac had no technical backup
facilities.) Thus, when our grant came up for renewal in 1976, we asked the site visitors to
arrange for us to rent permanent lab space at Pierce, which they did. So started a long and
fruitful time. From that point, my subsequent adventures in human physiology is largely on
the published record. However, there is one more tale I have to tell. That is how and why I
first became interested in fatigue.

HOW FATIGUE SET IN

Until 1974 my research, like that of most others at the time, was directed towards
neuromuscular responses evident in short-term exercise. I had not given much thought to
how these responses might change as exercise continued. Then chance stepped in again. My
husband and I went to Cambridge to attend a meeting of the British Physiological Society,
and then have lunch with A.y. Hill. By that time, he was living there in retirement. His son
David Hill was also there, together with a young and energetic Welshman, Richard Edwards,
whom I had not met before. When I mentioned that I was looking for somewhere to work
the following spring when Murdoch would be in Cambridge on sabbatical leave, Richard
invited me to join him at the Royal Postgraduate Medical School at Hammersmith. There I
also met and worked with David Jones. We had a fabulous time that year. I became fascinated
with the changes in muscle properties induced by prolonged activity, and more particularly,
by the strategies that the nervous system must adopt if effective control is to be maintained.
Looking Back 9

Thus my visit to Richard's laboratory completely changed the direction of all my future
research, and all because of one lucky lunch-time encounter.

MY MORE RECENT COLLEAGUES

I have talked about a few of the many people who set me on my way in my early
days, and whose examples I have tried to emulate in later life. But it is the folk with whom
I have worked more recently who should, of course, get whatever credit the passage of time
and my peers may choose to attribute to the work on fatigue performed in my laboratory. It
is they who made everything possible and exciting.
First, Joe Woods, my stalwart ally from the beginning. I met him on my first day at
Quinnipiac. From then on, we instinctively turned to each other for advice and support in
every aspect of our professional lives - teaching, student problems, grant writing and all our
earlier experiments, both at Quinnipiac and later at the John B. Pierce Laboratory. Without
his help and encouragement, we could never have got started. It was a most unhappy day
when Quinnipiac chose to make him Dean and our collaboration had to end.
Then my thanks to my various post-doctoral students, each with me sequentially for
2-3 years: Carl Kukulka, Franyois Bellemare, Christine Thomas, Andy Fuglevand, and now
Michael Walsh. They all came from such different backgrounds, and brought with them such
different skills and personal philosophies. I have enjoyed knowing each one and have learned
so much from them. I also appreciate, more than I can say, the contributions of the many
scientists from other laboratories with whom I have collaborated, either when they visited
us or we went to them. There are too many to name individually, but among them I want to
mention, with particularly deep gratitude and affection, Olof Lippold, Roland Johansson,
Nina V011estad and Simon Gandevia who have collaborated with us repeatedly. In addition
to their well-known scientific talents, they also gave to me a very special friendship. And,
of course, my husband, Murdoch Ritchie. Without his constant support, encouragement and
often bullying I would never have had the courage to think that I could overcome the many
obstacles in the way of a mature, immigrant woman wishing to return to academia after being
absent from the action for so long.
THE SCIENTIFIC CONTRIBUTIONS OF
BRENDA BIGLAND-RITCHIE

C. K. Thomas, I R. M. Enoka,2 S. C. Gandevia,3 A. J. McComas,4 and

D. G. Stuart5

I The Miami Project to Cure Paralysis

University of Miami School of Medicine
Miami, FL 33136
2 Department of Biomedical Engineering, Cleveland Clinic Foundation
Cleveland, Ohio 44195
3Prince of Wales Medical Research Institute
Randwick, New SouthWales 2031, Australia
4Department of Biomedical Sciences, McMaster University
Hamilton, Ontario, L8N 3Z5, Canada
5 Department of Physiology, University of Arizona College of Medicine
Tucson, Arizona 85724

ABSTRACT

Brenda Bigland-Ritchie has made seminal contributions to our understanding of skeletal

muscle physiology - the energy cost of muscle when it shortens or is forcibly stretched. the
relationship between EMG and force, the behavior of single motor units, and above all, the
processes underlying neuromuscular fatigue. More than this, she has stimulated inquiry into
the search for reflex mechanisms which may serve to balance the activity of the spinal cord
with that of the fatiguing muscles. Her use of human volunteers for much of this work is
extraordinary, and represents a major strength. Equally important are her well known and
widely cited manuscripts. Not only are her findings clearly described and depicted, but every
attempt is made to relate her results to the fatigue processes measured in animal or isolated
tissue preparations.

INTRODUCTION

Fatigue is common, complex, and controversial (Bigland-Ritchie, 1981 b). Common

in that it confronts us in our daily tasks. Complex because the processes which fail vary with
the exercise performed. Controversial due to discrepancies in its definition and interpreta-
tion. Despite this complexity, Brenda Bigland-Ritchie has heightened our awareness of
neuromuscular fatigue as a subject for scientific study and has shown it to be approachable
through physiological experimentation. It has taken a careful mix of inquiry, use of effective

11
12 c. K. Thomas et al.
techniques, and collaboration. In this chapter, five themes are addressed. These include:
understanding muscle function from measures of oxygen uptake, examination of the sites of
fatigue, exploration of the difficulties in interpreting EMG signals, the reciprocity of
influences between spinal cord and muscle, and examples of how effective methods and
collaboration can enhance our understanding of muscle fatigue. In viewing these areas, we
highlight some of the findings of Brenda Bigland-Ritchie and propose why we think they
are so valuable.

PARTNERSHIPS BETWEEN MUSCLE, HEART AND LUNG

Brenda Bigland began her research when there was a systemic view towards under-
standing muscle function. At the time, work done was typically correlated to oxygen uptake.
But with the elegant demonstrations of the force-velocity relationships of striated muscle
(Hill, 1938), it became possible to relate work in terms of the number of active muscle fibers
and their activation rate. Since each fiber exerted a larger force during stretch than during
shortening (Hill, 1938; Katz, 1939), varying the numbers of active fibers could explain any
differences in energy cost for these two activities. Brenda Bigland's first scientific challenge
was to determine the oxygen requirements of shortening and lengthening contractions
(positive and negative work, respectively). To do this, it was necessary to find movements
which were mirror images of each other. Two bicycles were linked together back to back
(see Chapter, Looking back). To ensure that the placement and timing of activity in each
limb was identical, one person pedaled forward while the other pedaled in reverse. Not only
was the energy cost of positive work much greater than that of negative work (Abbott et aI.,
1952), but for the same rate of work, oxygen uptake was many times greater for work
involving great force and low speed compared to work involving small force and high speed
(Abbott & Bigland, 1953). All these data were explained in terms of fewer fibers being
needed to perform negative compared to positive work. In terms of muscle energetics, they
also substantiated the concept that muscle shortening was not the converse of muscle
lengthening (Hill, 1938; Abbott et aI., 1951).
These same concepts were explored further some twenty years later when Brenda
Bigland-Ritchie re-entered science. The energy requirements of the fibers (rate of metabolic
heat production and rate of ATP and phosphocreatine breakdown) had been shown to fall
substantially during stretch compared to during shortening (Wilkie, 1968; Curtin & Davies,
1970). Thus, during positive and negative work, were there differences in the number of
active fibers and in the energy requirements per fiber? With a new and improved motor-
driven ergometer which better confined the pedaling work to the specific muscles involved
(Bigland-Ritchie et aI., 1973), oxygen uptakes were compared to measures of integrated
surface EMG (IEMG) (Bigland-Ritchie & Woods, 1974). If the IEMG could provide a
reliable measure of muscle activity during voluntary contractions, it was argued that it would
be possible to examine changes in oxygen uptake per active muscle fiber. Interestingly, the
IEMG and particularly the oxygen uptake were less during muscle stretching versus
shortening (Fig. 1, open and closed symbols respectively). Thus, those muscle fibers which
were active required less oxygen when they were stretched compared to when they were
shortened (Bigland-Ritchie & Woods, 1976).
These manuscripts also signaled the reappearance of a well defined scientific style
which has continued in ever-sharpening focus in her work on neuromuscular fatigue.
However, in examining detail at the whole muscle and motor unit levels, Brenda Bigland-
Ritchie is unusual in that she never assumes that the particular test muscle works in isolation.
Rather, the entire organism may actually determine how well a particular part functions. For
example, co-ordination between different body systems was the mechanism proposed to
The Scientific Contributions of Brenda Bigland-Ritchie 13

3 A c 800 B
j
§ 600
8
gw 400
Figure 1. Mean oxygen uptake (A) and -g
integrated EMG (B) versus mean torque ~ 200

1
on the pedals (or work rate) for positive
(filled symbols) and negative work
(open symbols). All data are from one I

subject (n=75 experiments conducted 2 3

over several months) and only data from Torque on pedals, kg.m
the left leg are shown in B. Pedaling was i i i I i i I

at 50 rev/min. (adapted from Bigland-

300 600 900 300 600 900 1200
Ritchie & Woods, 1976). Work rate, [Link]

explain why dynamic but not isometric exercise performance was disproportionately im-
paired under hypoxic conditions (Bigland-Ritchie & Vellestad, 1988). These kinds of
hypotheses are invaluable in that they pose novel ways of interpreting data and they serve
to inspire new experiments.

FINDING THE FACTORS THAT ARE RATE LIMITING IS AS

IMPORTANT AS FINDING THOSE THAT ARE NOT
One hallmark of Brenda Bigland-Ritchie is her strong advocacy for evaluating each
possible fatigue site rather than assuming only one is at fault and others are intact. Most
discussions of fatigue sites focus on three possible areas: central fatigue, failure at the
neuromuscular junction, and deficits in the muscle. Each area can be further subdivided to
include other processes as shown in Fig 2A. But which of these processes fail, when, and
how? Furthermore, is this rather unidirectional depiction of events, starting with the
excitatory input to higher motor centers and ending with the energy-dependent interaction
of actin and myosin, much more complex and interactive?

Central Fatigue
Early on, there were objective methods to evaluate whether a contraction was the
strongest possible, and whether voluntary drive could be maintained during sustained efforts.
For example, when a contraction was produced by voluntary effort or by tetanic nerve
stimulation, similar forces were generated (Fig. 2B; Mosso 1915; Bigland & Lippold, 1954b;
Merton 1954). Such findings were taken to indicate that fatigue was a peripheral phenome-
non. The method of twitch occlusion originally suggested by Denny-Brown (1928) was
adapted by Merton (1954) to assess contraction maximality. As voluntary force increased
within a muscle, that force which could be evoked artificially by stimulation was reduced,
then occluded completely during a maximal voluntary contraction. Could these initial
findings be reproduced for different muscles, and quring other types of contractions?
It is usually possible to activate a muscle maximally during brief contractions but
during fatiguing contractions this is much more difficult (Bigland-Ritchie et aI., 1978;
Belanger & McComas, 1981; Gandevia & McKenzie 1988b; Thomas et aI., 1989), particu-
larly for the diaphragm during expulsive maneuvers (Bellemare & Bigland-Ritchie, 1987).
While the latter difficulty may arise because of the rather unphysiological nature of this
14 C. K. Thomas et al.

A C 0

'ih
Time,s
a 0 20
o per
5 o ATP
15 ~ .La
e-... 15
9 ~
0
d./ 25 E
E
35 c 10
musdefiber

B
45 ~
'E

1flill
60·· CD
0
c: 5
0
90 0

105
20 40 60
120N !2mV Time, min
10s 4ms
Figure 2. A, potential fatigue sites: a, excitatory input to the motor cortex; b, excitatory drive to the lower
motoneuron; c, motoneuron excitability; d, neuromuscular transmission; e, sarcolemma excitability; f, exci-
tation-contraction coupling; g, contractile mechanism; h, metabolic energy supply. B, 60 s maximum voluntary
contraction of adductor pollicis which was interrupted periodically with 50 Hz stimulation (cross-hatching).
Similar forces were always generated by stimulation and voluntary effort indicating an absence of central
fatigue. C, intramuscular M-waves recorded during a fatiguing co-contraction of both adductor pollicis and
first dorsal interosseous. At 60s (**) the M-wave declined. It was restored by repositioning the stimulating
electrode (90s). D, mean metabolite concentrations (lactate, ATP, Per) changes during repeated target force
(30% MVC) contractions performed for 6s followed by 4s rest. Exhaustion was reached at 55 minutes. Data
are from 5 subjects. (A-D adapted from Bigland-Ritchie, 1981a, 1983b, 1982, and 1986a, respectively).

behavior (McKenzie et aI., 1992), it may also relate to the need for the diaphragm to continue
functioning without contractile failure. As such, these data raise the intriguing possibility
that there are substantial interactions between the potential failure sites, a phenomenon which
needs our attention.
In this type of research there was a slow evolution of the methods needed for the
assessment of central fatigue. To avoid injury, supramaximal nerve stimulation was replaced
by percutaneous stimulation of a constant fraction of the muscle (Bigland-Ritchie et aI.,
1978). When bilateral tetanic phrenic nerve stimulation was too painful and inconsistent, the
already established method of twitch occlusion was extended to evaluate muscle activation
during sustained contractions (Bellemare & Bigland-Ritchie, 1987). Resolution of the small
evoked forces was also a concern. Thus, voltage clamp circuits (Belanger & McComas 1981)
or those which amplified only the AC force components were developed (Hales & Gandevia
1988; Westling, Johansson & Bigland-Ritchie, unpublished). In some cases, two briefly
separated stimulation pulses were delivered to enhance the force output (Gandevia &
McKenzie, 1985, 1988a). Another innovation was electrical stimulation of the motor cortex
during fatiguing contractions (Merton et aI., 1981). With no sign of electrical decrement
during these protocols, both central and neuromuscular junction excitability must have
remained intact, confirming support for the peripheral nature of fatigue. Such methods are
also critical for evaluating central fatigue in neurological disorders which can partially
disconnect the central and peripheral nervous systems (for example, spinal cord injury).
Then, the same intact innervation which can be driven by voluntary effort is also stimulated
(Thomas, 1993).
The Scientific Contributions of Brenda Bigland-Ritchie 15

Neuromuscular Junction and Fiber Plasmalemma

Demonstration of junction and/or plasmalemma failure requires comparisons be-
tween the responses evoked by nerve stimulation and direct muscle stimulation. Since the
latter is difficult in human experiments due to the large voltages required, measurements of
M-waves have largely been used to evaluate integrity of the neuromuscular junction and
plasmalemma (Merton, 1954). This in itself has brought controversy, primarily because
investigators have measured and interpreted the waveforms differently (Bigland-Ritchie,
1987). Thus, while many studies have concluded that there is little evidence to support
junction or plasmalemma failure as sites for fatigue (for example, Merton, 1954; Bigland-
Ritchie et aI., 1982; Duchateau & Hainaut, 1985; Thomas et aI., 1989), other studies conclude
the contrary (for example, Stephens & Taylor, 1972; Aldrich et aI., 1986; Milner-Brown &
Miller, 1986; Bellemare & Garzantiti, 1988). On several occasions Bigland-Ritchie has
attempted to resolve these differences by repeating the experiments and eliminating meth-
odological differences. Two important factors seem to have arisen from this. First, the
M-wave amplitude can often be restored by repositioning of the stimulating electrode so
depressed M-waves do not always imply failure (Fig. 2C). Second, any M-wave measure-
ments must take into account the slowing of impulse conduction velocity which occurs with
fatigue.
An alternative way of evaluating junction integrity has involved examining the
regularity of motor unit interspike intervals. While continuity of impulses suggests junction
integrity, the possibility of central fatigue must be precluded. That motor unit firing rates
during maximal voluntary contractions rarely seem to exceed those stimulation rates where
failure has been shown to occur (Krnjevic & Miledi, 1958; Bellemare et aI., 1983), also lends
plausibility to the idea that neuromuscular junctiOli block may indeed be rare during
voluntary contractions.

Muscle
Use of muscle biopsies, NMR spectroscopy and skinned fiber preparations have
added much to our understanding of muscle metabolism during human fatigue. However,
study of other processes beyond the neuromuscular junction has come largely from animal
studies (reviewed by Fitts, 1994). In this area, Bigland-Ritchie has constantly urged others
to proceed, but with standardized protocols. Then direct comparisons between human and
animal data are possible, and human data can be more fully interpreted.
Perhaps the most difficult issue to address has been the potential failure of excita-
tion-contraction coupling. It is often singled out as a contributing factor because of the
disproportionately low force in response to low versus high frequencies of activation
(Edwards et aI., 1977). Alternatively, it has been identified as a failure site by exclusion.
During the first 30 min of repeated low intensity submaximal intermittent contractions of
quadriceps, little change was seen in muscle metabolites in the presence of intact central
drive and neuromuscular junction function (Fig. 2D). Only failure of excitation-contraction
coupling was left to explain the data (Bigland-Ritchie et aI., 1986a; V 0llestad et aI., 1988).
However, as the exercise proceeded, metabolic changes did occur and revealed the novel
finding that exhaustion was closely related to depletion of phosphagen stores (V0llestad et
aI., 1988).
However, no matter what the conclusion, Brenda Bigland-Ritchie has always left
room to question the importance of the findings. For example, are the proposed fatigue sites
linked directly to the force loss, or are the observed changes by-products of other processes?
Similarly, only a few studies have examined fatigue during dynamic contractions (for
example, de Haan et aI., 1989) but the relative importance of central and peripheral factors
16 C. K. Thomas et al.

in these conditions deserves the same systematic investigation that has been given to
isometric contractions.

UNSCRAMBLING THE ELECTROMYOGRAM

SurfaceEMG
Throughout the years Bigland-Ritchie has made careful and informative campaigns
to clarify our understanding ofEMG-force relationships. During brief voluntary isometric
or constant velocity, shortening or lengthening contractions, the relationship between the
integrated surface EMG and force can be linear (Lippold, 1952; Bigland & Lippold, 1954a;
Milner-Brown & Stein, 1975) or non-linear depending on the muscle examined (Fuglsang-
Fredriksen, 1981; Woods & Bigland-Ritchie, 1983). Cautionary notes suggest that Bigland-
Ritchie, as well as others, thought some often overlooked factors which could explain the
discrepant results. For example, the surface record may not reflect the activity of the muscle
in general because fiber distributions are uneven for different units and the signals from deep
units are attenuated at the surface (Clamann, 1970). Potentials may also sum in a non-linear
way yet produce linear results (Milner-Brown & Stein, 1975). Similarly, if some motor units
are already firing at optimal rates for force development when the whole muscle force is
submaximal, further effort may increase the impulse frequencies of these units without
adding tension (Tanji & Kato, 1973a,b). Alternatively, force and/or EMG signals may be
contaminated from synergists or antagonists. Different roles of recruitment and rate coding
in various muscles or during different tasks may also change the overall EMG-force
relationship (Bigland-Ritchie, 1981a; Woods & Bigland-Ritchie, 1983).
Surface EMG changes with fatigue are equally complex. While motor unit synchro-
nization (Lippold et aI., 1960; Lippold, 1973) and muscle conduction velocity slowing with
fatigue may lead to increases in EMG, reductions in either motor unit rate or in recruitment
may lead to decreases. How these factors influence the surface EMG, and change for different
muscles, types of exercise, degrees of ischemia and recovery is still unclear. Thus surface
EMG recordings alone leave too much room for speculation on the possible fatigue site
(Bigland-Ritchie et aI., 1978; Bigland-Ritchie, 1981b, 1987; Bigland-Ritchie & Woods,
1984).
Careful evaluation of much published data from evoked or voluntary contractions
has also brought the whole EMG-force relationship into question. For example, how can the
EMG potentiate while the force fails? Alternatively, how can the EMG almost disappear with
minor decrements in force? What mechanisms underlie dramatic changes in EMG-force
relations which accompany minor increases in use? (see Kernell et aI., 1987; Hicks et aI.,
1989; Thomas et aI., 1991a). These kinds of data have prompted a thought-provoking
conclusion: "no fixed relation is to be expected between force and EMG" (Bigland-Ritchie,
1988).

Intramuscular EMG
Surface EMG recordings cannot distinguish the importance of the number of active
muscle fibers versus the frequency of excitation in force generation (Bigland & Lippold,
1954b). But that kind of information was necessary to decipher the pattern of motor unit
recruitment during a brief voluntary contraction, as well as the rate at which units responded.
Early motor unit research documented an orderly sequence of motor unit recruitment,
showed that it was dependent on the task being performed, and characterized unit discharge
rates during weak contractions (Adrian, 1925; Adrian & Bronk, 1929; Smith, 1934; Lindsley,
The Scientific Contributions of Brenda Bigland-Ritchie 17

E 100 A 100 B
g 80
.~ 60
Figure 3. A , abductor digiti minimi tension ~
c- 40
(relative to maximum) compared to minimum
o
and maximum unit discharge frequencies (left 'iii
c
20
and right curves respectively) during voluntary CIJ
I-- 0
contractions (closed symbols) and after almost 0 10 20 30 40 50 50
complete pressure block of the ulnar nerve near Frequency, Hz
the elbow (open symbols). S, typical force·fre-
quency behavior proposed for single motor 30 D
units during voluntary contractions. C, poten-
tials from one motor unit that change in ampli- ~
tude as the tungsten electrode is moved slowly
:5 20
10 mV [ (ij
through the muscle. The instantaneous unit rate (5
is below. D. distributions of unit firing rates I-- 10
~
during maximum voluntary contractions ofbi- 40 [
ceps brachii for two subjects. (A.S adapted Hz
0
from Bigland & Lippold, 1954b; C.D adapted 20 .----.. : 0 20 40 60
0.55 Frequency, Hz
from Bellemare et aI., 1983).

1935; Denny-Brown & Pennybacker, 1938; Gilson & Mills, 1941). But, how did unit
discharge frequencies vary for different units, and how were these frequencies altered with
greater increases in muscle force?
Two tacks were taken to explore what was going on during stronger voluntary
contractions. The first was to develop selective recording electrodes from fine wire. The
other was to block activity in some axons by local nerve pressure (Bigland & Lippold,
1954b). These strategies taught us that units of higher threshold were recruited at higher
firing rates and that each unit followed a sigmoid-shaped force-frequency relation. Lower
threshold units reached the same maximal firing frequencies as high threshold units but at
lower whole muscle forces (Fig. 3A,B). Although the maximal firing rates were not measured
consistently, they suggested that only units recruited at the highest forces fired above 35 Hz.
All these data were important in providing the first detailed analysis of unit rate coding in
human muscle. Moreover, they gave a framework from which to examine further the role of
frequency modulation and recruitment in force generation.
Substantial contributions have subsequently been made by others. For example, how
recruitment and rate coding varies for different muscles (Kukulka & Clamann, 1981),
additional detail on how unit frequency changes with force (Monster & Chan, 1977), the
finding that human units are activated in order of increasing force output during isometric
contractions which are slow (Milner-Brown et aI., 1973) or fast (Desmedt & Godaux, 1977),
and the observation that this pattern can be altered during eccentric contractions (Nardone
et aI., 1989; Howell et aI., 1994). However, despite the current vast literature on these topics,
it is interesting to note that our current models of motor unit behavior (Heckman & Binder
1993) still rely greatly on the data described by Bigland and Lippold (1954b).
Extracting motor unit potentials from the complex interference pattern of the elec-
tromyogram (EMG), particularly at maximal forces, still represents a major challenge.
Although comparisons between voluntary and stimulated forces had shown that human
voluntary drive can maximally activate all motor units in a muscle (Mosso, 1915), it was
unclear what the unit firing rates accompanying such efforts actually were. Experiments
involving partial nerve block, aberrant motor innervation, and partial denervation suggested
that unit firing rates may range from 35 to 60 Hz, and go up to 200 Hz (Bigland & Lippold
18 c. K. Thomas et al.
1954b; Marsden et aI., 1971, Grimby & Hannerz, 1977; 1983). But, were these extraordinary
rates maintained beyond the first few impulses, and were these compromised situations
providing accurate information? Use of fine, high impedance, tungsten electrodes typically
used for microneurographic recordings (Vallbo et aI., 1979) gave the answers (Fig. 3C,D).
Not only did units vary markedly in their maximal rates (12-60 Hz for adductor pollicis and
biceps brachii units), between subjects (mean maximal rates between 22-35 Hz for biceps
brachii) and between muscles (5-20 Hz for soleus), but these rates were rather slow compared
to those needed to evoke maximal muscle force by stimulation (Bellemare et aI., 1983;
Bigland-Ritchie et aI., 1983c). Thus, for the first time, the limits of frequency modulation
during maximal voluntary contractions had been defined for the population of units in a
muscle. A future challenge is to determine which firing rates belong to units recruited at low
versus high forces.
With a convenient method to measure motor unit firing rates during maximal
voluntary contractions, it was now possible to evaluate whether these rates declined during
sustained contractions as predicted. Earlier experiments involving stimulation showed that
the force and EMG decline mimicked voluntary force and EMG loss if the stimulation rates
were gradually reduced (Fig. 4A-C; Marsden et aI., 1969; Bigland-Ritchie et aI., 1979; Jones
et aI., 1979). The improved recordings showed that unit discharge did decline during
maximal voluntary contractions such that the time courses of the force, contractile speed and
rate reductions closely followed each other (Fig. 4D). As there was no evidence for central
fatigue or neuromuscular junction failure, the force loss must have resulted from contractile
failure (Bigland-Ritchie et aI., 1983a,b). How then could muscle be activated optimally with
lower and lower motor unit firing rates? The obvious hypothesis was that contractile slowing
meant lower frequencies were adequate for full tetanic fusion (Fig. 4C). Maintenance of
higher rates would actually be disadvantageous, since energy would be wasted in additional
ion pumping and excitation failure would be more likely to occur. However, were these
changes in firing rate mere coincidence, or were they under regulatory control?

Figure 4. A, force from adductor pollicis during 80 Hz and 20 Hz (solid lines) stimulation of the ulnar nerve.
Muscle ischemia was maintained throughout the protocol. Note that 20 Hz stimulation produces more force
than 80 Hz towards the end of the contraction. B, mechanical responses to ulnar nerve stimulation after a 5s
MVC (upper records) and after a 60s MVC (lower records) with blood occlusion. Stimulation was at 1, 7 and
50 Hz in each case (left to right re-
A cords respectively). After fatigue the
B Control
7 Hz tetanus/twitch ratio increased

~yMM].J\J. from 1.67 to 2.50. The force in re-

sponse to 7 Hz stimulation increased

nI.
I i After 60s MVe~
(a/c) by 15%, while the fraction of the

~
~ .-r l' l _[SN
100 me
J
108
maximum available force (alb) in-
['5N creased from 0.17 to 0.24%. c.
smoothed rectified EMG record from
adductor pollicis during a maximum
c o voluntary contraction and a stimu-
'21 lated contraction (smoother line) in
~C which the frequency was decreased as
::J '00
shown (the amplification was four
~
I!! times greater for the MVC). D,
:e
~'.[
changes (% initial) in mean mo-
toneuron firing rates and relaxation
21 during a 60s MVC. (A-D adapted
a: ~
W 0
(JJ ... from Jones et ai., 1979, Bigland-
10 • • ~ ~ m Jo 40 eo 80 100 Ritchie et ai., 1983b, 1979 and 1983a
Hz Time,s respectively).
The Scientific Contributions of Brenda Bigland-Ritchie 19

MUSCLE CONTROL, CNS CONTROL OR BOTH?

Fatigue has long been known to involve force loss and contractile slowing (Hill,
1913; Abbott, 1951). Yet what regulates and limits these processes? Moreover, how do these
muscle responses integrate with other changes, and particularly those in the central nervous
system? For example, what factors underlie the decline in motoneuron firing rates during
sustained maximal voluntary contractions when there is adequate central drive and
neuromuscular transmission? On some level, the possibility of reflex control from the muscle
was evident to many. However, it needed to be shown. With the extraordinary dedication of
the research volunteers, it was shown that the force and motoneuron discharge reductions
which occur in response to sustained maximal voluntary contractions fail to recover if the
fatigued muscle is kept ischemic (Fig. 5). Thus, it was proposed that factors from the
exercised muscle must signal its fatigued character, and that discharge of motoneurons at
the initial activation rates would confer no functional advantage to the slowed muscle
(Bigland-Ritchie et aI., 1986b). Viewed as such, these data provide a way to understand how
a fatiguing muscle can continue to work optimally despite a decline in the firing rates of its
constituent motor units. The argument for this reflex was then substantiated further by
eliminating the possibility that transmission fails at the neuromuscular junction (Woods et
aI., 1987). These results perhaps provide the most persuasive evidence we have that dynamic
factors in the muscle can play critical roles in motoneuron regulation.
One gauge of the importance of this work was the impetus it gave for new experi-
ments. A flurry of new questions were immediately apparent. For example, which afferents
contribute to these effects and under what circumstances? How widespread are these reflex
effects in the spinal cord and higher brain centers? What triggers induce these reflex effects?
Do independent changes in motoneuron excitability (Kernell & Monster, 1982) also playa
role? Furthermore, how important are these reflexes when the exercise is submaximal or
induced by stimulation? We soon learned that the reflex effects were not specific to the
fatigued muscle but also carried over to synergistic muscles (Hayward et aI., 1988; McComas
et aI., 1994), though the effect may vary in magnitude for different muscles or tasks (Dacko
& Cope, 1994). Similarly, muscle slowing induced by cooling or length changes failed to
reproduce the changes in motor unit firing rates typically measured during fatigue (Bigland-
Ritchie et aI., 1992a,b). Whether ischemia or metabolic factors are of more importance still
needs to be evaluated. Similarly, amid the speculation that group III and IV afferents were

A. Activation B. Force C. Firing rate

MVC: 2 3 2 3 2 3

02040020020 02040020020 02040 020020

Time,s

Figure 5. Mean quadriceps muscle activation (A), force (B) and unit firing rate (C) expressed relative to initial
control values. Subjects performed a 40s maximum voluntary contraction (MYC I) then two 20s maximum
voluntary contractions (MYC2, MYC3) with a 3 minute rest between each contraction. The muscle was kept
ischemic during MYCI, the 3 min rest and during MYC2 as indicated by the solid line (adapted from Woods
et aI., 1987).
20 C. K. Thomas et aI.

involved in the reflex (Bigland-Ritchie et aI., 1986b; Woods et aI., 1987), came evidence
that group Ia spindle afferents may also contribute (Hagbarth et aI., 1986) and that both
excitatory and inhibitory influences seem to playa role (Gandevia et aI., 1990). Clearly the
entire significance of fatigue-induced control of motoneuron firing rates will only be
appreciated after the completion of many more experiments.

METHODS AND COLLABORATIONS YIELD REWARDS

Apart from brief forays into experiments which examined the effects of growth
hormone on muscle performance (Bigland & Jehring, 1952), how temperature influences
the effects of neuromuscular blocking drugs (Bigland & Zaimis, 1958; Bigland et aI., 1958),
and the functional characteristics of local anesthetics (Ritchie et aI., 1965a,b; Ritchie &
Ritchie, 1968), Brenda Bigland-Ritchie's scientific contributions have all come from human
experiments. Thus, what some perceived as limiting, that is, data from human experiments,
she used to great advantage. Some of the value lay in viewing whole organism behavior
without anesthesia and surgery, and some in examining the actual limits of performance and
how it was controlled. The experimental strategy included the willingness to experience these
limits first hand, the ability to recognize the talents of other people and to pull them into
collaborations, and effective methodology. For example, when challenged to find a task
involving the same work, movements and speed to measure the oxygen cost of positive and
negative work, what more elegant solution could there have been than the two bicycles
depicted on the frontispiece of this book (Abbott et aI., 1952)? Similarly, when individual
motor unit potentials could not be discerned from the interference pattern during high force
voluntary contractions, a steady stream of innovations brought her closer to the eventual
goal: the development of more selective electrodes inserted intramuscularly, partial nerve
block from pressure (Bigland & Lippold, 1954b), counting unit potentials over a fixed
amplitude (Bigland-Ritchie & Lippold, 1979), and the use of tungsten microelectrodes to
record the firing rates of populations of motor units during maximal voluntary contractions
(Bellemare et aI., 1983; Bigland-Ritchie et aI., 1983a,c). When bilateral phrenic nerve
stimulation was needed to describe the contractile properties of the diaphragm, and tetanic
stimulation was too painful and inconsistent, the method oftwitch occlusion (Denny-Brown,
1928; Merton, 1954; Belanger & McComas, 1981) was extended to evaluate muscle
activation during brief and fatiguing voluntary contractions (Fig. 6A; Bellemare & Bigland-
Ritchie, 1984, 1987; Bellemare et aI., 1986). More recently, Bigland-Ritchie initiated and
saw to fruition the successful development of an alternative technique for measuring human
motor unit contractile properties. Intraneural motor axon stimulation permitted the first
detailed measurements of the entire twitch and tetanic force responses of the popUlation of
units in a muscle without signal averaging (Fig. 6B; Westling et aI., 1990). Moreover, it
became possible to determine how the force and EMG of units respond to defined activation
patterns (Fig. 6C; Thomas et aI., 1991a,b), bringing us closer to understanding differences
between human and other mammalian muscles.

PICTURES, WORDS AND QUESTIONS

It is a special privilege to know Brenda Bigland-Ritchie. With both word and graph,
her papers address well defined issues, are characterized by cautious interpretation and, at
the same time, show her willingness to share ideas and to inspire new experiments. These
manuscripts are worth investigating. And, if we ask who has advanced our knowledge of
The Scientific Contributions of Brenda Bigland-Ritchie 21

A c
25

0.5mV [ --fII--I'fIr-~Ir--"fV"-"",,",~-
240
200

&. 160 .e 15
::c 120 8'
5 I
I
CD
10
I
~80 ::::J

~ ~o
I
40
o ~ 5
100ms 2
Run

"'" II' IIII!I ,

1s 0.5s 0.5s

Figure 6. A, left EMG (upper) and Pdi (lower) records when maximal shocks were delivered to the phrenic
nerve bilaterally during combined Mueller-expulsive maneuvers at different %Pdimax in the sitting position.
The vertical dashed lines mark the time of stimulation delivery. Note the magnitude of the response reduces
as Pdi level increases. B, resultant force records from one thenar motor unit when its axon was stimulated with
trains of pulses at 1,15 and 50 Hz. C, frequencies that produced 50% of initial maximum tetanic force for 12
different thenar motor units before (run 1) and after (run 2) a Burke fatigue test. (A adapted from Bellemare
& Bigland-Ritchie, 1984, B-C adapted from Thomas et aI., 1991a).

neuromuscular fatigue, Dr. Bigland-Ritchie is not the only scientist who has. But clearly, she
has made significant and imaginative contributions, perhaps more than anyone.

ACKNOWLEDGMENTS
The authors thank Ms. Bette Mas and Dr. Mary Pocock for editorial assistance and
much library work, and Drs. Vladimir Esipenko and James Broton for help with the figures.
The author's research is currently supported by: United States Public Health Service
(USPHS) grant NS 30226, the Department of Defense, and The Miami Project to Cure
Paralysis (C.K.T.); USPHS grants AG 09000 and NS 20544 ([Link].); the National Health
& Medical Research Council of Australia and the Asthma Foundation of New South Wales,
Australia (S C. G.); the Medical Research Council and the Natural Sciences and Engineering
Research Council (NSERC) of Canada ([Link].), and USPHS grants GM 08400, NS
20577, NS 01686 and NS 07309 (D.G.S). The authors' attendance at the 1994 Bigland-
Ritchie conference was supported, in part, by NSERC ([Link].), NS 30226 (C.K.T.), the
Cleveland Clinic Foundation ([Link].), the Muscular Dystrophy Association (USA; Sc.G.),
University of Arizona Regents' Professor funds (D.G.S), and the University of Miami
(S C. G. and [Link].).
22 c. K. Thomas et aI.

REFERENCES
Abbott BC (1951). The heat production associated with the maintenance of a prolonged contraction and the
extra heat produced during large shortening. Journal ofPhysiology (London) 112,438-445.
Abbott BC, Aubert X & Hill AV (1951). Changes of energy in a muscle during very slow stretches. Proceedings
of the Royal Society, Series B 139, 104-117.
Abbott BC & Bigland B (1953). The effects of force and speed changes on the rate of oxygen consumption
during negative work. Journal of Physiology (London) 120, 319-325.
Abbott BC, Bigland B & Ritchie JM (1952). The physiological cost of negative work. Journal of Physiology
(London) 117, 380-390.
Adrian ED (1925). Interpretation of the electromyogram. Lancet 208, 1229-1233.
Adrian ED & Bronk DW (1929). The discharge of impulses in motor nerve fibres. Part II. The frequency of
discharge in reflex and voluntary contractions. Journal of Physiology (London) 67,119-151.
Aldrich TK, Shander A, Chaudhry I & Nagashima H (1986). Fatigue of isolated rat diaphragm: role of impaired
neuromuscular transmission. Journal of Applied Physiology 61, 1077- 1083.
Belanger AY & McComas AJ (1981). Extent of motor unit activation during effort. Journal of Applied
PhYSiology 51, 1131-1135.
Bellemare F & Bigland-Ritchie B (1984). Assessment of human diaphragm strength and activation using
phrenic nerve stimulation. Respiration Physiology 58, 263-277.
Bellemare F & Bigland-Ritchie B (1987). Central components of diaphragmatic fatigue assessed by phrenic
nerve stimulation. Journal ofApplied PhYSiology 62,1307-1316.
Bellemare F, Bigland-Ritchie B & Woods JJ (1986). Contractile properties of the human diaphragm in vivo.
Journal ofApplied Physiology 61, 1153-1161.
Bellemare F & Garzaniti N (1988). Failure of neuromuscular propagation during human maximal voluntary
contraction. Journal ofApplied Physiology 64, 1084-1093.
Bellemare F, Woods JJ, Johansson R & Bigland-Ritchie B (1983). Motor-unit discharge rates in maximal
voluntary contractions of three human muscles. Journal of Neurophysiology 50, 13 80-1392.
Bigland B, Goetzee B, MacLagan J & Zaimis E (1958). The effect oflowered muscle temperature on the action
of neuromuscular blocking drugs. Journal of Physiology (London) 141,425-434.
Bigland B & Jehring B (1952). Muscle performance in rats, normal and treated with growth hormone. Journal
ofPhysiology (London) 116, 129-136.
Bigland B & Lippold OCJ (1954a). The relation between force, velocity and integrated electrical activity in
human muscles. Journal of Physiology (London) 123, 214-224.
Bigland B & Lippold OCI (l954b). Motor unit activity in the voluntary contraction of human muscle. Journal
ofPhysiology (London) 125, 322-335.
Bigland B & Zaimis E (1958). Factors influencing limb temperature during experiments on skeletal muscle.
Journal of Physiology (London) 141,420-424.
Bigland-Ritchie B (1981a). EMG and fatigue of human voluntary and stimulated contractions. In: Porter R,
Whelan I (eds.), Human Muscle Fatigue: Physiological Mechanisms, pp. 130-156. London: Pitman
Medical.
Bigland-Ritchie B (1981b). EMG/force relations and fatigue of human voluntary contractions. In: Miller D
(ed.), Exercise and Sports Sciences Reviews (Vol. 9), pp. 75-117. Franklin Institute Press.
Bigland-Ritchie B (1987). Respiratory muscle fatigue: posters, methods, and more. In: Sieck GC, Gandevia
SC, Cameron WE (eds.), Respiratory Muscles and Their Neuromotor Control, pp. 379-389. New York:
Alan R. Liss.
Bigland-Ritchie B (1988). Nervous system and sensory adaptation. In: Bouchard C, Shephard RJ, Stephens T,
Sutton JR, McPherson BD (eds.), Exercise, Fitness and Health, pp. 377- 383. Champaign, IL: Human
Kinetics.
Bigland-Ritchie B, Cafarelli E & Vallestad NK (1986a). Fatigue of submaximal static contractions. Lars
Hermansen Memorial Symposium: Exercise in Human Physiology. Acta Physiologica Scandinavica
128 (suppl. 556), 137-148.
Bigland-Ritchie B, Dawson NJ, Johansson RS & Lippold OCJ (1986b). Reflex origin for the slowing of
motoneurone firing rates in fatigue of human voluntary contractions. Journal ofPhysiology (London)
379,451-459.
Bigland-Ritchie B, Furbush FH, Gandevia SC & Thomas CK (1992a). Voluntary discharge frequencies of
human motoneurons at different muscle lengths. Muscle & Nerve 15, 130- 137.
Bigland-Ritchie B, Graichen H & Woods JJ (1973). A variable-speed motorized bicycle ergometer for positive
and negative work exercise. Journal ofApplied Physiology 35, 739- 740.
The Scientific Contributions of Brenda Bigland-Ritchie 23

Bigland-Ritchie B, Johansson R, Lippold OCJ, Smith S & Woods JJ (1983a). Changes in motoneurone firing
rates during sustained maximal voluntary contractions. Journal ofPhysiology (London) 340, 335-346.
Bigland-Ritchie B, Johansson R, Lippold OCJ & Woods JJ (1983b). Contractile speed and EMG changes
during fatigue of sustained maximal voluntary contractions. Journal ofNeurophysiology 50,313-324.
Bigland-Ritchie B, Johansson R & Woods JJ (1983c). Does a reduction in motor drive necessarily result in
force loss during fatigue? In: Knuttgen HG, Vogel JA, Poortrnans, J (eds.), Biochemistry ofExercise
(International Series on Sport Sciences Vol. 13), pp. 864-870. Champaign, IL: Human Kinetics.
Bigland-Ritchie B, Jones DA, Hosking GP & Edwards RHT (1978). Central and peripheral fatigue in sustained
maximum voluntary contractions of human quadriceps muscle. Clinical Science and Molecular
Medicine 54,609-614.
Bigland-Ritchie B, Jones DA & Woods JJ (1979). Excitation frequency and muscle fatigue: Electrical
responses during human voluntary and stimulated contractions. Experimental Neurology 64, 414-427.
Bigland-Ritchie B, Kukulka CG, Lippold OCJ & Woods JJ (1982). The absence of neuromuscular transmission
failure in sustained maximal voluntary contractions. Journal of Physiology (London) 330, 265-278.
Bigland-Ritchie B & Lippold OCJ (1979). Changes in muscle activation during prolonged maximal voluntary
contractions. Journal ofPhysiology (London) 292, 14P-15P.
Bigland-Ritchie B, Thomas CK, Rice CL, Howarth JV & Woods JJ (1992b). Muscle temperature, contractile
speed, and motoneuron firing rates during human voluntary contractions. Journal of Applied Physi-
ology 73, 2457-2461.
Bigland-Ritchie B & VliJllestad NK (1988). Hypoxia and fatigue: How are they related? In: Sutton JR, Houston
CS, Coates G (eds.), Hypoxia: The Tolerable Limits, pp. 315-328. Indianapolis, IN: Benchmark Press.
Bigland-Ritchie B & Woods JJ (1974). Integrated EMG and oxygen uptake during dynamic contractions of
human muscles. Journal ofApplied Physiology 36, 475-479.
Bigland-Ritchie B & Woods JJ (1976). Integrated electromyogram and oxygen uptake during positive and
negative work. Journal of Physiology (London) 260, 267-277.
Bigland-Ritchie B & Woods JJ (1984). Changes in muscle contractile properties and neural control during
human muscular fatigue. Muscle & Nerve 7, 691-699.
Clamann HP (1970). Activity of single motor units during isometric tension. Neurology 20, 254-260.
Curtin NA & Davies RE (1970). Chemical and mechanical changes during stretching of activated frog skeletal
muscle. Cold Spring Harbor Symposia on Quantitative Biology 37,619-626.
Dacko SM & Cope TC (1994). Patterns ofMG motor unit activity following fatigue ofa close synergist in the
decerebrate cat. Abstracts, International Symposium on Neural and Neuromuscular Aspects ofMuscle
Fatigue (Miami, FL, November 10-13, 1994, pp. 35. Miami, FL: Miami Project to Cure Paralysis,
University of Miami.
de Haan A, Jones DA & Sargeant AJ (1989). Changes in velocity of shortening, power output and relaxation
rate during fatigue of rat medial gastrocnemius muscle. European Journal of Physiology 413,
422-428.
Denny-Brown D (1928). On inhibition as a reflex accompaniment of the tendon jerk and of other forms of
active muscular response. Proceedings of the Royal Society Series B 103, 321-336.
Denny-Brown D & Pennybacker JB (1938). Fibrillation and fasciculation in voluntary muscle. Brain 61,
311-333.
Desmedt IE & Godaux E (1977). Fast motor units are not preferentially activated in rapid voluntary
contractions in man. Nature 267,717-719.
Duchateau J & Hainaut K (1985). Electrical and mechanical failures during sustained and intermittent
contractions in humans. Journal ofApplied Physiology 58, 942-947.
Edwards RHT, Hill DK, Jones DA & Merton PA (1977). Fatigue of long duration in human skeletal muscle
after exercise. Journal ofPhYSiology (London) 272, 769-778.
Fitts RH (1994). Cellular Mechanisms of Muscle Fatigue. Physiological Reviews 74, 49-94.
Fuglsang-Frederiksen A (1981). Electrical activity and force during voluntary contraction of normal and
diseased muscle. Acta Neurologica Scandinavica 63 (suppl. 83), 1-60.
Gandevia SC & McKenzie DK (1985). Activation of the human diaphragm during maximal static efforts.
Journal of Physiology (London) 367, 45-56.
Gandevia SC & McKenzie DK (1988a). Activation of human muscles at short muscle lengths during maximal
static efforts. Journal of Physiology (London) 407, 599-613.
Gandevia SC & McKenzie DK (1988b). Human diaphragmatic endurance during different maximal respiratory
efforts. Journal ofPhysiology (London) 395, 625-638.
Gandevia SC, Macefield G, Burke D & McKenzie DK (1990). Voluntary activation of human motor axons in
the absence of muscle afferent feedback. The control of the deafferented hand. Brain 113, 1563-1581.
24 C. K. Thomas et al.

Gilson AS Jr & Mills WB (1941). Activities of single motor units in man during slight voluntary efforts.
American Journal of Physiology 133, 658-669.
Grimby L & Hannerz J (1977). Firing rate and recruitment order of toe extensor motor units in different modes
of voluntary contraction. Journal of Physiology (London) 264 , 865-879.
Hagbarth K, Kunesch EJ, Nordin M, Schmidt R & Wallin EU (1986). Gamma loop contributing to maximal
voluntary contractions in man. Journal ofPhysiology (London) 380, 575-59l.
Hales JP & Gandevia SC (1988). Assessment of maximal voluntary contraction with twitch interpolation: an
instrument to measure twitch responses. Journal of Neuroscience Methods 25, 97-102.
Hayward L, Breitbach D & Rymer WZ (1988). Increased inhibitory effects on close synergists during muscle
fatigue in the decerebrate cat. Brain Research 440,199-203.
Heckman CJ & Binder MC (1993). Computer simulations of motoneuron firing rate modulation. Journal of
Neurophysiology 69,1005-1008.
Hicks A, Fenton J, Garner S & McComas AJ (1989). M wave potentiation during and after muscle activity.
Journal ofApplied Physiology 66, 2606-2610.
Hill AV (1913). The heat production in prolonged contractions of an isolated frog's muscle. Journal of
Physiology (London) 47, 305-324.
Hill AV (1938). The heat of shortening and the dynamic constants of muscle. Proceedings ofthe Royal Society
Series B 126,136-195.
Howell IN, Fuglevand AJ, Walsh ML & Bigland-Ritchie B (1994). Motor unit firing during shortening and
lengthening contractions. Society for Neuroscience Abstract 20, 1759.
Jones DA, Bigland-Ritchie B & Edwards RHT (1979). Excitation frequency and muscle fatigue: Mechanical
responses during voluntary and stimulated contractions. Experimental Neurology 64, 401-413.
Katz B (1939). The relation between force and speed in muscular contraction. Journal ofPhysiology (London)
96,45-64.
Kernell D, Donselaar Y & Eerbeek 0 (1987). Effects of physiological amounts of high- and low-rate chronic
stimulation on fast-twitch muscle of the cat hindlimb. II. Endurance-related properties. Journal of
Neurophysiology 58, 614-627.
Kernell D & Monster AW (1982). Motoneurone properties and motor fatigue. An intracellular study of
gastrocnemius motoneurones of the cat. Experimental Brain Research 46, 197- 204.
Krnjevic K & Miledi R (1958). Failure of neuromuscular propagation in rats. Journal of Physiology (London)
140, 440-46l.
Kukulka CG & Clamann HP (1981). Comparison of the recruitment and discharge properties of motor units
in human brachial biceps and adductor pollicis during isometric contractions. Brain Research 219,
45-55.
Lindsley DB (1935). Electrical activity of human motor units during voluntary contraction. American Journal
of Physiology 114, 90-99.
Lippold OCJ (1952). The relation between integrated action potentials in a human muscle and its isometric
tension. Journal ofPhysiology (London) 117, 492-499.
Lippold OCI (1973). The Origin of the Alpha Rhythm. London: Churchill Livingston.
Lippold OCJ, Redfearn JWT & Vuo J (1960). The electromyography offatigue. Ergonomics 3, 121-13!.
Marsden CD, Meadows IC & Merton PA (1969). Muscular wisdom. Journal ofPhysiology (London) 200, 15P.
Marsden CD, Meadows IC & Merton PA (1971). Isolated single motor units in human muscle and their rate
of discharge during maximal voluntary effort. Journal of Physiology (London) 217, 12-13P.
Marsden CD, Meadows JC & Merton PA (1983). "Muscular wisdom" that minimizes fatigue during prolonged
effort in man: Peak rates of motoneuron discharge and slowing of discharge during fatigue. In:
Desmedt IE (ed.), Motor Control Mechanisms in Health and Disease, pp. 169-211. New York: Raven
Press.
McComas AJ, McFadden L, Newberry R & Sacco P (1994). Abstracts, International Symposium on Neural
and Neuromuscular Aspects of Muscle Fatigue (Miami, FL, November 10-13,1994, pp. 35. Miami,
FL: Miami Project to Cure Paralysis, University of Miami.
McKenzie DK, Bigland-Ritchie B, Gorman RB & Gandevia SC (1992). Central and peripheral fatigue of
human diaphragm and limb muscles assessed by twitch interpolation. Journal ofPhysiology (London)
454, 643-656.
Merton PA (1954). Voluntary strength and fatigue. Journal of PhYSiology (London) 123,553-564.
Merton PA, Hill DK & Morton HB (1981). Indirect and direct stimulation offatigued human muscle. In: Porter
R & Whelan J (eds.), Human Muscle Fatigue: Physiological Mechanisms. London: Pitman Medical.
Milner-Brown HS & Miller RG (1986). Muscle membrane excitation and impulse propagation velocity are
reduced during muscle fatigue. Muscle & Nerve 9, 367-374.
The Scientific Contributions of Brenda Bigland-Ritchie 25

Milner-Brown HS & Stein RB (1975). The relation between the surface electromyogram and muscular force.
Journal of Physiology (London) 246, 549-569.
Milner-Brown HS, Stein RB & Yemm R (1973). The orderly recruitment of human motor units during
voluntary isometric contractions. Journal of Physiology (London) 230, 359-370.
Monster AW & Chan H (1977). Isometric force production by motor units of extensor digitorum communis
muscle in man. Journal ofNeurophysiology 40, 1432-1443.
Mosso A (1915). Fatigue, 3rd Ed. Translated by Drummond M & Drummond WD. London: Allen and Unwin.
Nardone A, Romano C & Schieppati M (1989). Selective recruitment of high-threshold human motor units
during voluntary isotonic lengthening of active muscles. Journal of PhYSiology (London) 409,
451-471.
Ritchie IM & Ritchie BR (1968). Local anesthetics: Effect of pH on activity. Science 162, 1394-1395.
Ritchie IM, Ritchie B & Greengard P (1965a). The active structure oflocal anesthetics. Journal of Pharma-
cology & Experimental Therapeutics 150,152-159.
Ritchie IM, Ritchie B & Greengard P (1965b). The effect of the nerve sheath on the action oflocal anesthetics.
Journal of Pharmacology & Experimental Therapeutics ISO, 160-164.
Smith OC (1934). Action potentials from single motor units in voluntary contraction. American Journal of
Physiology 108, 629-638.
Stephens JA & Taylor A (1972). Fatigue of maintained voluntary muscle contractions in man. Journal of
Physiology (London) 220,1-18.
Tanji J & Kato M (1973a). Recruitment of motor units in voluntary contraction of a finger muscle in man.
Experimental Neurology 40,759-770.
Tanji J & Kato M (1973b). Firing rate of individual motor units in voluntary contraction of abductor digiti
minimi muscle in man. Experimental Neurology 40,771-783.
Thomas CK (1993). Muscle fatigue after incomplete human cervical spinal cord injury. Journal of the
American Paraplegia Society 16, 87.
Thomas CK, Bigland-Ritchie B & Johansson RS (199Ia). Force-frequency relationships of human thenar
motor units. Journal ofNeurophysiology 65, 1509-1516.
Thomas CK, Johansson RS & Bigland-Ritchie B (199Ib). Attempts to physiologically classify human thenar
motor units. Journal ofNeurophysiology 65,1501-1508.
Thomas CK, Woods JJ & Bigland-Ritchie B (1989). Impulse propagation and muscle activation in long
maximal voluntary contractions. Journal ofApplied Physiology 67, 1835-1842.
Vallbo AB, Hagbarth KE, Torebjork HE & Wallin BG (1979). Somatosensory, proprioceptive, and sympathetic
activity in human peripheral nerves. Physiological Reviews 59, 919-956.
VlIJllestad NK, Sejersted OM, Bahr R, Woods JJ & Bigland-Ritchie B (1988). Motor drive and metabolic
responses during repeated submaximal voluntary contractions in humans. Journal ofApplied Physi-
ology 64,1421-1427.
Westling G, Johansson RS, Thomas CK & Bigland-Ritchie B (1990). Measurement of contractile and electrical
properties of single human thenar motor units in response to intraneural motor-axon stimulation.
Journal of Neurophysiology 64,1331-1338.
Wilkie DR (1968). Heat work and phosphory\creatine break-down in muscle. Journal ofPhysiology (London)
195, 157-183.
Woods JJ & Bigland-Ritchie B (1983). Linear and non-linear surface EMG/force relationships in human
muscles: An anatomical/functional argument for the existence of both. American Journal ofPhysical
Medicine 62, 287-299.
Woods JJ, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal ofNeurophysiology 58, 125-137.
SECTION I
Fatigue of Single Muscle Fibers

While the bulk of muscle fatigue can be attributed to processes beyond the neuromus-
cular junction, it is the central nervous system (CNS) that actually decides when the process
of muscle fatigue will occur. Chapters in this section examine the crucial processes that take
place beyond the neuromuscular junction. A common definition of fatigue is used in these
chapters, namely a contraction-induced reduction in the ability of the muscle to produce
force and/or to shorten. The experimental work relies primarily on the use of single
muscle-fiber preparations to study the key events that impair force generation.
Chapter 1 (Edman) assesses the extent to which force is impaired due to events that
occur at the crossbridges, or to more proximal events within the cell which impair crossbridge
activation. It is argued that calcium-induced activation is less likely to be impaired than
events at the myofibrils, at least during moderate fatigue. The relative importance of these
events depends on the exact protocol and stimulus regimen that is used to induce fatigue.
In Chapter 2 (Stephenson and colleagues), the processes that transpose the sarcolem-
mal action potential to the actomyosin motor are investigated. Many of the studies here use
a preparation in which a single muscle fiber is skinned so that the intracellular environment
can be manipulated to produce effects on force. This permits many crucial steps in excita-
tion-contraction coupling to be examined. As the processes involved in excitation-contrac-
tion coupling are not all-or-none unitary events, emphasis must be placed on the molecular
events involved in voltage sensing and initiation of signal transmission from the t-tubular
system to the sarcoplasmic reticulum (via dihydropyridine receptors) and on the variety of
factors that modify calcium release from the sarcoplasmic reticulum. An elevation in
intracellular Mg2+, which occurs in fatigue, may significantly impair calcium release and
ultimately crossbridge activation.
Many examples of muscle fatigue are associated with intracellular acidosis. In
Chapter 3 (Allen and colleagues), the processes impaired by a fatigue-induced rise in
hydrogen ion concentration are considered. While many features of fatigue can be mimicked
by intracellular acidosis (such as an impaired maximal shortening velocity and slowed rate
of relaxation), it is not the only factor that can produce them. This view is supported by
observations on patients with myophosphorylase deficiency. They do not generate lactate
during exercise, yet they fatigue faster than normal subjects. This chapter also reiterates the
biochemical reality that hydrogen ions generated by exercise will affect many critical steps
in force generation, including the turnover of crossbridges and the force produced by each
attached crossbridge.
Ion exchange mechanisms have long been known to change in exercising muscle. In
Chapter 4 (Sj0gaard & McComas), the local and distant effects produced by the initial rise
28 Fatigue of Single Muscle Fibers

in interstitial potassium are reviewed. Local increases in extracellular potassium will be most
marked in isometric exercise. While they can impair sarcolemmal and t-tubular function,
regulatory mechanisms exist that range from local activation of the Na+/K+ pump in both
active and adjacent quiescent fibers to the reflex adjustment of blood flow and ventilation.
This discussion introduces the theme that specific mechanisms exist to inform the eNS about
the progress of muscle fatigue.
The processes reviewed in this section define those crucial biochemical events that
must underlie fatigue of single fibers acting not only in vitro but also in vivo. However,
particular challenges, taken up in later sections, are to assess the relative importance of these
events and to decide which are more important when force is produced by whole motor units
in vivo.
 1

MYOFIBRILLAR FATIGUE VERSUS FAILURE

OF ACTIVATION

K. A. P. Edman

Department of Phannacology
University of Lund
S61vegatan 10
S-223 62 Lund
Sweden

ABSTRACT

Two principal mechanisms underlying fatigue of isolated muscle fibers are described: failure
of activation of the contractile system and reduced performance of the myofibrils due to
altered kinetics of crossbridge function. The relative importance of these two mechanisms
during development of fatigue is discussed.

It is common knowledge in muscle physiology that the mechanical performance of

an isolated muscle preparation is greatly dependent on the time that the muscle is allowed
to rest between successive contractions. For example, a single twitch interposed in a series
of isometric tetani can affect both the time course and the amplitude of one or more of the
subsequent tetani. This indicates that the extra contraction causes some disturbance within
the contractile machinery or in the excitation-contraction coupling, which takes the muscle
a relatively long time to overcome. Thus, the mechanical performance of striated muscle is
quite dependent on the stimulation history of the muscle fibers, and this likely holds true for
muscles operating in situ in the body as well as muscle preparations studied in an organ bath.
Progressive shortening ofthe resting intervals between contractions may greatly reduce the
force generating capability of the muscle. This reversible state offorce depression, referred
to as fatigue, also includes a lower rate of rise of force and a slower relaxation (e.g., Edwards
et aI., 1975; Fitts & Holloszy, 1978; Edman & Mattiazzi, 1981). Two main causes are thought
to underlie fatigue in an isolated muscle preparation: 1) failure of activation of the contractile
system; and 2) reduced capacity of the myofibrils to produce force at any given state of
activation (myofibrillar fatigue). The relative importance of these two factors during devel-
opment of fatigue in isolated muscle will be discussed below. Myofibrillar fatigue will be
discussed first.

29
30 K. A. P. Edman

MYOFIBRILLAR FATIGUE

An obvious difficulty in simulating human or animal fatigue in an isolated muscle

preparation is that the motor control mechanisms normally operating in vivo are lacking in
the organ bath. Presently, there is no clear information on the extent a muscle fiber can be
stimulated by its motor nerve in situ in the body. Needless to say, an isolated muscle fiber
can easily be stimulated until some aspect of the excitation-contraction mechanism fails and
irreversible changes appear, but there is still no evidence that such a situation will ever appear
in vivo. As will be discussed, it is feasible to produce a moderate, and fully reversible state
of fatigue in an isolated muscle fiber by using a stimulation routine that does not lead to
failure of activation of the contractile system. Single fibers isolated from amphibian leg
muscles are most frequently used in these studies. The remarkable viability of isolated
amphibian muscle fibers makes these preparations particularly well suited for the long-last-
ing and demanding experiments involved in the study of muscle fatigue.

Isometric Force
Fig. I illustrates the characteristic changes of the isometric tetanus that are produced
during moderate fatigue in a frog muscle fiber. In the control series (Fig. IA), the fiber is
stimulated to produce a I-s isometric tetanus at regular 5- or IS-min intervals. Fatigue is
produced by reducing the intervals between the tetani to 15 s (Fig. IB). Isometric force, using
this stimulation regimen, declines to a steady level at approximately 75% of the original
value after twenty-five to thirty contractions. The shape of the tetanus myogram is also
changed in a characteristic way during fatigue; there is a lower rate of rise of force during
the onset of the tetanus and a slowing of the relaxation phase (note the prolonged time to the
shoulder of the relaxation phase). It is of interest to note, however, that these changes are all
fully reversed after return to the control stimulation frequency (Fig. 1C).
There is now good evidence that moderate fatigue does not involve failure of
activation ofthe contractile system. In studying this problem, caffeine has served as a useful
tool based on its ability to potentiate the release of activator calcium from the sarcoplasmic
reticulum. Fig. 2 illustrates the effects of 0.5 mM caffeine on the isometric twitch and tetanus
after the tetanic force has been reduced by approximately 25% by frequent tetanization (15
s intervals between tetani). Caffeine can be seen to increase the rate of rise of force in both
twitch and tetanus and to potentiate the isometric twitch. However, caffeine does not affect
the amplitude of the tetanus. These results suggest that even in the fatigued state there is a
capacity for further release of calcium from the sarcoplasmic reticulum. The lack of effect
of caffeine on the tetanus amplitude clearly shows, however, that the contractile system is
maximally activated during a tetanus in moderate fatigue.
Caffeine, added to the external medium in higher than twitch-potentiating concen-
trations, induces graded contracture of intact muscle fibers by releasing calcium from the
sarcoplasmic reticulum.· This effect of caffeine is independent of membrane excitation and
provides a means of activating the contractile system to the desired level by side-stepping
the excitation-contraction coupling (e.g., Delay et aI., 1986; Klein et aI., 1990). Fig. 3 shows
an experiment in which contractures induced by 3 and 15 mM caffeine have been interposed
between I-s tetani under control (rested state) conditions, and after development of moderate
fatigue in a single muscle fiber. As can be seen in the inset diagram of Fig. 4, 3 mM caffeine
induces submaximal contracture whereas 15 mM caffeine is more than sufficient to mechani-

• Contracture is the term used for shortening of muscle that does not involve membrane excitation. It should
not be confused with a permanent shortening of soft tissues.
Myofibrillar Fatigue versus Failure of Activation 31

)
8

)
c

I
H
C\I

Figure 1. Isometric tetani of a single muscle fiber

recorded after attainment of steady state at two contrac-
tion frequencies. Interval between contractions: A, 15
j d

min; B, 15 s; C, return to 15 min intervals. Temperature,

2.2 °C (reproduced with permission from Edman & ~
Mattiazzi, 1981). 400 ms

A 8

'"E
E

l...--....J
J~
500 ms
Figure 2. Effects of caffeine on isometric twitch (left) and tetanus (right) after development of fatigue.
Continuous lines, no caffeine. Dashed lines, presence of 0.5 mM caffeine. Horizontal bar above tetanus
myogram: maximum tetanic force recorded before fatigue (reproduced with permission from Edman & Lou,
1990).
32 K. A. P. Edman

Part 1 Part 2

Control Fatigue Rest

,
Control Fatigue Rest Control
I I I

100 at!' "ll. <II> aIJ

•

'.
0
~
~ co 0 "'"
0
0
80

~ 60
CD
tJ
0
u.. 40

20
Rest Rest
I' , I 2h I , II' , I 2 h f---1--
0 60 120130 140 300 360 420 480505 515 660
Time (min)
Figure 3. Relative changes of tetanic force and contracture responses to caffeine during development of
moderate fatigue in a frog single muscle fiber. 0, isometric tetanus; ... and., contracture responses to 15 and
3 mM caffeine, respectively. The sequence of control, fatigue and rest periods is indicated by marked intervals
above the data points. The time scale is expanded during the fatigue period. Only some of the tetanic responses
are plotted in the diagram for clarity. Tetanic tensions throughout the experiment are normalized to the mean
value of tetanic force derived during the control period of part 1. Contracture tensions being induced by two
different caffeine concentrations are normalized to values derived during the control period in each respective
part of the experiment. Note that tetani and caffeine contractures are depressed to approximately the same
degree (from Edman & Lou, 1992).

cally saturate the contractile system of the intact frog muscle fiber. However, independent
of the caffeine concentration used, the amplitude of the caffeine-induced contracture is
reduced to nearly the same degree as is the isometric tetanus by fatiguing stimulation (Figs.
3 and 4). Similar to the situation during the tetanus, the rate of rise offorce during the caffeine
contracture is also reduced in the fatigued state (Edman & Lou, 1992). These results strongly
suggest that the decrease in mechanical performance during moderate fatigue is not attrib-
utable to failure of activation of the contractile system. The fatigue process apparently makes
the myofibrils less capable of producing force even though they are fully activated by
calcium.

Vmax, Force-Velocity Relation and Power

Fatigue does not merely involve the fiber's capacity to produce force; the maximum
speed of shortening (Vo) and the power output are affected as well. Vo is preferably measured
by the slack-test (Edman, 1979) in an isolated muscle preparation as this ensures that
shortening occurs at virtually zero load and, furthermore, that the measured velocity is not
affected by recoil of elastic components acting in series with the contractile units. The first
detailed study ofthe relative changes ofVo and isometric force during development of fatigue
was performed by Edman and Mattiazzi (1981). In a more recent publication (Curtin &
Edman, 1994) the effects of moderate fatigue on the entire force-velocity relationship,
including measurements at loads exceeding maximum tetanic force (Po), have been de-
scribed. Vo is not significantly affected during the initial stage of muscle fatigue but starts
Myofibrillar Fatigue versus Failure of Activation 33

~ 100

r-· _.
Q)

u
~ 90
c
0
u
Q)
c 100
80 • 0
~ctI .. 0-
~
'~"
U 50
Cl
c
.;:: 70 0 0
u..
:l
"1j 0
Q) 0 5 10 15
u
(; 60 Caffeine (mM)
u..
60 70 80 90 100
Tetanic force (%)

Figure 4. Relation between maximal tetanic force and maximal contracture response to caffeine during
development of moderate fatigue. Data expressed as percentage of tetanus and contracture tensions recorded
under control (prefatigue) conditions in respective fibers. The control value (indicated by large open circle)
is the mean of two to four separate recordings in each fiber. Caffeine concentrations (mM): 0, 3; 0, 6; L'l., 9;
.... , 12; ., 15. Continuous line, linear regression based on all data points. Inset: relation between peak
contracture tension and caffeine concentration; each data point is the mean of two to nine contractures
performed in nine fibers (reproduced with permission from Edman & Lou, 1992).

to decline at a point where the tetanic force has been reduced by approximately 10% of its
rested-state value (Fig. 5). As the fatiguing process goes on, however, Vo becomes increas-
ingly affected and finally reaches approximately the same degree of depression as Po does
during moderate fatigue.
The fact that V0 is reduced during fatigue provides a relevant piece of information
to the evaluation of the cellular mechanisms involved in muscle fatigue. The maximal speed

35

30

-
~
c
o
()
25
cfi
Figure 5. Decrease in velocity of~o
unloaded shortening (L'[Link]) in rela- >
tion to force depression (L'l.P o) dur- '0 20
ing development of moderate c
fatigue in single muscle fibers. g
Each set of data connected by a g 15
solid line is from a single fiber. ~
Dashed line drawn from the equa- a:
tion L'[Link] = 0.006 L'l.P 02.48 - 1.0 de- 10
rived by regression analysis from
all data points. Co-ordinates ex-
pressed as percent of Vo and te- 5
tanic force recorded under control
(pre fatigue ) conditions in respec-
tive fibers (reproduced with per-
mission from Edman & Mattiazzi, o 5 10 15 20 25 30 35
1981). Reduction of tetanic force (% control)
34 K. A. P. Edman

A
2·0
0
0
0

1·5 A

R
A
lUI 0
....J 1·0 A
~ 0
>. A
0
~
0 0·5 A 0
A
~ AA
0
00 0
A 0
A
0·0 .!hA0 0 0

-0·5
0·0 0·1 0·2 0·3 0·4
Force (N mm- 2 )

8 0·00

,,
,
'*
I
I
Figure 6. A, force-velocity relation
of frog muscle fiber at rested state and
A
, moderate fatigue. FL is the fiber
%l.
length at resting sarcomere length of
2.1 [Link]. Positive velocities for shor-
tening, negative velocities for length-
ening. Control (0), fatigue (~); B,
normalized fonn of the force-velocity
relation for lengthening. Force ex-
pressed as a fraction of the isometric
-0.12 L - _ - . . I ._ _.....L.._ _- ' -_ _.L.-_--' force measured in control and fatigue,
2·0 respectively (from Curtin & Edman,
1·0 1·2 1·4 1·6 1·8
1994).
Force (PIPo)

of shortening in vertebrate skeletal muscle is, unlike the isometric force, independent of the
degree of activation of the contractile system. This has been demonstrated in both frog intact
muscle fibers (Edman, 1979) and skinned fiber preparations of amphibian and mammalian
muscles (Podolsky & Teichholz, 1970; Brenner, 1980; Ferenczi et aI., 1984. Vo is further-
more insensitive to the amount of overlap between the thick and thin filaments (Edman,
1979) supporting the view (Huxley, 1957) that the speed of shortening at zero load is
independent of the actual number of myosin crossbridges interacting with the thin filaments.
The finding that Vo does change during the course of fatigue thus strongly suggests that
fatigue involves a specific change of the kinetic properties of the myosin crossbridge that is
not attributable to altered metabolism of calcium in the excitation-contraction coupling.
Vo may be presumed to reflect the maximal cycling rate of the crossbridges, and this
property is now generally thought to be determined by the rate constant of detachment of
Myofibrillar Fatigue versus Failure of Activation 35

A
0·08

7CI)
0-06

'"IE
E ,,
ii 0·04 l!.
z ~
....
~
a. 0·02
l!.
\
\

\
l!. \
0
0
l!. 0
0·00
0·0 0·1 0·2 0·3
Force (N mm- 2 )

B
0·20

.-.. 0·15
/',.
~ .... --,
~E tr' l;,

Figure 7. Power-force relations for *rt: /

~,
shortening in the control state and in
/

0
,
the moderately fatigued state. A. ~ 0·10
power calculated as the product of
e:,.
force and velocity expressed in N
mm· 2 and fiber lengths (FL) s·l, re-
spectively. B, power calculated after
normalizing force (P) and vlocity (V)
0·05
data with respect to maximum force
(Po *) and maximum speed of shorten-
ing (Vmax), respectively. Control
(0), fatigue (A) (from Curtin & Ed- 0·00
man, 1994). 0·0 0·2 0·4 0·6 0·8 1·0
Force (PIPo*)
the crossbridges at negative strain, that is when the crossbridges are in a position where they
brake the sliding motion of the myofilaments (Huxley, 1957).
Another expression of the altered cross bridge function during muscle fatigue is the
change in shape of the force-velocity relation during fatiguing stimulation. The force-veloc-
ity relation for shortening has been shown to consist of two distinct curvatures, both having
an upwards concave shape, located on either side of a break-point near 78% of Po (Edman,
1988). Moderate fatigue reduces the main curvature of the force-velocity relation but does
not affect the biphasic nature of the force-velocity curve or shift the location of the breakpoint
relative to Po (Fig. 6). As is evident from Fig. 7A, there is a true reduction of the power output
36 K. A. P. Edman

during fatigue at all force levels. It is essential to point out, however, that since the
force-velocity relation becomes straighter during fatigue, the fiber is able to maintain a
higher power output than would otherwise be the .case.
This is demonstrated in Fig. 7B where the mechanical power at rested state and during
moderate fatigue has been normalized with respect to isometric force and maximum speed
of shortening. Note that the relative power points for fatigue lie above the controls and that
maximal power is attained at a higher relative force in the fatigued state. The decrease in
curvature of the force-velocity relation can be presumed to make the muscle less efficient
(Woledge, 1968). The muscle may thus have to spend somewhat more energy for a certain
amount of work. This may be the price the muscle has to pay to better maintain its power
output during fatigue.
The force-velocity relation at loads exceeding Po expresses the muscle's ability to
resist stretching and forms a smooth continuation of the force-velocity relation during
shortening (Edman, 1988). The flat force-velocity relationship around Po confers a high
degree of stability upon the contractile system. Any tendency of a fiber segment to be
stretched during contractile activity (due to pull by stronger segments in series) is thus
counteracted by the build-up of a resistive force in the weaker (yielding) part. The recruit-
ment of extra force during stretching is only to a minor part attributable to an increased
number of attached crossbridges judging from the modest increase in fiber stiffness (Lom-
bardi & Piazzesi, 1990; K.A.P. Edman, unpublished observations). Apparently the cross-
bridges are able to hold a greater force when the muscle is stretched and the bridges are
dragged along the thin filaments against their normal course than when the bridges operate
under isometric conditions. As is evident from the following, fatigue does not reduce the
bridges' ability to resist stretch.
Fig. 6A shows the effects offatigue on the force-velocity relation during stretching,
that is for loads higher than Po, along with results for shortening. Force is here expressed in
absolute values. While fatigue reduces Po and lowers the speed of shortening at each load
less than the isometric force, the force-velocity relations for stretching nearly overlap in the
control and fatigued state. Thus, for a given speed of elongation, the fiber produces a
proportionately higher force after fatigue than in the control state. This is further illustrated
in Fig. 6B in which the force-velocity relation for lengthening is shown with force normalized
to the corresponding isometric value, Po, in control and fatigue. This change in contractile
behavior during fatigue is likely to be of significance under in vivo conditions as a muscle,
weakened by fatiguing exercise, may yet be able to resist stretch nearly as well as in the
non-fatigued state.

Fiber Stiffness
The instantaneous stiffness of active muscle provides information on the number of
myosin crossbridges that are interacting with the thin filaments (Huxley & Simmons, 1971).
For this measurement a fast,low-amplitude length perturbation is applied to the muscle either
in the form of a single, very fast length step (Huxley & Simmons, 1971) or a high-frequency
sinusoidal length oscillation (e.g., Julian & Morgan, 1981; Cecchi et aI., 1986; Edman &
Lou, 1990). Simultaneous measurements of force and instantaneous stiffness have recently
been carried out during development of fatigue in frog single muscle fibers (Edman & Lou,
1990, 1992; Curtin & Edman, 1994). Results from such measurements are illustrated in Fig.
8 for fibers subjected to moderate fatigue by repetitive tetanization at 15-s intervals. It is
clear that stiffness is much less depressed than force is during fatiguing stimulation. This is
readily seen from the inset myograms of Fig. 8 and from the pooled data in the main diagram,
which shows maximum tetanic force plotted against maximum tetanic stiffness during
development of fatigue in six different fibers. A regression analysis based on these results
Myofibrillar Fatigue versus Failure of Activation 37

100

90

80

70 80 90 100
Tetanic force (%)
Figure 8. Relation between maximum tetanic stiffness and maximum tetanic force during development of
moderate fatigue in six single muscle fibers of the frog. Data normalized with respect to maximum force and
maximum stiffness recorded under control conditions in respective fibers. The control value (indicated by
large open circle) is the mean of five separate recordings in each fiber. Data from a given fiber are denoted by
the same symbol. Line, least squares regression of stiffness upon force. Insets: superimposed records of tetanic
force (A) and tetanic stiffness (B) under control conditions (traces 1) and after fatiguing stimulation (traces 2)
(from Edman & Lou, 1990).

shows that for a 25% depression of the maximum tetanic force during fatigue there is merely
9% reduction in fiber stiffness. Another discrepancy between force and stiffness concerns
the rising phase of contraction. As pointed out before, the initial rate of rise of force during
tetanus is reduced during moderate fatigue. By contrast, the rate of rise of stiffness was found
to remain virtually unchanged (Edman & Lou, 1990).
These observations would seem to make clear that the decrease in force during
moderate fatigue is only partly due to fewer crossbridges being attached to the thin
filaments. Fatigue evidently reduces the force output of the individual crossbridges and
this accounts for the major part of the force decline during moderate fatigue. As already
discussed, fatigue also reduces Vo, the maximum speed of shortening, which is yet another
expression of the modified performance of the crossbridge during development of fatigue.
It should be noted that these changes occur while the contractile system is fully activated
(see earlier).
The above changes in crossbridge performance during fatigue, including the slight
decrease in number of attached crossbridges, therefore seem to be unrelated to the myoplas-
mic free calcium concentration. There is reason to believe, on the other hand, that the altered
kinetics of crossbridge function is attributable to the intracellular metabolic changes that are
known to arise during muscle fatigue.

Metabolic Changes during Fatigue and Their Influence on Muscle

Function
Fatiguing exercise leads to, among other things, reduction of ATP and accumulation
of products of the ATP hydrolysis, i.e., ADP, orthophosphate (Pi) and H+ (Spande &
Schottelius, 1970; Edwards et aI., 1975; Dawson et aI., 1978, 1980; Nassar-Gentina et aI.,
1978; Kawano et aI., 1988). The change in MgATP concentration is quite small during
38 K. A. P. Edman

moderate fatigue (Dawson et aI., 1978, 1980) and is unlikely to significantly affect cross-
bridge function (Ferenczi et aI., 1984). However, the ATP hydrolysis products, in the
concentrations reached during fatigue, are all known to affect the kinetic properties of the
contractile system (see Stephenson et aI., Chapter 2).
Results from skinned fiber experiments show that increased Pi concentration reduces
isometric force (Altringham & Johnston, 1985; Chase & Kushmerick, 1988; Cooke et aI.,
1988; Godt & Nosek, 1989; Stienen et aI., 1990, 1992; Dantzig et aI., 1992. It is believed
the power stroke of the crossbridge is initiated as Pi dissociates from the myosin head. An
increased Pi concentration would thus be expected to increase the fraction of crossbridges
that are in a low-force producing state. The influence of Pi on maximum speed of shortening,
however, is less well understood. There is agreement that Pi does not affect Vo at a pH of
7.0-7.4 (Altringham & Johnston, 1985; Chase & Kushmerick, 1988; Cooke et aI., 1988). At
pH 6.00, on the other hand, Pi was found to have no effect in one study (Cooke et aI., 1988)
but to have a marked depressant effect in another (Chase & Kushmerick, 1988). Mg-ADP
has been stated to slightly increase the isometric force of skinned fibers, in the presence of
millimolar concentrations ofMg-ATP, and to exert a small inhibitory effect on Vo (Cooke et
aI., 1988). Finally, lowered pH has been convincingly shown to reduce both Vo and isometric
force in skinned muscle fibers (Chase & Kushmerick, 1988; Cooke et aI., 1988; Godt &
Nosek, 1989). A valuable study that has yet to be performed on the skinned fiber preparation
would be to explore the combined action of the hydrolysis products on the force-velocity
relation using concentrations corresponding to those recorded at various degrees of fatigue.
The relative changes in Vo and isometric force that are known to occur during fatigue in the
intact fiber (see further Edman & Mattiazzi, 1981) would serve as a useful guide in such a
study.
The effects oflowered intracellular pH (PHi) are testable in intact single muscle fibers
by varying the concentration of the permeant acid CO2 in the bathing fluid. Based on this
approach the contractile effects of intracellular acidification have been compared with the
mechanical changes that are induced by fatiguing stimulation in frog muscle fibers (Edman
& Mattiazzi, 1981; Edman & Lou, 1990, 1992; Curtin & Edman, 1989, 1994). Overall there
is a striking similarity between the changes induced by the two interventions. Lowering of
the intracellular pH reduces both Vo and maximal tetanic force (Edman & Mattiazzi, 1981).
Intracellular acidification also induces the characteristic changes of the isometric myogram
that are seen in fatigue, that is reduced rate of rise of force during the onset of tetanus and
slowing of the linear phase of force relaxation (Edman & Mattiazzi, 1981; Edman & Lou,
1990). Similar to the situation in moderate fatigue, the decrease in active force during
intracellular acidification is associated with a relatively small reduction in fiber stiffness
suggesting that lowered pH does reduce the force output of the individual bridge (Edman &
Lou, 1990). Furthermore, reduced pHi, like fatigue, increases the fiber's ability to resist
stretch presumably by making the myosin crossbridges adhere more firmly to the actin
filaments during activity (Curtin & Edman, 1989, 1994).
Although the changes induced by fatigue are, in large measure, reproducible by
intracellular acidification, there is evidence that lowered pHi does not fully account for the
contractile change during moderate fatigue (see Allen et aI., Chapter 3). This is evident when
the relative changes in force and stiffness are considered for the two interventions. For
example, a given decline in tension during intracellular acidification is associated with an
even smaller drop in stiffness than during fatigue (Edman & Lou, 1990). Other examples of
such differences between the two interventions are given by Edman and Lou (1990) and
Curtin and Edman (1994). One or more of the other metabolic factors discussed above,
notably Pi, are therefore also likely to playa part, in addition to lowered pHi, during muscle
fatigue.
Myofibrillar Fatigue versus Failure of Activation 39

Myoplasmic Free Calcium Concentration

Measurements of the intracellular free calcium concentration, [Ca2+]io have been
performed during fatiguing stimulation using various calcium indicators such as aequorin,
Fura-2, indo-l (Allen et aI., 1989; Westerblad & Allen, 1991, 1993) and Fluo-3 (Edman et
aI., unpublished data). The results of these studies show that the tetanic [Ca2+]j undergoes
only slight changes during moderate fatigue, in line with the above conclusion that the
contractile system remains maximally activated under these conditions. In a recent study,
using Fluo-3 and accounting for the on- and off-rate constants for the calcium-dye complex
(see further Caputo et aI., 1994), the tetanic [Ca2+]j during moderate fatigue did not drop
below the level required for maximum tetanic force at rest (Edman et aI., unpublished
results). Westerblad and Allen (1991), using Fura-2 in a study of mouse muscle fibers,
observed no decline in tetanic free calcium concentration as the force was reduced to
approximately 80% of the rested-state value by fatiguing stimulation. At lower degrees of
fatigue, tetanic [Ca2+]j was actually found to rise above the control level (Lee et aI., 1991;
Westerblad & Allen, 1991). The situation is quite different, however, when the fiber is
brought into a more profound state offatigue by frequent stimulation. Under such conditions,
as described in the following section, there is a marked reduction of the [Ca2+]j transient
during tetanus (Lee et aI., 1991).

FAILURE OF ACTIVATION

Most previous studies of fatigue in isolated muscle preparations have employed a

stimulation program which produces a more profound depression of force than that described
in the previous section. The fatiguing protocols employed left very little time, generally no
more than 1-3 s, for recovery of the muscle between the stimuli. As a result of such an intense
treatment, the force produced during a I-s tetanus is reduced within 15 min to approximately
40% of the rested-state value; over the same period of time the isometric twitch response
declines to merely 5-10% of the control value (Eberstein & Sandow, 1963; Edman & Lou,
1992). This state of force depression will be referred to in the following as excessive fatigue
(Edman & Lou, 1992) in order to distinguish it from the moderate, myofibrillar form of
fatigue (approximately 70% of rested-state force), discussed in the preceding section.
The first convincing evidence that excessive fatigue involves failure of activation of
the contractile system was presented by Eberstein and Sandow (1963). They demonstrated
that a single muscle fiber, nearly exhausted by frequent stimulation and whose twitch
response was virtually abolished, was nevertheless capable of producing an almost maximal
contracture response to caffeine. This finding has subsequently been confirmed in several
studies on amphibian and mammalian muscles (Kanaya et aI., 1983; Jones & Sacco, 1989;
Liinnergren & Westerblad, 1989; Edman & Lou, 1992). High-frequency stimulation appar-
ently leads to some failure of the excitation-contraction coupling that is overcome when the
fiber is activated by caffeine, which releases calcium from its storage site in the sarcoplasmic
reticulum (see further Delay et aI., 1986; Klein et aI., 1990). Another conclusion that may
be drawn from this finding is that the intracellular calcium store is not depleted during
high-frequency stimulation. In fact, the above results would seem to make clear that there
is enough calcium available in the store during excessive fatigue to fully activate the
contractile system.
There is now evidence that fast repetitive stimulation leads to failure of the inward
spread of activation. This was first reported by Gonzalez-Serratos and colleagues (1981) and
later, in a more detailed form, by Garcia and colleagues (1991). These authors observed that
myofibrils in the center of the fiber were unable to actively shorten below slack length during
40 K. A. P. Edman

tetanus in the fatigued state indicating that they were not properly activated by the electrical
stimulus.
Using an experimental approach similar to that employed by Garcia and colleagues
(1991), combined with rapid freeze-fixation, Edman and Lou (1992) did not find any sign
of defective activation of the central myofibrils during moderate fatigue, that is under
conditions when the tetanic force was depressed by less than 30% of the control value.
However, as the tetanic force was depressed further, central myofibrils exhibited a wavy
appearance when the fiber shortened below slack length at zero, indicating inactivity. In fact,
at an advanced state of fatigue (tetanic force reduced below 50% of the control) only a
relatively thin layer ofmyofibrils in the periphery of the fiber remained active while the rest
of the fiber showed clear signs of being inactive or submaximally activated (Edman & Lou,
1992; their Fig. 8).
The mechanism underlying the failure of inward spread of activation during exces-
sive fatigue is not yet fully understood. There is good reason to presume, however, that
repetitive electrical stimulation will cause accumulation of potassium in the T-tubules (Hnik
et aI., 1976; Juel, 1986) when too short a time is available to restore the potassium
concentration between the stimulation volleys. The increased potassium concentration will
gradually depolarize the tubular membrane and this will impair the inward conduction of the
electrical impulse. At high stimulation frequencies only the outer layers of the fiber may be
accessible for activation by the action potential while the interior of the fiber stays totally
inactive. Although this mechanism would seem to adequately explain the experimental
results, the possibility cannot yet be excluded that fatiguing stimulation leads to reduced
calcium sensitivity and that this effect somehow becomes more pronounced towards the
center of the fiber (Allen et aI., 1993). It is known that Pj and pHj both reduce the myofibrillar
calcium sensitivity (Fabiato & Fabiato, 1978; Kentish, 1986; Godt & Nosek, 1989; Metzger
& Moss, 1990; Lynch & Williams, 1994), but it would require that a concentration gradient
for these two products develops across the fiber during fatiguing stimulation. At present there
is no experimental evidence to suggest that such a gradient arises.
As described earlier, the instantaneous stiffness of the muscle fiber is only slightly
reduced during moderate fatigue when the decrease in mechanical performance is mainly
attributable to altered kinetics of the myosin crossbridges (see earlier). However, there is a
drastic change of the force-stiffness relationship at a point where the inward spread of
activation begins to fail during fatiguing stimulation. Thus, as the tetanic force is reduced
below approximately 70% of the rested-state value (interval between contractions 1-2 s),
any further decrease in force is associated with a nearly proportional reduction in fiber
stiffness (Edman & Lou, 1992). The progressive decrease in force at this stage of fatiguing
stimulation is therefore largely due to fewer bridge attachments being formed. This supports
the conclusion (see above) that failure of activation finally becomes the main cause of the
force decline in a fiber subjected to frequent stimulation.
Further evidence in support of this view comes from intracellular calcium measure-
ments. Thus, as observed in both frog and mammalian muscle fibers (Allen et aI., 1989; Lee
et aI., 1991; Westerblad & Allen, 1991), the tetanic [Ca2+]j is progressively reduced as the
isometric force is depressed below the level (near 70% of the control value) at which stiffness
starts to decline steeply. In one study, Allen and coworkers have indeed been able to
demonstrate a transient of tetanic [Ca 2+]j across the fiber during intense fatiguing stimula-
tion. Using a stimulation program in which relatively long tetani were produced at short
intervals, these authors recorded a lower concentration of Ca2+j in the center than in the
periphery of the fiber (Westerblad et aI., 1990). This observation accords with the idea that
frequent stimulation eventually leads to failure of the inward spread of activation. However,
in a later study the same laboratory, using a somewhat different stimulation routine, found
no indication of a [Ca2 +]j gradient across the fiber during excessive fatigue (Westerblad et
Myofibrillar Fatigue versus Failure of Activation 41

aI., 1993). More experimental evidence is apparently needed to settle the problem of whether
failure of inward activation during fatiguing stimulation is associated with an uneven
distribution of [Ca2+]j within the fiber.

SUMMARY
While myofibrillar fatigue is likely to playa relevant part during muscular exercise
in vivo (see earlier), the physiological significance offailure of activation is less clear. For
activation failure to arise the muscle fiber has to undergo an intense stimulation program
that may not be permitted in a muscle operating in situ in the body. There is experimental
evidence suggesting that the motor input to the muscle is progressively reduced during
development of fatigue in vivo (Marsden et aI., 1971; Bigland-Ritchie et aI., 1986; Garland
& McComas, 1990; Gandevia et aI., Chapter 20). This implies that the altered contractile
state of the muscle is somehow sensed and that this information is utilized in a feed-back
loop to modulate the frequency of motor stimuli to the muscle. Failure of activation may
therefore playa less significant role during development of fatigue in vivo than it does when
a muscle is fatigued in an organ bath where no restrictions are placed on the stimulation
program.

ACKNOWLEDGMENT

The author wishes to thank Dr. Edwin E. Gilliam for his helpful comments on the
manuscript. The author's laboratory has been supported by grants from the Swedish Medical
Research Council (project No. 14X-184), the Crafoord Foundation, and the Medical Faculty,
University of Lund, Sweden. Attendance of the author at the 1994 Bigland-Ritchie confer-
ence was supported, in part, by the Muscular Dystrophy Association (USA) and the American
Physical Therapy Association (Research and Analysis Division).

REFERENCES
Allen D, Duty S & Westerblad H (1993). Letter to the editors. Journal of Muscle Research and Cell Motility
14, 543-544.
Allen DG, Lee JA & Westerblad H (1989). Intracellular calcium and tension during fatigue in isolated single
muscle fibres from Xenopus laevis. Journal ofPhysiology (London) 415, 433-458.
Altringham JD & Johnston IA (1985). Effects of phosphate on the contractile properties offast and slow muscle
fibres from an antarctic fish. Journal of Physiology (London) 368, 491-500.
Bigland-Ritchie BR, Dawson NJ, Johansson RS & Lippold OCJ (1986). Reflex origin for the slowing of
motoneurone firing rates in fatigue of human voluntary contractions. Journal ofPhysiology (London)
379,451-459.
Brenner B (1980). Effect of free sarcoplasmic Ca2+ concentration on maximum unloaded shortening velocity:
measurements on single glycerinated rabbit psoas muscle fibres. Journal ofMuscle Research and Cell
Motility 1, 409-428.
Caputo C, Edman KAP, Lou F & Sun Y-B (1994). Variation in myoplasmic Ca2+ concentration and relaxation
studied by Fluo-3 in frog muscle fibres. Journal of Physiology (London) 478, 13 7-148.
Cecchi G, Griffiths PJ & Taylor S (1986). Stiffness and force in activated frog skeletal muscle fibers.
Biophysical Journal 49, 437-451.
Chase PB & Kushmerick MJ (1988). Effects of pH on contraction of rabbit fast and slow skeletal muscle fibers.
Biophysical Journal 53, 935-946.
Cooke R, Franks K, Luciani GB & Pate E (1988). The inhibition of rabbit skeletal muscle contraction by
hydrogen ions and phosphate. Journal o/Physiology (London) 395, 77-97.
42 K. A. P. Edman

Curtin NA & Edman KAP (1989). Effects of fatigue and reduced intracellular pH on segment dynamics in
'isometric' relaxation of frog muscle fibres. Journal of Physiology (London) 413, 150-174.
Curtin NA & Edman KAP (1994). Force-velocity relation for frog muscle fibres: effects of moderate fatigue
and of intracellular acidification. Journal ofPhysiology (London) 475, 483-494.
Dantzig JA, Goldman YE, Millar NC, Lacktis J & Homsher E (1992). Reversal of the cross-bridge force-gen-
erating transition by photogeneration of phosphate in rabbit psoas muscle fibres. Journal of Physiol-
ogy (London) 451, 247-278.
Dawson MJ, Gadian DG & Wilkie DR (1978). Muscular fatigue investigated by phosphorus nuclear magnetic
resonance. Nature 274,861-866.
Dawson MJ, Gadian DG & Wilkie DR (1980). Mechanical relaxation rate and metabolism studied in fatiguing
muscle by phosphorus nuclear magnetic resonance. Journal of Physiology (London) 299,465-484.
Delay M, Ribalet B & Vergara J (1986). Caffeine potentiation of calcium release in frog skeletal muscle fibres.
Journal of Physiology (London) 375, 535-559.
Eberstein A & Sandow A (1963). Fatigue mechanisms in muscle fibres. In: Gutman E, Hnik P (eds.), The Effect
of Use and Disuse on Neuromuscular Functions, pp. 515-526. Prague: Nakladatelstvi Ceskoslovenske
akademie ved Praha.
Edman KAP (1979). The velocity of unloaded shortening and its relation to sarcomere length and isometric
force in vertebrate muscle fibres. Journal of Physiology (London) 291, 143-159.
Edman KAP (1988). Double-hyperbolic force-velocity relation in frog muscle fibres. Journal of Physiology
(London) 404, 301-321.
Edman KAP & Lou F (1990). Changes in force and stiffness induced by fatigue and intracellular acidification
in frog muscle fibres. Journal of Physiology (London) 424, 133-149.
Edman KAP & Lou F (1992). Myofibrillar fatigue versus failure of activation during repetitive stimulation of
frog muscle fibres. Journal of Physiology (London) 457, 655-673.
Edman KAP & Mattiazzi A (1981). Effects of fatigue and altered pH on isometric force and velocity of
shortening at zero load in frog muscle fibres. Journal ofMuscle Research and Cell Motility 2,321-334.
Edwards RHT, Hill DK & Jones DA (1975). Metabolic changes associated with the slowing of relaxation in
fatigued mouse muscle. Journal of Physiology (London) 251, 287-301.
Fabiato A & Fabiato F (1978). Effects of pH on the myofilaments and the sarcoplasmic reticulum of skinned
cells from cardiac and skeletal muscles. Journal of Physiology (London) 276, 233-255.
Ferenczi MA, Goldman YE & Simmons RM (1984). The dependence of force and shortening velocity on
substrate concentration in skinned muscle fibres from Rana temporaria. Journal of Physiology
(London) 350, 519-543.
Fitts RH (1994). Cellular mechanisms of muscle fatigue. Physiological Reviews 74, 49-94.
Fitts RH & Holloszy JO (1978). Effects of fatigue and recovery on contractile properties of frog muscle.
Journal ofApplied Physiology 45, 899-902.
Garcia M, Gonzalez-Serratos H, Morgan JP, Perreault CL & Rozycka M (1991). Differential activation of
myofibrils during fatigue in phasic skeletal muscle cells. Journal ofMuscle Research and Cell Motility
12,412-424.
Garland SJ & McComas AJ (1990). Reflex inhibition of human soleus muscle during fatigue. Journal of
Physiology (London) 429, 17-27.
Godt RE & Nosek TM (1989). Changes of intracellular milieu with fatigue or hypoxia depress contraction of
skinned rabbit skeletal and cardiac muscle. Journal of Physiology (London) 412, 155-180.
Gonzalez-Serratos H, Garcia M, Somlyo A, Somlyo AP & McClellan G (1981). Differential shortening of
myofibrils during development of fatigue. Biophysical Journal 33, 224a.
Hnik K, Holas M, Krekule I, Kriz N, Mejsnar J, Smiesko V, Ujec E & Vyskocil F (1976). Work-induced
potassium changes in skeletal muscle and effluent venous blood assessed by liquid ion-exchanger
microelectrodes. Pfliigers Archiv 362,84-95.
Huxley AF (1957). Muscle structure and theories of contraction. Progress in Biophysics and Biophysical
Chemistry 7, 255-318.
Huxley AF & Simmons RM (1971). Proposed mechanism of force generation in striated muscle. Nature 233,
533-538.
Jones DA & Sacco P (1989). Failure of activation as the cause of fatigue in isolated mouse skeletal muscle.
Journal of Physiology (London) 410, 75P.
Juel C (1986). Potassium and sodium shifts during in vitro isometric muscle contraction, and the time course
of the ion-gradient recovery. Pfliigers Archiv 406, 458-463.
Julian FJ & Morgan DL (1981). Variation of muscle stiffness with tension during tension gradients and constant
velocity shortening in the frog. Journal ofPhysiology (London) 319,193-203.
Myofibrillar Fatigue versus Failure of Activation 43

Kanaya H, Takauyi M & Nagai T (1983). Properties of caffeine- and potassium-contractures in fatigued frog
single twitch muscle fibres. Japanese Journal of Physiology 33, 945-954.
Kawano T, Tanokura M & Yamada K (1988). Phosphorus nuclear magnetic resonance studies on the effect of
duration of contraction in bull-frog skeletal muscle. Journal of PhYSiology (London) 407, 243-261.
Kentish JC (1986). The effects of inorganic phosphate and creatine phosphate on force production in skinned
muscles from rat ventricle. Journal ofPhysiology (London) 370, 585-604.
Klein MG, Simon BJ & Schneider MF (1990). Effects of caffeine on calcium release from the sarcoplasmic
reticulum in frog skeletal muscle fibres. Journal of Physiology (London) 425, 599-626.
Liinnergren J & Westerblad H (1989). Maximum tension and force-velocity properties of fatigued single
Xenopus muscle fibres studied by caffeine and high K+. Journal ofPhYSiology (London) 409, 473-490.
Lee JA, Westerblad H & Allen DG (1991). Changes in tetanic and resting [Ca2+]j during fatigue and recovery
of single muscle fibres from Xenopus laevis. Journal ofPhysiology (London) 433, 307-326.
Lombardi V & Piazzesi G (1990). The contractile response during steady lengthening of stimulated frog muscle
fibres. Journal ofPhysiology (London) 431, 141-171.
Lynch GS & Williams DA (1994). The effect of lowered pH on the Ca2+-activated contractile characteristics
of skeletal muscle fibres from endurance-trained rats. Experimental Physiology 79, 47-57.
Marsden CD, Meadows JC & Merton PA (1971). Isolated single motor units in human muscle and their rate
of discharge during maximal voluntary effort. Journal ofPhysiology (London) 217, 12-13P.
Metzger JM & Moss RL (1990). pH modulation of the kinetics ofa Ca2+-sensitive cross-bridge state transition
in mammalian single skeletal muscle fibres. Journal ofPhysiology (London) 428, 751-764.
Nassar-Gentina V, Passonneau N, Vergara JL & Rapoport SJ (1978). Metabolic correlates of fatigue and of
recovery from fatigue in single frog muscle fibers. Journal of General Physiology 72,593-606.
Podolsky RJ & Teichholz LE (1970). The relation between calcium and contraction kinetics in skinned muscle
fibres. Journal ofPhysiology (London) 211,19-35.
Spande JJ & Schottelius BA (1970). Chemical basis of fatigue in isolated mouse soleus muscle. American
Journal of Physiology 219, 1490-1495.
Stienen GJM, Roosemalen MCM, Wilson MGA & Elzinga G (1990). Depression of force by phosphate in
skinned skeletal muscle fibers of the frog. American Journal ofPhysiology 259, C349-357.
Stienen GJM, Versteeg PGA, Papp Z & Elzinga G (1992). Mechanical properties of skinned rabbit psoas and
soleus muscle fibres during lengthening: effects of phosphate and Ca2+. Journal of Physiology
(London) 451, 503-523.
Westerblad H & Allen DG (1991). Changes of myoplasmic calcium concentration during fatigue in single
mouse muscle fibers. Journal of General Physiology 98,615-635.
Westerblad H & Allen DG (1993). The contribution of [Ca2+]j to the slowing of relaxation in fatigued single
fibres from mouse skeletal muscle. Journal of Physiology (London) 468, 729-740.
Westerblad H, Duty S & Allen DG (1993). Intracellular calcium concentration during low-frequency fatigue
in isolated single fibers of mouse skeletal muscle. Journal ofApplied Physiology 75,382-388.
Westerblad H, Lee JA, Lamb AG, Bolsover SR & Allen DG (1990). Spatial gradients of intracellular calcium
in skeletal muscle during fatigue. Pjliigers Archiv 415, 734-740.
Woledge RC (1968). The energetics of tortoise muscle. Journal of Physiology (London) 197, 685-707.
 2

MECHANISMS OF
EXCITATION-CONTRACTION COUPLING
RELEVANT TO SKELETAL MUSCLE
FATIGUE

D. G. Stephenson,1 G. D. Lamb,1 G. M. M. Stephenson,2 and M. W. Fryer3

I School of Zoology, La Trobe University

Bundoora, Victoria 3083, Australia
2Department of Chemistry and Biology, Victoria University of Technology
Footscray, Victoria 3011, Australia
3 School of Physiology and Pharmacology, University of New South Wales
Kensington, New South Wales 2033, Australia

ABSTRACT

This review examines recent progress in elucidation of the excitation-contraction coupling

in skeletal muscle with particular reference to processes which may play an important role
in muscle fatigue.

INTRODUCTION

For the purpose of this review excitation-contraction (E-C) coupling refers to the
whole sequence of events which occur in the twitch skeletal muscle fiber between the
generation of an action potential and the activation of the contractile apparatus (Sandow,
1965). This sequence of events comprises 1) the initiation and propagation of an action
potential along the sarcolemma and into the transverse tubular system (t-system); 2) signal
transmission from the t-system to the sarcoplasmic reticulum (SR) where Ca2+ is stored; 3)
Ca2+-release from the SR into the myoplasm and the rise in myoplasmic [Ca2+] ([Ca2+]i); and
4) Ca2+-binding to the regulatory system and activation of the contractile apparatus.
The ionic composition of the t-system and of the myoplasmic environment, as well
as the functional state of major participants in the E-C coupling may change markedly during
and immediately after a period of intense stimulation of the muscle fiber. These cellular
differences brought about by sustained activation of a muscle fiber can cause a decline in
mechanical performance (power output, force, velocity of shortening) which is generally

45
46 D. G. Stephenson et al.

known as muscle fatigue (Edwards, 1983; Big1and-Ritchie & Woods, 1984; Westerb1ad et
aI., 1991; Fitts, 1994).
There are a number of excellent recent reviews which explore aspects of the E-C
coupling which are relevant to muscle fatigue (Ashley et aI., 1991; Ebashi, 1991; Rios &
Pizarro, 1991; Westerblad et aI., 1991; Dulhunty, 1992; Lamb, 1992; Rios et aI., 1992; Riiegg,
1992; Fitts, 1994; Franzini-Armstrong & Jorgensen, 1994; Meissner, 1994; Schneider, 1994;
Melzer et aI., 1995). The purpose of this overview is to provide an update on mechanisms
of E-C coupling which can play an important role in muscle fatigue.

INITIATION AND PROPAGATION OF ACTION POTENTIALS

ALONG THE SARCOLEMMA AND INTO THE T-SYSTEM

Under physiological conditions, skeletal muscle contracts in response to motoneuron

stimulation. The motoneurons conduct action potentials to the neuromuscular junction where
they cause release of acetylcholine from the nerve terminals. The transmitter activates
nonselective cationic channels in the motor end-plate region of the fiber causing a local
depolarization of the sarcolemma. This depolarization activates voltage-gated Na+ - (and K+-)
channels, triggering action potentials in the surface of twitch muscle fibers which spread out
from the motor end plate along the sarcolemma and into the t-system.
The neuromuscular transmission does not appear to be blocked during muscle fatigue
(Bigland-Ritchie et aI., 1982), but there is clear evidence that repetitive stimulation ofthe
muscle is associated with disturbances in sarcolemmal and t-system membrane excitability
(for recent reviews see Sjegaard, 1991; Fitts, 1994) which are caused by modifications in
the functional properties of ionic channels and/or by changes in the ionic gradients for K+,
Ca2+, Cl- and Na+. For example, Ca2+-dependent K+ and Cl- channels become activated (Hille,
1992) when [Ca2+]j rises above a certain level. Activation of these types of channels will
always cause a repolarization or a hyperpolarization if the K+ and Cl- gradients across the
membrane are not also modified. Since [Ca2+]j between stimulations is more elevated in the
fatigued fibers than in the resting fibers (Lee et aI., 1991; Westerblad & Allen, 1991), it is
possible that Ca2+-dependent K+ and Cl- channels would be more activated during fatigue
and thus contribute to the change in the shape of the action potential. Also, ATP-sensitive
K+ -channels become activated in fatigued fibers due to a synergistic effect of decreased
[ATP] and acidification (Fink & Liittgau, 1976; Light et aI., 1994), and mechanosensitive
K+ - (and Ca2+-) channels (Burton & Hutter, 1990) may be affected by swelling of the t-system
and of the whole fiber (Gonzales-Serratos et aI., 1978). Changes in ionic gradients across
external membranes are brought about by increased ion flow through ionic channels in
combination with altered activities of the Ca2+-ATPase pump and the Na+/K+ pump, the latter
of which is activated mainly by a rise in intracellular [Na+] and inhibited by a decrease in
[ATP] below ImM (see Fitts, 1994).
The extent to which the above complex modifications in surface and t-system
membrane excitability contribute to fatigue is difficult to quantitate because it depends on a
great number of parameters, including the precise geometry of the sarcolemmal and t-system
membranes, diffusional pathways from the surface of the fiber, density of specific ion
channels in the sarcolemma and t-system, density ofNa+/K+ pumps in the t-system, pattern
of stimulation, type of muscle fiber and type of contraction. Nevertheless, it is likely that
altered membrane excitability within the fiber plays a major role in fatigue caused by
high-frequency stimulation (>100Hz), where significant recovery of force takes place after
a rest of only a few seconds, which is consistent with the time required for the diffusion of
ions in and out of the t-system (evidence summarized by Westerblad et aI., 1991). Several
Mechanisms of Excitation-Contraction Coupling and Muscle Fatigue 47

lines of evidence indicate that altered membrane excitability plays only a minor role in
fatigue induced by prolonged intermittent stimulation. For example, the action potential was
found to recover considerably faster than force after low-frequency fatiguing stimulation
(e.g., Metzger & Fitts, 1986). Also, muscle fibers stimulated by action potentials at moderate
frequency showed similar fatigability to fibers stimulated by voltage steps under voltage-
clamp conditions using otherwise the same stimulation protocol (Gyorke, 1993). Finally,
there is conclusive evidence that other factors, which more directly determine the level of
force production, such as the sensitivity to Ca2+ of the contractile apparatus (Lee et aI., 1991;
Westerblad & Allen, 1991), and the average force produced per cross-bridge (Edman & Lou,
1990) are affected during fatigue (for further discussion see Fuglevand, Chapter 6).

SIGNAL TRANSMISSION FROM THE T-SYSTEM TO THE SR

The t-system and the SR are distinctly separate membrane compartments with the
t-tubules forming, for most of their length, specialized junctions with two neighboring SR
compartments and producing a structure which is known as a triad. Detection of t-system
depolarization and signal transmission from the t-system to the SR takes place at these
junctional regions between the t-system and the SR.
There is strong evidence (see recent reviews by Rios & Pizarro, 1991; Lamb, 1992;
Franzini-Armstrong & Jorgensen, 1994; Schneider, 1994; Melzer et aI., 1995) that the
dihydropyridine (DHP) receptors found in the junctional domain of the t-tubules are the
intramembrane molecules that serve as voltage sensors and are responsible for the asymmet-
ric charge movement associated with depolarization to the contractile threshold. The DHP
receptors appear as tetrads that are orderly arranged and there are suggestions that one tetrad
consists offour DHP receptors (see review by Franzini-Armstrong & Jorgensen, 1994). At
least some of the DHP receptors in skeletal muscle are slowly activating L-type Ca2+-chan-
nels (there is dispute about the proportion of DHP receptors acting as Ca2+-channels (see
Lamb, 1992)), but Ca2+-inflow through these channels is not necessary for contraction. Each
DHP receptor is an oligomer consisting of'five subunits Ul (185kD), U2 (143kD), f3 (54kD),
y (26kD), () (30kD), where the Ul subunit has the binding site for DHP and other Ca2+-an-
tagonists and also forms part of the L-type Ca2+-channel when incorporated into lipid bilayers
(see reviews by Catterall, 1991; Lamb, 1992). The structure of the Ul subunit resembles that
of the Na+-channel with four repeats (I-IV), each consisting of 6 hydrophobic segments
(sl-s6) thought to span the t-system membrane. The fourth segment (s4) in each of the four
repeats is positively charged and is believed to be the voltage-sensitive element of the DHP
receptor (Glossmann & Striessnig, 1990) responsible for the asymmetric charge movement
in muscle (Rios & Pizarro, 1991; Schneider, 1994). The cytoplasmic loop between repeats
II and III of the skeletal muscle DHP receptor appears to be essential for signal transmission
to the SR (Tanabe et aI., 1990). The f3 subunit of the DHP receptor plays an important
regulatory role in the Ca2+-channel function and it interacts with the Ul subunit through the
loop between repeats I and II (Pragnell et aI., 1994). The functions of the other subunits
U2, y and () are not fully determined. Recent experiments using prolonged intermittent
fatiguing stimulation have shown that charge movement is fatigue resistant (Gyorke, 1993),
suggesting that the voltage sensor in the DHP receptor is not affected in fatigue.
There are multiple and distinct sites of phosphorylation on the Ul and f3 subunits of
the DHP receptor which are differentially phosphorylated by several protein kinases (Chang
et aI., 1991), and phosphorylation of certain sites alters the channel activity of the DHP
receptors (Sculptoreanu et aI., 1993). Since the state of phosphorylation ofthe DHP receptors
may change following repetitive activation, it is conceivable that a change in the state of
48 D. G. Stephenson et al.

activation ofDHP receptors, in their Ca2+ -channel function, may play some role in muscle
fatigue (see also section on ([Ca2+] Rise in the Myoplasm)).
Electron microscope studies have revealed that the four DHP receptor particles
comprising a tetrad are located in the immediate proximity of alternate triadic feet spanning
the gap between the t-tubules and the SR membrane (Franzini-Armstrong & Jorgensen,
1994). The triadic feet are homotetrameric (-560kD/subunit) protein structures, with each
subunit having several transmembrane segments at the carboxy-terminus and a large cyto-
plasmic domain. This homotetramer functions as a single Ca2+-channel in the SR membrane
and binds one molecule of the plant alkaloid ryanodine with very high affinity. Therefore,
these tetrameric complexes are known either as the foot-spanning proteins, or as the
ryanodine receptors, or as the SR Ca2+-release channels. The SR Ca2+-release channels are
found mostly on the junctional region of the SR membrane but a small fraction have been
identified in extrajunctional regions (see Dulhunty, 1992).
If it is true that a) the DHP receptors/voltage sensors activate the SR Ca2+-release
channels by some direct physical interaction (Chandler et aI., 1976) (e.g., via the loop
between repeats II and III - see above); and b) the DHP tetrads only oppose every alternate
Ca2+-release channel, it would imply that at least half of the Ca2+-release channels are not
under the direct control ofthe voltage sensors. Nevertheless, physiological experiments have
indicated that, with the possible exception of a fast transient component, the SR Ca2+-release
in skeletal muscle is under the tight control of the DHP receptors (Jacquemond et aI., 1991;
Rios & Pizarro, 1991; Schneider, 1994). Therefore, it appears either that adjacent Ca2+-re-
lease channels without direct voltage sensor control either rapidly inactivate or are non-func-
tional.
Some reports suggest that the Ca 2+-release channels may be activated by a linking
protein such as triadin (-95kD) (Kim et aI., 1990; but see Franzini-Armstrong & Jorgensen,
1994 for a recent critical account) or via a diffusible second messenger such as inositol
1,4,5-trisphosphate (IP 3) or Ca2+, which are known to be the major messengers for activation
ofCa 2+release in smooth muscle (and non-muscle cells, Berridge & Irvine, 1989) and cardiac
muscle (Fabiato, 1985; Stern & Lakatta, 1992), respectively. However, there is a growing
amount of evidence indicating that IP 3 is unlikely to playa primary role in the physiological
coupling between the DHP receptors and Ca2+-release in skeletal muscle (see Schneider,
1994). G-protein activation, which is involved in the generation of IP 3 , is not necessary for
excitation-contraction coupling in skeletal muscle (Lamb & Stephenson, 1991 b) and heparin,
which blocks binding of IP 3 to its receptor, has little effect on the coupling in mammalian
skeletal muscle and causes a unique activation-dependent block of coupling in anuran muscle
fibers. This is inconsistent with a simple IP 3 second messenger hypothesis (Lamb et aI.,
1994c). Furthermore, the possibility that Ca2+ itself may be the primary factor in linking the
voltage sensors to the Ca2+release channels in the skeletal muscle is not likely (see Dulhunty,
1992; Melzer et aI., 1995; Schneider, 1994). Experiments in our laboratory have shown that
potent buffering of [Ca2+]j to very low levels does not interrupt normal coupling (Lamb &
Stephenson, 1990) emphasizing the view that Ca2+ is not necessary for signal transmission
between the DHP receptor and the SR Ca2+-release channel. However, it is likely that Ca2+
ions do have an important role in modulating the open time of the Ca2+-release channel (see
below).
An interesting possibility is that Mg2+ plays a key role in the coupling of the DHP
receptor to the SR Ca 2+-release channel (Lamb & Stephenson, 1991a, 1992). Under resting
physiological conditions, Mg2+ (-lmM) exerts a powerful inhibitory action on the SR Ca2+
release channels keeping them closed (Endo, 1985; Lamb & Stephenson, 1991 a; Meissner,
1994) despite a strong stimulatory effect of ATP on channel opening. This inhibitory effect
ofMg2+ is mainly due to Mg2+ binding to a low-affinity inhibitory site on the SR Ca2+-release
channel (Meissner, 1994; Lamb, 1993). It has been proposed that activation of the DHP
Mechanisms of Excitation-Contraction Coupling and Muscle Fatigue 49

receptors decreases the affinity of the inhibitory site for Mg2+ by a factor of 10-20 fold,
causing dissociation of Mg2+ and removal of the Mg2+ block. Without Mg2+ inhibition, the
Ca2+-release channels open in the presence of ATP a substantial propOJ tion of time, allowing
Ca2+ release from the SR (Meissner et aI., 1986). An increase in myoplasmic [Ca2+] would
cause Ca2+ binding to an activation site on the channel which in tum could further increase
the fractional open time of the channel (see section on (Modulators ofCa2+ release)). In this
way, Ca2+ can have a positive feedback effect on Ca2+ release. However, this type of
Ca2+-induced Ca2+-release is totally under the control of the DHP receptor, by virtue of the
Mg2+ inhibition which can be reimposed at any time by the deactivation of the DHP receptor.
This proposal extends the earlier proposal made by Ashley and Moisescu (1973; see also
Liittgau & Stephenson, 19861; Ashley et aI., 1991) where the rate ofCa2+release is a function
of both [Ca2+]j and the membrane depolarization and readily explains all findings to date.
The model also reconciles and unifies the apparently contradictory hypotheses of Ca 2+ -in-
duced Ca 2 +-release (Endo, 1985; Fabiato, 1985), which has received strong support by direct
studies of the SR Ca2+-release channel (Meissner, 1994; Gyorke et aI., 1994), and of voltage
control of the non-inactivating Ca2+-release component which is well established (Rios &
Pizarro, 1991; Schneider, 1994).

CA2+-RELEASE INTO MYOPLASM AND THE RISE IN

MYOPLASMIC [CA2+]

After transmission of excitation from the t-system to the SR, Ca2+ is released from
the SR into the myoplasm. The rate of Ca2+-release from the SR will directly depend on both
the average open time of the SR Ca2+-release channels and on the electrochemical gradient
for Ca2+. In the previous section, we have discussed how depolarization of the t-system may
remove the Mg2+ block, causing activation of SR Ca2+-release channels in the presence of
ATP. Here we will discuss how factors which are known to change in fatigued fibers may
further modulate the average probability of opening of the SR Ca2+ release channels and
modify the electrochemical Ca2+ gradient across the SR membrane.

Modulators of SR Ca2+-release channels relevant to fatigue

In vitro studies have indicated that the activity of the SR Ca2+-release channels can
be altered by many physiological factors including Mg2+, adenine nucleotides, Ca2+, pH,
calmodulin, inorganic phosphate [Pd, phosphorylation, lipid metabolites, IP 3, polyamines
and some muscle proteins (Meissner, 1994, as well as the membrane potential across the SR
membrane (Stein & Palade, 1988). If the activation of the Ca2+-release channels by the DHP
receptors involves simply the removal of the Mg2+ block, then the channel opening time will
be determined by the overall influence of all modulators present.

As mentioned earlier, the binding of Mg2+ to a low-affinity inhibitory site on the SR

Ca2+-release channel powerfully inhibits channel opening. Mg2+ can also reduce the Ca2+
flow through the channel by competing with Ca2+ for a common binding site within the
channel and by competing with Ca2+ for the high affinity Ca2+ activation site (Lamb, 1993;
Meissner, 1994). We have shovvn that raising [Mg2+] in myoplasm from ImM to 3mM can
have a marked inhibitory effect on the amount of Ca2+ released by depolarization and that
at 10mM Mg2+ the transmission of excitation was completely blocked (Lamb & Stephenson,
50 D. G. Stephenson et al.

1991 a). This inhibitory effect ofMg2+can play an important role later in fatigue when [Mg2+]
begins to rise well above resting levels (Westerblad & Allen, 1992).

[ATP]
The coupling between depolarization and Ca2+-release was also blocked when [ATP]
in the fiber was very low « 1JlM) but was not obviously affected when [ATP] was changed
between 2 and 8mM (Lamb & Stephenson, 1991a). In extreme fatigue, the average [ATP]
is unlikely to decrease below 2mM (for review see Fitts, 1994), but it is possible that the
local [ATP] is much lower and quite different in the junctional area than in the rest of the
myoplasm. This idea is supported by the experiments of Han and colleagues (1992), which
have indicated that isolated skeletal muscle triads contain glycolytic enzymes that can
synthesize ATP which is not readily exchangeable with the bulk myoplasmic ATP.

Ca2+ ions may not be necessary for skeletal muscle coupling between the DHP
receptors and the SR Ca2+-release channels but they do play important roles in modulating
channel opening. Thus, Ca2+ binding to a high-affinity activating site of the SR Ca2+-release
channels promotes opening, whereas Ca2+-binding to the low-affinity Mg2+ inhibitory site
prevents the channels from opening (see Lamb, 1993; Meissner, 1994). Recently, Gyorke
and colleagues (1994) have shown that SR Ca2+-release channels from skeletal muscle
activate rapidly (in the absence of ATP and Mg2+) following flash photolysis of caged Ca2+
and then deactivate slowly, a phenomenon which was interpreted to indicate channel
adaptation. However, it is still not clear whether this adaptation is a consequence of a brief,
large [Ca2+] spike generated by the rapid photolysis of caged Ca2+, rather than an intrinsic
property of the Ca2+-release channel (Lamb et aI., 1994a).
The resting [Ca2+]j in a fatigued fiber is elevated and this would cause an increase in
the opening time of the SR Ca2+-release channels particularly when activated by depolari-
zation, assuming that the influence on the release channel of all other factors remains the
same. However, the magnitude of the Ca2+ transient is normally reduced in a fatigued fiber,
indicating that there is reduced Ca2+ flow through the channel which may be due, for
example, to a reduced SR luminal [Ca2+] (see below). A reduced SR luminal [Ca2+] may also
have a direct excitatory or inhibitory effect on the opening of the Ca2+-release channels, but
the results are not yet conclusive (see Meissner, 1994).
Raised [Ca2+]j may also be important in a quite different way, as we have recently
described a novel phenomenon in which raised myoplasmic [Ca2+] uncouples the DHP
receptor from the SR-Ca2+-release channel, and coupling cannot be restored for the duration
of the experiment (Lamb et aI., 1994b). This phenomenon may underlie low-frequency
fatigue, which occurs after prolonged exercise and persists for more than one day (Wester-
blad et aI., 1993).

pH
Experiments with SR vesicles and isolated channels have also shown that pH can be
a powerful modulator of the SR-Ca2+-release channels (Rousseau & Pinkos, 1990; Meissner,
1994). The channels are open for a considerable amount of time at pH 8.0 but activation by
Ca2+ is almost abolished by decreasing the pH below 6.5. Since muscle fatigue is often
accompanied by a decrease in pH it has been suggested (see Fitts, 1994) that acidification
of the myoplasm was directly responsible for the reduced Ca2+ release observed in fatigued
fibers. However, in contrast to this potent inhibitory effect of acid pH on Ca2+-activation of
Mechanisms of Excitation-Contraction Coupling and Muscle Fatigue 51

the SR Ca2+-release channels, our experiments with functional skinned fibers showed that
normal voltage-sensor control of Ca2+ release was virtually unaffected by pH in both
mammalian and anuran muscle (pH 6.1 to 8.0; Lamb et aI., 1992; Lamb & Stephenson, 1994).
This result shows that the acidification observed in fatigued fibers is not directly responsible
for the reduced Ca2+ release and highlights the need for examining channel release properties
in functioning fibers as well as isolated systems.

Calmodulin
Calmodulin is known to exert a direct inhibitory effect on the SR Ca2+-release channel
at physiological concentrations (see Meissner, 1994) by reducing the channel opening time.
The concentration of free calmodulin is likely to be lower in a fatigued fiber than in a rested
fiber because the higher [Ca2+]j in the fatigued fiber and the very high affinity constant of the
Ca4 -calmodulin complex for the regulatory units of the Ca2+-calmodulin dependent protein
kinases, will increase the amount ofthe bound form ofcalmodulin. Therefore, one would expect
a net stimulatory effect on the opening time of the depolarization-activated Ca2+-release
channel in the fatigued fiber due to the reduced concentration of free calmodulin.

Another modulator of Ca2+-release channel activity which plays multiple roles in

muscle fatigue is Pj (see later). Recently, it has been shown that inorganic phosphate in the
mM range concentrations has a stimulatory effect on the Ca2+-release channels of skeletal
muscle (Fruen et aI., 1994). Since [Pi] rises early during fatigue, this observation may explain
why [Ca2+]j between stimulations also increases early in fatigue (Lee et aI., 1991; Westerblad
& Allen, 1992).

Phosphorylation
Phosphorylation of the SR Ca2+-release channel of skeletal muscle may play an
important physiological role in muscle fatigue, where Ca2+-calmodulin dependent protein
kinases would be more active due to increased [Ca2+]i. In one patch-clamp study, phospho-
rylation caused inactivation of the SR Ca2+-release channel (Wang & Best, 1992) while in
other studies phosphorylation caused activation of the channel, seemingly by removal of the
Mg2+ block (Hain et aI., 1993) or by increasing the channel's sensitivity for Ca2+ and ATP
(Herrmann-Frank & Varsanyi, 1993}.1t is therefore difficult to predict at this stage whether
phosphorylation of the skeletal muscle Ca2+-release channel will contribute to an increase
or decrease in depolarization-induced release of Ca2+ from the SR.

Fiber Volume
The volume offatigued fibers can be as high as 180% of that at rest (Gonzales-Ser-
ratos et aI., 1978), raising the possibility that a change in fiber morphology associated with
this change in volume may affect Ca2+-release. However, the coupling between the depolari-
zation of the t-system and Ca2+-release was not modified by changes in the myofilament
lattice similar to those known to occur in fatigued fibers (Lamb et aI., 1993).

Other Modulators
Of the other known modulators of SR Ca2+-release channels (see Meissner, 1994),
the fatty acids and their derivatives (El-Hayek et aI, 1993) may play some role in muscle
52 D. G. Stephenson et aI.

fatigue. This is because the skeletal muscle is known to switch from a fatty acid degradation
pathway in the rested state to a glycogenolytic/glycolytic pathway during exercise (Stryer,
1988). Therefore, the concentration of the fatty acids and their derivatives which have a
potent stimulatory effect on the Ca2+ release channel (EI-Hayek et aI., 1993) will be likely
to be decreased in fatigue.

Modifications in the Electrochemical Ca2+ Gradient across the SR

Membrane
An often overlooked, but most important factor determining the electrochemical Ca2+
gradient across the SR membrane and thus the rate ofCa2+-release from the SR, is the [Ca2+]
within the junctional SR. The SR luminal [Ca2+] is dependent on several parameters
including the activity of the SR Ca2+ pump (Inesi & de Meis, 1989), rate ofCa2+ efflux from
the SR as well as the binding properties of the sarcotubular Ca2+-binding proteins (calse-
questrin and calreticulin in the junctional SR and the 53 and 160kD glycoproteins in the
longitudinal SR). The amount of calcium in the SR is not significantly altered by an increase
in myoplasmic [Mg2+] from 1 to 3 mM, similar to that known to occur in fatigued fibers
(Kabbara & Stephenson, 1994) but it may be affected by a decreased pH in fatigued fibers.
An acidic pH inhibits the SR Ca2+-pump, reduces the opening probability of the SR
Ca2+-release channels and lowers the affinity constant of calsequestrin (see Lamb et aI., 1992
for references).
More importantly, the SR luminal [Ca2+] is dependent on the concentration ofPj. As
[Pj] builds up during muscle contraction, a threshold is reached above which a precipitate
of CaHP0 4 is formed within the SR, causing a reduction in SR luminal [Ca2+]. We have
shown that 1) significant Ca precipitation occurred within the SR in the presence ofPj levels
observed in fatigue; and 2) a stimulus-induced response was prolonged and of reduced
amplitude when conditions were favorable for the formation of a CaHP04 precipitate (Fryer
et aI., 1995). The reduction in the amplitude of the response is due to the decrease of [Ca2+]
in the SR lumen caused by precipitation, whereas the prolongation of the response is due to
the CaHP0 4 precipitate acting as a slow Ca2+-buffer.

[Ca2+] Rise in the Myoplasm

The magnitude and the time course of the [Ca2+]j changes in the myoplasm following
stimulation are the result of the interplay between Ca2+ release from the SR, Ca2+ uptake by
the SR, Ca2+ entry through the DHP channels in the sarcolemma, Ca2+ removal by the
Na+/Ca2+ exchanger and the Ca2+ pump in the sarcolemma, and the Ca2+ movements
associated with various Ca2+binding sites in the myoplasm, including parvalbumin, troponin
C, SR Ca2+ pump sites and regulatory myosin light chains (Luttgau & Stephenson, 1986;
Ashley et aI, 1991; Schneider, 1994). The magnitude of the Ca2+ transient is only a very
small fraction (-1 %) ofthe total amount ofCa2+ released from the SR (Luttgau & Stephen-
son, 1986; Ashley et aI., 1991) and that the peak of the Ca2+ transient occurs before positive
force generation (Claflin et aI., 1994).
Late during fatigue, Ca2+ transients become more prolonged and of reduced magni-
tude (Westerblad & Allen, 1991). These changes in the time course and magnitude of the
Ca2+ transient can be accounted for to a large extent by changes in the Ca2+ fluxes associated
with the SR brought about by changes in the myoplasmic ionic composition with respect to
[Pj], pH, and [Mg 2+] (see above).
It is of interest that Ca2+-inflow through L-type Ca2+-channels in the sarcolemma may
significantly contribute to larger Ca2+-transients and therefore cause a delay in the onset of
Mechanisms of Excitation-Contraction Coupling and Muscle Fatigue 53

fatigue when they are phosphorylated by cAMP dependent protein kinase, because then the
channels show marked voltage-dependent potentiation (Sculptoreanu et aI., 1993).

Ca 2+-BINDING TO THE REGULATORY SYSTEM AND

ACTIVATION OF CONTRACTION

The contractile apparatus of vertebrate skeletal muscle becomes activated following

the binding of Ca2+ ions to specific sites on the Ca 2+ regulatory system comprising troponin
C, troponin I, troponin T and tropomyosin (for reviews see Luttgau & Stephenson, 1986;
Ashley et aI., 1991; Ebashi, 1991; Ruegg, 1992). Sensitivity to Ca2+ is conferred by troponin
C but interactions between different myofibrillar components including myosin and actin
(Lehrer, 1994) can have marked modulatory effects on the relationship between the isometric
force response and [Ca2 +]i. Of particular relevance to fatigued muscle fibers is the depression
in sensitivity to Ca2+ of the contractile activation caused by an increase in [Pi] (Cooke et aI.,
1988; Godt & Nosek, 1989; Ruegg, 1992; Fitts, 1994; Fryer et aI., 1995), increase in [Mg2+]
(Lamb & Stephenson 1991a, 1992; Ruegg, 1992) and decrease in pH (Metzger & Moss,
1987; Chase & Kushmerick, 1988; Lamb et aI., 1992; Ruegg, 1992; Fitts; 1994). An increase
in muscle temperature which occurs during intense exercise could further reduce the
sensitivity to Ca2+ of the contractile response (Stephenson & Williams, 1981, 1985).
However, it is generally ignored that other changes occurring in the fatigued fibers, such as
the Ca2+-calmodulin dependent phosphorylation of the myosin regulatory light chains
(Ruegg, 1992; Stephenson & Stephenson, 1993) and the decrease in creatine phosphate
concentration (Fryer et aI., 1995) can actually enhance the Ca 2+-sensitivity of contractile
activation. For example, we have shown that when the concentration of creatine phosphate
and [Pi] were varied reciprocally (as should occur in vivo), the inhibitory effects of Pi on the
Ca2+ sensitivity of the contractile apparatus were greatly reduced.

CONCLUDING REMARKS

There are many events in the E-C coupling of a skeletal twitch muscle fiber which
are altered by intense muscle fiber activity. Individually, these alterations could lead to either
a decrease or a potentiation of the contractile response, but their net effect is a decrease in
force and power output that is known as muscle fiber fatigue.

ACKNOWLEDGMENTS

Support from the National Health & Medical Research Council and the Research
Council of Australia is gratefully acknowledged. Attendance of [Link]. at the 1994
Bigland-Ritchie conference was supported, in part, by the American Physical Therapy
Association (Section on Research), and the Muscular Dystrophy Association (USA).

REFERENCES
Ashley CC & Moisescu DG (1973). The mechanism of the free calcium change in single muscle fibres during
contraction. Journal ofPhysiology (London) 231, 23-25P.
Ashley CC, Mulligan IP & Lea TJ (1991). Ca 2+ and activation mechanisms in skeletal muscle. Quarterly
Reviews of Biophysics 24, 1-73.
54 D. G. Stephenson et al.

Berridge MJ & Irvine RF (1989). Inositol phosphates and cell signalling. Nature 341, 197-205.
Bigland-Ritchie B, Kukulka OC, Lippold OCJ & Woods 11 (1982). The absence of neuromuscular transmission
failure in sustained maximal voluntary contractions. Journal ofPhysiology (London) 330, 265-278.
Bigland-Ritchie B & Woods 11 (1984). Changes in muscle contractile properties and neural control during
human muscular fatigue. Muscle & Nerve 7, 691-699.
Burton FL & Hutter OF (1990). Sensitivity to flow of intrinsic gating in inwardly rectifying potassium channel
from mammalian skeletal muscle. Journal of Physiology (London) 424, 253-261.
Catterall WA (1991). Excitation-contraction coupling in vertebrate skeletal muscle: a tale of two calcium
channels. Cell 64, 871-874.
Chandler WK, Rakowski RF & Schneider MF (1976). Effects of glycerol treatment and maintained depolari-
zation on charge movement in skeletal muscle. Journal ofPhysiology (London) 254, 285-316.
Chang CF, Gutiener LM, Meudina-Weilenmann C & Hosey MM (1991). Dihydropyridine-sensitive calcium
channels from skeletal muscle. II, Functional effects of differential phosphorylation of channel
subunits. Journal of Biological Chemistry 266, 16395-16400.
Chase PB & Kushmerick MJ (1988). Effects of pH on contraction of rabbit fast and slow skeletal muscle fibres.
Biophysical Journal 53, 935-946.
Claflin DR, Morgan DL, Stephenson DG & Julian FJ (1994). The intracellular Ca2+-transient and tension in
frog skeletal muscle fibres measured with high temporal resolution. Journal of Physiology (London)
475,319-325.
Cooke R, Franks K, Luciani GB & Pate E (1988). The inhibition of rabbit skeletal muscle contraction by
hydrogen ion and phosphate. Journal ofPhysiology (London) 395, 77-97.
Dulhunty AF (1992). The voltage-activation of contraction in skeletal muscle. Progress in Biophysics and
Molecular Biology 57,181-223.
Ebashi S (1991). Excitation-contraction coupling and the mechanism of muscle contraction. Annual Review
ofPhysiology 53,1-16.
Edman KAP & Lou F (1990). Changes in force and stiffness induced by fatigue and intracellular acidification
in frog muscle fibres. Journal ofPhysiology (London) 424,133-149.
Edwards RHT (1983). Biochemical bases of fatigue in exercise performance: catastrophe theory of muscular
fatigue. In: Knuttgen HG (ed.), Biochemistry ofExercise, pp. 3-28. Champaign, IL: Human Kinetics.
EI-Hayek R, Valdivia C, Valdivia HH, Hogan K & Coronado R (1993). Palmitoyl camitine: Activation of the
Ca2+ release channel of skeletal muscle sarcoplasmic reticulum by palmitoyl camitine and related
long chain fatty acids derivatives. Biophysical Journal 65, 779-789.
Endo M (1985). Calcium release from sarcoplasmic reticulum. Current Topics in Membranes and Transport
25, 181-230.
Fabiato A (1985). Time and calcium dependence on activation and inactivation of calcium induced release of
calcium from the sarcoplasmic reticulum ofa skinned canine cardiac Purkinje cell. Journal ofGeneral
Physiology 85,247-289.
Fink R & Luttgau HCh (1976). An evaluation ofthe membrane constants and the potassium conductance in
metabolically exhausted muscle fibres. Journal of Physiology (London) 263, 215-238.
Fitts RH (1994). Cellular mechanisms of muscle fatigue. Physiological Reviews 74, 49-94.
Franzini-Armstrong C & Jorgensen AO (1994). Structure and development ofE-C coupling units in skeletal
muscle. Annual Review of Physiology 56, 509-534.
Fruen BR, Mickelson JR, Shomer NH, Roghair TJ & Louis CF (1994). Regulation of the sarcoplasmic
reticulum ryanodine receptor by inorganic phosphate. Journal ofBiological Chemistry 269, 192-198.
Fryer MW, Owen VJ, Lamb GD & Stephenson DG (1995). Effects of creatine phosphate and Pi on force
development and Ca2+ movements in rat skinned skeletal muscle fibres. Journal of PhYSiology
(London) 482, 123-140
Glossmann H & Striessnig J (1990). Molecular properties of calcium channels. Reviews in Physiology.
Biochemistry and Pharmacology 114, 1-105.
Godt RE & Nosek TM (1989). Changes of intracellular milieu with fatigue or hypoxia depress contraction of
skinned rabbit skeletal and cardiac muscle. Journal of Physiology (London) 412, 155-180.
Gonzales-Serratos H, Somlyo AV, McClellan G, Shuman H, Borrero LM & Somlyo AP (1978). Composition
of vacuoles and sarcoplasmic reticulum in fatigued muscle: electron probe analysis. Proceedings of
the National Academy of Sciences USA 75, 1329-1333.
Gyorke S (1993). Effects of repeated tetanic stimulation on excitation-contraction coupling in cut muscle fibres
of the frog. Journal ofPhysiology (London) 464, 699-710.
Gyorke S, Velez P, Suarez-Isla B & Fill M (1994). Activation of single cardiac and skeletal ryanodine receptor
channels by flash photolysis of caged Ca2+. (Biophysical Journal) 66, 1879-1886.
Mechanisms of Excitation-Contraction Coupling and Muscle Fatigue 55

Hain J, Schindler H, Nath S & Fleischer S (1993). Phosphorylation of the skeletal muscle calcium release
channel removes block by magnesium ions. Biophysical Journal 64, A 151.
Han JW, Thieleczek R, VarsBnyi M & Heilmeyer LMO (1992). Compartmentalized ATP synthesis in skeletal
muscle triads. Biochemistry 31, 377-384.
Herrmann-Frank A & VarsBnyi M (1993). Enhancement of Ca 2+ release channel activity by phosphorylation
of the skeletal muscle ryanodine receptor. FEBS Letters 333, 237-242.
Hille B (1992). Ionic Channels of Excitable Membranes, pp. 115-139. Sutherland, MA: Sinauer.
Inesi 0 & De Meis L (1989). Regulation of steady-state filling in sarcoplasmic reticulum. Journal ofBiological
Chemistry 264, 5929-5936.
Jacquemond V, Csernock L, Klein MO & Schneider MF (1991). Voltage-gated and calcium-gated release
during depolarization of skeletal muscle. Biophysical Journal 60, 867-873.
Kabbara AA & Stephenson DO (1994). Effects of Mg2+ on Ca 2+ handling by the sarcoplasmic reticulum in
skinned skeletal and cardiac muscle fibres. Pflugers Archiv 428,331-339.
Kim KC, Caswell AH, Talvenheimo JA & Brandt NR (1990). Isolation of a terminal cisterna protein which
may link the dihydropyridine receptor to the junctional foot protein in skeletal muscle. Biochemistry
29,9283-9289.
Lamb OD (1992). DHP receptors and excitation-contraction coupling. Journal of Muscle Research and Cell
Motility 13, 394-405.
Lamb OD (1993). Ca 2+-inactivation, Mg2+-inhibition and malignant hyperthermia. Journal ofMuscle Research
and Cell Motility 14, 554-556.
Lamb OD, Fryer MW & Stephenson DO (1994a). Technical comment: Ca2+-induced Ca 2+-release in response
to flash photolysis. Science 263, 986-987.
Lamb OD, lunankar P & Stephenson DO (1994b). Abolition of excitation-contraction coupling in skeletal
muscle by raised intracellular [Ca2+]. Proceedings ofthe Australian Physiological and Pharmacologi-
cal Society 25, 76P.
Lamb OD, Posterino OS & Stephenson DO (1994c). Effects of heparin on excitation-contraction coupling in
skeletal muscle fibres of toad and rat. Journal of Physiology (London) 474, 319-329.
Lamb OD, Recupero E & Stephenson DO (1992). Effect of myoplasmic pH on excitation-contraction coupling
in skeletal muscle fibres of the toad. Journal of Physiology (London) 448, 211-224.
Lamb OD & Stephenson DO (1990). Control of calcium release and the effect of ryanodine in skinned muscle
fibres of the toad. Journal of Physiology (London) 423, 519-542.
Lamb OD & Stephenson DO (199Ia). Effect ofMg2+ on the control ofCa 2+ release in skeletal muscle fibres
of the toad. Journal of Physiology (London) 434, 507-528.
Lamb OD & Stephenson DO (1991 b). Excitation-contraction coupling in skeletal muscle fibres of rat in the
presence of OTPyS. Journal of Physiology (London) 444, 65-84.
Lamb OD & Stephenson DO (1992). Importance ofMg2+ in excitation-contraction coupling in skeletal muscle.
News in Physiological Sciences 7, 270-274.
Lamb OD & Stephenson DO (1994). Effect of intracellular pH and [Mg2+] on excitation-contraction coupling
in skeletal muscle fibres of the rat. Journal of Physiology (London) 478, 331-339.
Lamb OD, Stephenson DO & Stienen OJM (1993). Effects of osmolality and ionic strength on the mechanism
of Ca2+ release in skinned skeletal muscle fibres of the toad. Journal of Physiology (London) 464,
629-648.
Lee lA, Westerblad H & Allen DO (1991). Changes in tetanic and resting [Ca 2+]i during fatigue and recovery
of single muscle fibres from Xenopus laevis. Journal of Physiology (London) 433, 307-326.
Lehrer SS (1994). The regulatory switch of the muscle thin filament: Ca 2+ or myosin heads? Journal ofMuscle
Research and Cell Motility 15, 232-236.
Light PE, Comtois AS & Renaud JM (1994). The effect of glibenclamide on frog skeletal muscle: evidence
for K+ ATP channel activation during fatigue. Journal of Physiology (London) 475, 495-507.
Liittgau HCh & Stephenson DO (1986). Ion movements in skeletal muscle in relation to the activation of
contraction. In: Andreoli TE, Hoffman JF, Fanestil DD Schultz SO (eds.), Physiology of Membrane
Disorders, pp. 449-468. New York: Plenum Publishing Corporation.
Meissner 0 (1994). Ryanodine receptor/Ca2+ release channels and their regulation by endogenous effectors.
Annual Review of Physiology 56, 485-508.
Meissner 0, Darling E & Eveleth J (1986). Kinetics of rapid Ca2+ release by sarcoplasmic reticulum. Effects
ofCa2+, Mg2+ and adenine nucleotides. Biochemistry 25,236-244.
Melzer W, Herrmann-Frank A & Liittgau HCh (1995). The role of Ca 2+ ions in excitation-contraction coupling
in skeletal muscle fibres. Biochimica et Biophysica Acta In press.
Metzger JM & Fitts RH (1986). Fatigue from high- and low-frequency muscle stimulation: role of sarcolemma
action potentials. Experimental Neurology 93, 320-333.
56 D. G. Stephenson et al.

Metzger 1M & Moss RL (1987). Greater hydrogen ion induced depression of tension and velocity in skinned
single fibres of rat fast than slow muscles. Journal ofPhysiology (London) 393, 727-742.
Pragnell M, De Waard M, Mori Y, Tanabe T, Snutch TP & Campbell KP (1994). Calcium channel beta-subunit
binds to a conserved motif in the I-II cytoplasmic linker of the alpha I-subunit. Nature 368,67-70.
Rios E & Pizarro G (1991). Voltage sensor of excitation-contraction coupling in skeletal muscle. Physiological
Reviews 71, 849-908.
Rios E, Pizarro G & Stefani E (1992). Charge movement and the nature of signal transduction in skeletal
muscle excitation-contraction coupling. Annual Review ofPhysiology 54, 109-133.
Rousseau E & Pinkos J (1990). pH modulates conducting and gating behaviour of single calcium release
channels. Pjlugers Archiv 415, 645-647.
Riiegg JC (1992). Calcium in Muscle Contraction. Cellular and Molecular Physiology, 2nd Edition, 354pp.
Berlin, Heidelberg: Springer Verlag.
Sandow A (1965). Excitation-contraction coupling in skeletal muscle. Pharmacological Reviews 17, 265-320.
Schneider MF (1994). Control of calcium release in functioning muscle fibres. Annual Review ofPhysiology
56, 463-484.
Sculptoreanu A, Scheuer T & Catterall WA (1993). Voltage-dependent potentiation ofL-type Ca2+ channels
due to phosphorylation by cAMP-dependent protein kinase. Nature 364, 240-243.
Sjegaard G (1991). Role of exercise-induced potassium fluxes underlying muscle fatigue: a brief review.
Canadian Journal of Physiology and Pharmacology 69, 238-245.
Stein P & Palade P (1988). Sarcoballs: direct access to sarcoplasmic reticulum Ca2+-channels in skinned frog
muscle fibres. Biophysical Journal 54, 357-363.
Stephenson DG & Williams DA (1981). Calcium-activated force responses in fast- and slow-twitch skinned
muscle fibres from the rat. Journal ofPhysiology (London) 317, 281-302.
Stephenson DG & Williams DA (1985). Temperature-dependent calcium sensitivity changes in skinned muscle
fibres of rat and toad. Journal ofPhysiology (London) 360, 1-12.
Stephenson GMM & Stephenson DG (1993). Endogenous MLC2 phosphorylation and Ca2+-activated force
in mechanically skinned skeletal muscle fibres of the rat. Pjlugers Archiv 424,30-38.
Stem MD & Lakatta EG (1992). Excitation-contraction coupling in the heart: the state of the question. FASEB
Journal 6, 3092-3100.
Stryer L (1988). Biochemistry, 3rd Edition, pp.634-635. New York: Freeman and Co.
Tanabe T, Beam KG, Adams BA, Nicodome T & Numa S (1990). Regions of the skeletal muscle dihydropyrid-
ine receptor critical for excitation-contraction coupling. Nature 346,567-569.
Wang J & Best PM (1992). Inactivation of the sarcoplasmic reticulum by protein kinase. Nature 359,739-741.
Westerblad H & Allen DG (1991). Changes in myoplasmic calcium concentration during fatigue in single
mouse muscle fibres. Journal of General Physiology 98,615-635.
Westerblad H & Allen DG (1992). Myoplasmic free Mg2+ concentration during repetitive stimulation of single
fibres from mouse skeletal muscle. Journal ofPhysiology (London) 453, 413-434.
Westerblad H, Duty S & Allen DG (1993). Intracellular calcium concentration during low-frequency fatigue
in isolated single fibres of mouse skeletal muscle. Journal ofApplied Physiology 75,382-388.
Westerblad H, Lee JA, Liinnergren J & Allen DG (1991). Cellular mechanisms of fatigue in skeletal muscle.
American Journal oJPhysiology 261, CI95-C209.
 3

THE ROLE OF INTRACELLULAR ACIDOSIS

IN MUSCLE FATIGUE

D. G. Allen,l H. Westerblad,2 and J. Lannergren2

1 Department of Physiology
University of Sydney
New South Wales 2006, Australia
2 Department of Physiology and Pharmacology
Karolinska Institutet
S-171 77 Stockholm, Sweden

ABSTRACT

Muscle fatigue is often accompanied by an intracellular acidosis of variable size. The

variability reflects the involvement of different metabolic pathways, the presence or absence
of blood flow and the effectiveness of pH-regulating pathways. Intracellular acidosis affects
many aspects of muscle cell function; for instance it reduces maximal Ca2 +-activated force
and Ca2+ sensitivity, slows the maximal shortening velocity and prolongs relaxation. How-
ever, acidosis is not the only metabolic change in fatigue which causes each of the above.
and there are important aspects of muscle fatigue (e.g .• the failure ofCa2 + release) which do
not appear to be caused by acidosis.

INTRODUCTION

During repetitive activity, muscle pH generally falls slowly acid due to accumulation
oflactic acid. Simultaneously, muscle performance usually declines, and there is, therefore,
often a correlation between the intracellular acidosis and fatigue. This much is generally
accepted; the acidity of muscles in a fatigued deer was observed by Berzelius in 1807, and
he showed that the acid involved was the same as that in sour milk (cited in Needham, 1971).
A clear statement of the hypothesis that lactic acid accumulation is the cause of muscle
fatigue was advanced by Hill and Kupalov (1929) though, of course, the possible mecha-
nisms were unclear at that time. The questions we will address in the present review are: to
what extent, and by what mechanism(s), does acidosis cause fatigue? The literature on this
topic is large and often contradictory; there are many attempts with varying success to
correlate fatigue with pH change, and many attempts to prove or disprove that a particular
component of fatigue is caused by acidosis.
We believe that the true situation is conceptually simple though often complex in
practice; acidosis does indeed cause a number of features of fatigue though in virtually all

57
58 D. G. Allen et al.

cases there are additional causes for these features. Furthermore, the degree of acidosis is
quite variable in different muscle types and in different muscle activities. If these points are
accepted, much of the contradictory literature is easier to understand. Thus, there are now a
number of well-documented cases in which many of the features offatigue are observed e.g.,
reduced force, and slowed relaxation but there is no acidosis. Such observations do not, of
course, disprove that acidosis is important in other types of fatigue, but raise interesting
questions about the cause of the features of fatigue and about pH regulation in the muscle
under study. Equally, the fact that an acidosis occurs in a particular muscle does not prove
that it causes the reduced force that also occurs. We need a version of Koch's postulates to
help us prove cause and effect. For instance, to prove that acidosis is the sole cause of the
reduced force during fatigue, we need to show the following I) acidosis and reduced force
occur together during fatigue with similar time courses for both onset and recovery of fatigue;
2) a mechanism, usually established in a simpler preparation, linking acidosis and reduced
force; 3) quantitative agreement between the magnitude of the pH change and the reduced
force in the intact and simplified preparation; 4) when pH decreases in the absence offatigue,
it should cause a corresponding change in force; and 5) repetitive contractions produced in
such a way that there is no pH change should not lead to any change in force. Once we accept
that other mechanisms also reduce force, the approach becomes more difficult and relies
more heavily on quantitative analysis.
Fortunately, recent experimental advances have minimized some of the difficulties
in earlier approaches. It is now possible to measure the pH change and the mechanical activity
in an isolated single fiber. This is advantageous since in multicellular preparations both the
pH change and the force production may vary in different fiber types. It is also easy to change
the pH of the fiber under study quickly and reversibly so that the consequences of acidosis
independent of fatigue can be studied. Finally, in some preparations, it is possible to study
fatigue with and without acidosis by inhibiting the key pH regulatory mechanism(s).
In the present account, we first describe the range of pH changes that can be observed
in different muscles and activities. The differences in lactic acid production, in pH buffering
and in proton extrusion mechanisms which lead to these differences are considered. In a
separate section, the various muscle functions which are affected by acidosis are then
considered. Many aspects of fatigue have been dealt with at length in earlier reviews and
can be consulted particularly about issues other than pH (Westerblad et aI., 1991; Fitts, 1994).

pH CHANGES DURING FATIGUE

In a recent review, Fitts (1994) lists a number of studies in which the resting pHi was
-7.00 and the value recorded in fatigued muscle was around 6.3. These are representative
values from the literature, and one way in which they can occur is if all the lactic acid
produced by the muscle stays in the cell. If lactate accumulation is L (in mmolll) and the
buffer power over the relevant pH range is B (in mmolll/pH unit), then the ~pHi = LIB.
Typical values ofB are 40 mmolll/pH unit (Curtin, 1987) so that an increase in intracellular
lactic acid of 26 mmol/l would produce the observed acidosis of O. 7 units. Where we depart
from Fitts (1994) is the implication that other values in the literature are either artifacts or
minor departures from a general rule. In the next section, we will review the variations in
the simple story given above and give examples from the literature which demonstrate these
variations. As noted earlier, our conclusion is that the ~pHi can show large variations
(between about 0 and 0.8 pH unit), and these variations are both functionally important and
give valuable insights into fatigue mechanisms.
Role ofIntracellular Acidosis in Muscle Fatigue 59

Source of the Protons

Many intracellular reactions involve the production or absorption of protons, but for
the present purpose, we are concerned only with net reactions which occur to a large extent
in working muscle. Muscles which are designed to use oxidative metabolism as their
principle source of energy produce only CO 2 and H20 as net products, but these can leave
the cell rapidly and, therefore, do not usually accumulate. Muscles which are used in high
intensity activity generally have inadequate capacity for oxidative metabolism and rely on
anaerobic glycolysis, where the end product is lactic acid. Two further reactions which can
contribute to ~pHi are the breakdown ofPCr which absorbs a proton (Amorena et aI., 1990)
and the net breakdown of ATP which occurs when fatigue is severe (Nagesser et aI., 1993)
and which releases a proton (Smith et aI., 1993).
One important source of variation between fiber types is the relative contribution of
glycolytic vs. oxidative phosphorylation enzymes. This is the probable cause of the differ-
ence in ~pHi noted by Westerblad and Uinnergren (1988) in the different fiber types of
Xenopus muscle. They found that type II fibers had a much smaller ~pHi than type I fibers,
and Van der Laarse and coworkers (1991) subsequently showed thatthe level of the oxidative
enzyme, succinate dehydrogenase, was greater in type II fibers. A more dramatic example
occurs in glycolysis when inhibited with iodoacetate, which eliminates the ~pHi (Sahlin et
aI., 1981), or in human myophosphorylase deficiency in which the ~pHi is small (Cady et
aI., 1989).

Buffer Power
If the buffer power of different fibers was shown to vary, this could be another
possible source of ~pHi variations. The review by Roos and Boron (1981) gives a table of
buffer power from different fiber types showing a wide variation; however, the only careful
comparative study we know of showed no significant difference in deleted buffer power
between type I and II fibers ofXenopus (Curtin, 1987). It is well known that the buffer power
of any cell is greatly increased by the use of a C0 2IHCO-3 buffer (Roos & Boron, 1981), and
use of different external solutions in in vitro experiments could contribute to the ~pHi by
such a mechanism.

Removal of Protons
All cells have pH regulating mechanisms of varying degrees of complexity and
efficacy. These generally lead to removal of protons at the same rate as they are produced.
An intracellular acidosis develops only if the rate of proton production exceeds the rate of
removal. This is the probable reason that low intensity, long term muscular activity in humans
causes little or no acidosis (e.g., V011estad et aI., 1988). However, in short bursts of maximal
activity an acidosis develops because the proton removal rates cannot keep pace with the
rate of production of lactic acid. This is particularly true of contractions sufficiently strong
to collapse the arteries and cause ischemia or in human experiments in which ischemia is
intentionally induced. It also applies to lesser extent in isolated muscle studies in which,
even though superfusion continues, there is a long extracellular diffusion pathway from the
cell membrane to the perfusate. In the above examples, even if the extrusion rate of protons
across the surface membrane is adequate, the accumulation of extracellular protons inhibits
the extrusion mechanism. This is, no doubt, partly why increasing the concentration of
extracellular pH buffer can reduce the manifestations of fatigue (Mason et aI., 1986).
Several studies have attempted to identify the most important pH regulatory
pathways during fatigue. Juel (1988) studied mouse soleus muscle and found that lactate
60 D. G. AlIen'et al.

efflux (which is electroneutral and accompanied by a proton) carried part of the proton
efflux and that the Na/H exchanger also contributed to proton efflux. A subsequent study
on single, perfused mouse fibers found that there was no acidosis during repeated tetani,
presumably because in the absence of extracellular barriers the extrusion mechanisms
were sufficiently effective that there was no lactic acid accumulation (Westerblad & Allen,
1992a). This interpretation was confirmed when it was shown that blockage of the lactate
transporter with cinnamate led to the development ofa 0.4 pH unit acidosis in an identical
stimulation protocol. This study is interesting because it is possible to compare the decline
of force in a stimulation protocol which leads to fatigue both in the presence and absence
of acidosis. Fig, 1 shows that in the absence of acidosis, force fell to 67 % over 50 tetani;
whereas, after the addition of cinnamate, an acidosis of 0.4 pH unit developed and force
fell to 21 % over the same period. This experiment shows clearly that acidosis is only
one of the mechanisms which lead to force decline. In these experiments, inhibiting the
Na/H exchanger had little effect on the acidosis associated with activity; this does not
prove that it is absent, only that it is unimportant under these conditions. A similar analysis
of fatigue in the presence and absence of acidosis has been performed by Cady and
coworkers (1989). Normal humans were compared with a subject with myophosphorylase
deficiency; the latter showed no acidosis during ischemic muscle activity but fatigued
faster than the normal subjects. This reinforces the argument that other metabolic changes
also cause force decline and that these metabolic changes may be accelerated in the
absence of glycolysis.

4 mM-cinnamate

400
"iii
.............................. .............
a.. ...........
~
c:
0
'iii
c:
(I)
~
0

6·8
pH;
7·0

7·2

1 min
Figure 1. Force and intracellular pH in a single isolated mouse muscle fiber during repeated tetanic stimula-
tion. The small circles show peak force and the dashed line shows intracellular pH in a control period of
repeated tetani. Subsequently 4 mM a-cyano-4-hydroxycinnamate was applied to block the lactate transporter,
The vertical lines show force and the full line shows the intracellular pH when repeated tetani were produced
under these conditions. Note that in the absence of cinnamate there was a moderate fall of force but no
intracellular acidosis. After inhibition of the lactate transporter, the decline of force was greater and faster and
an intracellular acidosis of 0.4 pH-units developed (reproduced from Westerblad & Allen, 1992a).
Role oflntracellular Acidosis in Muscle Fatigue 61

CONSEQUENCES OF ACIDOSIS

The most obvious feature of fatigue is the accompanying decline in tetanic force.
Eberstein and Sandow (1963) first established that failure of Ca2+ release was an important
cause of the force decline by showing that caffeine and high K+, which increase Ca2+release,
could overcome much of the force decline. There is also abundant evidence from skinned
fiber studies that some of the metabolites that accumulate during fatigue can depress both
the maximum Ca2+-activated force and the myofibrillar Ca2+ sensitivity (reviewed in Wester-
blad et aI., 1991; Fitts, 1994). In this section, we, therefore, first consider the contribution
of ApH; to each of these processes. A variety of other changes in muscle function can be
observed in fatigue. Two changes that have important effects on the work output offatigued
muscles are a slowing of velocity of shortening and a reduction in the rate of relaxation. The
effects of acidosis on these functions are considered next. Finally we consider the role of
ApH; on metabolic pathways. All enzymes are pH sensitive, so it is to be expected that pH
might have an important influence here.

Failure of Ca2+ Release

Since the demonstration that fatigue involves a partial failure of excitation-contrac-
tion coupling, there has been speculation that acidosis interferes with the process (e.g.,
Nassar-Gentina et aI., 1981). Support for this hypothesis was increased when Ma and
coworkers (1988) showed that acidosis inhibited the opening of isolated sarcoplasmic
reticulum (SR) Ca2+ channels in lipid bilayers. Although this is an attractive hypothesis, the
arguments against it are strong. 1) Lamb and colleagues (1992), using skinned fibers with
intact SRit-tubular system, showed that depolarization-induced SR Ca2+ release is not
affected by pH. On the other, hand the leakage of Ca2+ from the SR (in the absence of
depolarization) was decreased by acidosis which is probably the physiological equivalent of
the observations by Ma and colleagues (1988); 2) single mouse fibers show the normal
decline in Ca2+ release in fatigue (Westerblad & Allen, 1991, 1993b), although fatigue in
these fibers is not accompanied by any marked acidosis (Fig. 1). Thus, acidosis is probably
not important for the declining tetanic [Ca2+Ji in fatigue; and 3} if intracellular acidosis is
produced in single fibers by increasing extracellular CO 2 (Fig. 2), it causes a substantial
increase in tetanic [Ca2+];. The probable cause of this increase is a reduced troponin binding
constant for Ca2+ so that the same Ca2+ release leads to a larger [Ca2+]; but a smaller force
(see subsequent section). At present, the cause of the failure ofCa2+ release during fatigue
remains unknown (see Stephenson et aI., Chapter 2). Two possibilities are that it is a
consequence of the fall in intracellular ATP concentration, which is known to occur at about
the same time (Westerblad & Allen, 1992b,c), or that it is a consequence of precipitation of
calcium phosphate in the SR (Fryer et aI., 1995).

Reduction of Maximal Isometric Force

One part of the force decline in fatigue cannot be overcome by increasing tetanic
[Ca2+]i with caffeine; and it has, therefore, been ascribed to reduced maximum force (Edman
& Lou, 1990; Liinnergren & Westerblad, 1991; Westerblad & Allen, 1991). This component
dominates during the early phases of fatigue. Studies on skinned fibers have shown that, of
the metabolic changes that can occur in fatigue, two make the major contribution to the
decline of maximum force: acidosis (due to lactic acid accumulation) and accumulation of
inorganic phosphate (Pi; due to PCr breakdown).
62 D. G. Allen et al.

1·0

[Ca2+]. 0.5
(PM) I

0·0

400 0% Figure 2. IntracelJular calcium and

force in 100 Hz tetani from isolated
single fibers of mouse muscle. Intra-
celJular pH was varied by changing
Tension the extracelJular CO2 concentration;
(kPa)
30 % CO 2 gives a pHi =6.8; 5 % CO 2
gives a pHi = 7.3; 0 % CO2 gives a
pHi = 7.9. Note that acidosis in-
o creases the [Ca2+]i during a tetanus
but reduces the tetanic force (repro-
duced from Westerblad & AlJen,
1993a).
0·2 s

A l\-
A

1~~5tR
Figure 3. The relation between [Ca2+]i
and force in intact isolated mouse
muscle fibers at various intracellular
pH. A shows selected [Ca2+]i and
0·0 force records from tetani at various
pHi produced by changing extracelJu-

400[fL Jn L
lar CO2 (0 % CO2 gives a pHi = 7.9; 5
% CO2 gives a pHi = 7.3; 30 % CO2
Tension gives a pHi =6.8). In this experiment,
the stimulus frequency was varied at
(kPa)
each pHi so as to obtain a series of
o tetani with varying [Ca2+]i during the
tetanic contraction. To obtain very
high [Ca2+]h 10 mM caffeine was
0·5 s added (filJed points in B). In A, re-
B cords were selected with a roughly
0% constant [Ca2+]i during the tetanus to
100 5% show the very different force obtained
, . _ - - - -.....-30% at different pHi' B shows the relation
o between [Ca2+]i and force at the vari-
Tension 50 ous pHi . Each point is the mean
(%) [Ca2+]i and force taken from the pla-
teau of one tetanus. Note the reduc-
tion in maximum Ca2+-activated force
and Ca2+ sensitivity in acid conditions
0·5 1·0 1·5 2·0 (reproduced from Westerblad & Al-
[Ca2+]. (PM) len, 1993a).
I
Role of Intracellular Acidosis in Muscle Fatigue 63

It is well established that acidosis depresses maximal force in skinned fibers (e.g.,
Fabiato & Fabiato, 1978; Metzger & Moss, 1987; Godt & Nosek, 1989). To show that this
also occurs in intact fibers, it is necessary to maximally activate the muscle. One approach
to this problem is to construct [Ca2+]j -force curves at different pHj as illustrated in Fig. 3
(for details see legend to Fig. 3). Our results show a clear reduction in maximal force with
acidosis in intact fibers, although the magnitude is somewhat smaller in intact than in skinned
fibers. There are, however, studies (Adams et aI., 1991) suggesting that acidosis has no effect
on tetanic force in intact muscle. In this latter study, the maximal force was not assessed and
the lack of an effect of acidosis can be partly explained by the fact that tetanic [Ca2+1 increases
with acidosis (Fig. 2). The latter would be expected to offset the reduction of force.
The exact mechanism by which acidosis reduces maximal force is not clear (Cooke
et aI., 1988; Kentish, 1991). Acidosis also reduces actomyosin ATPase rates (Cooke et aI.,
1988; Parkhouse, 1992) suggesting that crossbridge turnover rates are reduced. The fact that
acidosis generally reduces ATPase less than force suggests that there is also a reduction in
the force produced by each attached crossbridge (Orchard & Kentish, 1990). It seems
probable that protons are interacting with multiple sites associated with the actomyosin
ATPase.

Reduced Myofibrillar Ca2+ Sensitivity

In fatigue, there is often an acidosis and an accumulation ofPj. Skinned-fiber results

have shown that both of these metabolic changes reduce the myofibrillar Ca2+ sensitivity
(e.g., Fabiato & Fabiato, 1978; Metzger & Moss, 1987; Godt & Nosek, 1989). This reduction
could occur by one of two mechanisms: either via competition for the Ca2+ binding site on
troponin (Blanchard et aI., 1984) or via a reduction in strongly bound crossbridges (Bremel
& Weber, 1972; Guth & Potter, 1987). It was originally proposed that acidosis inhibited Ca2+
binding to troponin by competition between H+ and Ca2+ but recent evidence suggests that
a reduced number of strongly bound crossbridges may also contribute (Metzger & Moss,
1987). Pj, in contrast, does not change the Ca2+ binding to isolated troponin C (Kentish &
Palmer, 1989); and, hence, its depression of Ca2+ sensitivity is thought to occur solely
through reduced cooperativity due to a reduced number of strongly attached crossbridges
(Millar & Homsher, 1990).
Changes of the myofibrillar Ca2+ sensitivity during fatigue have been assessed in
single fibers from mouse and Xenopus (Allen et aI., 1989; Lee et aI., 1991; Westerblad &
Allen, 1991, 1993b). In most of these studies, a clear-cut reduction of the sensitivity was
observed. An exception was the earliest study (Allen et aI., 1989) in which the photoprotein
aequorin was used to measure [Ca2+k Aequorin light emission is reduced by Mg2+, and it
was subsequently shown that an increase in myoplasmic [Mg2+] ([Mg2+]j) occurs in fatigue
(Westerblad & Allen, 1992b,c ). When correction was made for the effect of [Mg2+]j on the
aequorin light emission, a reduction in myofibrillar Ca2+ sensitivity can be observed in these
experiments as well. There are also metabolic changes in fatigue that may increase myofi-
brillar Ca2+ sensitivity (e.g., increased ADP, Cooke & Pate, 1985; and myosin light chain
phosphorylation, Persechini et aI., 1985).
Thus to summarize, several changes in fatigue may significantly affect the myofi-
brillar Ca2+ sensitivity: acidosis and increased Pj will reduce it, while increased ADP and
myosin light chain phosphorylation may have the opposite effect. In fatigue, the net effect
of these changes is generally a reduced Ca2+ sensitivity. This reduction is especially
important for force production when tetanic [Ca2+]jdec1ines, i.e., towards the end offatiguing
stimulation.
64 D. G. Allen et al.

The Rate of Relaxation

Relaxation is a complex process involving 1) Ca2+uptake by the SR and buffers such
as parvalbumin; 2) dissociation ofCa2+ from troponin; and 3) crossbridge detachment. It is
well established that acidosis slows relaxation (Edman & Mattiazzi, 1981; their Figs. 2 and
4; Westerblad & Allen, 1993a; Westerblad & Liinnergren, 1991) but it is also known that
relaxation during fatigue can slow in the absence of acidosis (Cady et aI., 1989; Westerblad
& Allen, 1992a). There is evidence from both isolated SR and skinned fibers with functional
SR that acidosis inhibits SR Ca2+ pumping. For instance a recent study by Lamb and
colleagues (1992) estimated SR Ca2+ uptake rates by measuring mechanical responses in
skinned fibers with intact t-tubule SR connections. They showed that SR pump rate fell
roughly 2-fold between pH 7.1 and 6.6 and fell a further 2-fold between pH 6.6 and 6.1.
Such studies have led to the suggestion that the mechanism by which acidosis slows
relaxation involves inhibition of the SR Ca2+ pump. Recent measurements of [Ca2+]j during
acidosis (Westerblad & Allen, 1993a) showed little change in the fast phase of [Ca2+]j decline,
perhaps because this process is dominated by the turnoff of SR permeability, and a pro-
nounced increase in magnitude and slowing of the slow phase of [Ca2+]j decline. SR pump
analysis on the slow phase showed that the pump rate was reduced by about 3-fold between
pH 7.3 and 6.8. Similar results were seen in studies on skinned fibers with intact SR (see
above). However, to decide on the contribution of the reduced SR Ca2+ uptake to the slower
relaxation, it is necessary to consider also the reduction of myofibrillar Ca2+-sensitivity that
is caused by acidosis (see earlier section). In our study, this was done by converting the
[Ca2+]j to the appropriate steady force by means of a previously determined relation between
[Ca2+]j and force (for details see legend to Fig 4). It can be seen from Fig. 4 that despite the
elevated tetanic [Ca2+]j and the enhanced slow decline, the Ca2+-derived force in acidosis
was little changed from control because of the reduced Ca2+ sensitivity and maximum force
of the contractile proteins. Consequently it seems that the delay between the Ca2+-derived
force and the true force is increased in acidosis compared to control. This suggests that the

1000

750
~
.s
500
1:
~ 250

0 Figure 4. Relaxation in control and acid conditions follow-

ing a tetanic contraction in an isolated mouse muscle fiber. In
300 [Ca2+1rderived force
each panel the continuous line is control (pHj = 7.3) while the
os
a. dashed line shows results when an intracellular acidosis (pHj
~ 200
CD = 6.8) was produced by increased extracellular CO2 • Upper
~
panel shows the [Ca2+]j with the last stimulus shown as zero
If 100 time. Middle panel shows the Ca2+ -derived force in which
[Ca2+]j from the upper panel has been converted to force using
0~~~=-~---------1
the steady state [Ca2+]j -force relation at the appropriate pHj
300 Force
of the sort shown in Fig. 3B. Lower panel shows the measured
------,
~ 200
'\
force. Note that the [Ca2+]j is larger and initially declines
rapidly but later declines more slowly in acid conditions.
~
If 100 "- "'-
However, because of the reduced maximum force and Ca2+
sensitivity in acid conditions, the [Ca2+Jj-derived force is not
'- greatly different under acid conditions. Thus the much slower
OL---------J-~------~
decline of force in acid condition seems to be caused by
o 100 200
slower detachment of crossbridges rather than the slower rate
Time (ms after last stimulus) of[Ca2+]jdecline (modified from Westerblad & Allen, 1993a).
Role ofIntracellular Acidosis in Muscle Fatigue 65

detachment rate of crossbridges is slowed in acidosis, which is consistent with a reduced

myofibrillar ATPase in acidosis (Cooke et aI., 1988; Parkhouse, 1992).

Shortening Velocity
Fatigue is generally accompanied by a reduction of shortening speed (DeHaan et aI.,
1989; Curtin & Edman, 1994; Westerblad & Uinnergren, 1994). There is, however, a large
variability in the magnitude of this slowing, which presumably depends on the type of
preparation, stimulation regime etc. Also, there is no clear-cut relation between reductions
in isometric tension and shortening velocity in fatigue. The latter can be explained by the
fact that the metabolic changes in fatigue may have markedly different effects on isometric
force compared to shortening velocity. A striking example of this is that increased ADP
reduces the shortening velocity while isometric tension is increased (Cooke & Pate, 1985).
In an early study employing single, intact fibers of frog, Edman and Mattiazzi (1981)
observed similar changes of the shortening velocity at zero load (V0) in fatigue and in
acidosis induced by exposure to CO 2 , They concluded that reduced pHi makes a large
contribution to the reduction of shortening speed in fatigue. However, since pHi was not
measured in this study, it is not possible to say if the slowed shortening was exclusively due
to acidosis or if other factors contributed.
Studies on skinned fibers support the idea that acidosis is important for the reduced
shortening speed in fatigue. Ifwe assume a typical pHi change in fatigue of 0.7 pH-units (see
above), this gives a reduction ofVo in skinned fibers of about 25% (Metzger & Moss, 1987;
Cooke et aI., 1988), which is similar to the slowing often observed in fatigue. This result
indicates that acidosis may be responsible for all changes of the shortening velocity in
fatigue. However, in a recent study we followed changes of V0 during fatigue produced by
repeated short tetani in isolated Xenopus fibers (Westerblad & Liinnergren, 1994). In these
fibers, pHi in fatigue is reduced by about 0.6 pH-units (Westerblad & Liinnergren, 1988)
which according to skinned fiber results would give a reduction of Vo of about 20%.
However, the observed reduction in Vo was about 50%, thus suggesting that some other
factor also depresses the shortening speed.
In conclusion, acidosis is of great importance for the reduced shortening speed in
fatigue; but, at least in Xenopus fibers, there seems to be an additional factor which
contributes to the slowing.

Metabolism
Enzymes are generally pH-sensitive, and even small changes of pHi may have large
impact on enzymatic activity and hence cellular metabolism. This is of particular importance
in skeletal muscle fatigue which may be accompanied by a large decline of pHi (see above).
Acidosis reduces the glycogenolytic rate by inhibiting both glycogen phosphorylase and
phosphofructokinase (e.g., Roos & Boron, 1981). This would lead to a reduced rate of ATP
production, which might set an important limitation to muscle performance during fatigue.
However, acidosis will also reduce the rate of ATP utilization by inhibiting myofibrillar
ATPase (Cooke et aI., 1988; Parkhouse, 1992) and SR Ca 2+-pumping (Fabiato & Fabiato,
1978; Lamb et aI., 1992). Furthermore, a reduction of ATP derived from glycogenolysis will
accelerate phosphocreatine (PCr) breakdown and eventually result in a net ATP breakdown,
leading to an accumulation of Pi and AMP, both of which are known activators of glycogen
phosphorylase and phosphofructokinase (Chasiotis et aI., 1982). Thus, the overall conse-
quences of acidosis on energy metabolism during fatigue are difficult to predict. In fatigue,
the reduction of ATP is generally small (Fitts, 1994), showing that the inhibition of energy
supply by acidosis is to a large extent counteracted by other mechanisms. Nevertheless, in
66 D. G. Allen et al.

frog muscle repeated contractions at long intervals are associated with a marked depletion
ofPCr under acidic conditions while PCr remained unchanged at normal pHi (Nakamura &
Yamada, 1992). Thus, at least under these experimental conditions the inhibition of ATP
production by acidosis was not fully counteracted by a reduced rate of ATP utilization.

CONCLUDING REMARKS

Intracellular acidosis frequently, but not invariably, accompanies muscle fatigue. Its
presence or absence can be understood in terms of the pH regulatory mechanisms of the
muscle cell. The magnitude of the intracellular acidosis reflects the interplay between
production of acid metabolites, the pH buffering by the cell and the activity of the pHi
regulating mechanisms. Intracellular acidosis has multiple actions on cell function, and the
observed effects in whole muscles represent a complex summation of these effects. In a
single cell, in which pHi can be measured, it is often possible to separate the effects associated
with acidosis from those associated with the other metabolic changes occurring in fatigue.
In whole muscles, the activation pattern and the muscle fiber types are usually heterogenous,
and it is likely that the changes in pHi vary from cell to cell. These complications mean that
detailed quantitative analysis of the role of acidosis is much more difficult.

ACKNOWLEDGMENTS

We thank Dr. Jennifer McDonagh for editorial comments on the manuscript. Work
in the authors' laboratories is supported by grants from the National Health & Medical
Research Council of Australia and the Medical Research Council of Sweden. Their atten-
dance at the 1994 Bigland-Ritchie conference was supported, in part, by the Muscular
Dystrophy Association (USA), and the American Physical Therapy Association (Section on
Research; D.G.A.).

REFERENCES
Adams GR, Fisher MJ & Meyer RA (1991). Hypercapnic acidosis and increased H2P04• concentration do not
decrease force in cat skeletal muscle. American Journal of Physiology 260, C805-C812.
Allen DG, Lee JA & Westerblad H (1989). Intracellular calcium and force in isolated single muscle fibres from
Xenopus. Journal ofPhysiology (London) 415, 433-458.
Amorena CE, Wilding TJ, Manchester JK & Roos A (1990). Changes in intracellular pH caused by high K+
in normal and acidified frog muscle. Journal of General Physiology 96, 959-972.
Blanchard EM, Pan BS & Solaro RJ (1984). The effect of acidic pH on the ATPase activity and troponin Ca2+
binding of rabbit skeletal myofilaments. Journal ofBiological Chemistry 259, 3181-3186.
Bremel RD & Weber A (1972). Cooperation within actin filaments in vertebrate skeletal muscle. Nature 238,
97-101.
Cady EB, Elshove E, Jones DA & Moll A (1989). The metabolic causes of slow relaxation in fatigued human
skeletal muscle. Journal of Physiology (London) 418, 327-337.
Chasiotis D, Sahlin K & Hultman E (1982). Regulation of glycogenolysis in human muscles at rest and during
exercise. Journal ofApplied Physiology 53,708-715.
Cooke R, Franks K, Luciani GB & Pate E (1988). The inhibition of rabbit skeletal muscle contraction by
hydrogen ion and phosphate. Journal ofPhysiology (London) 395, 77-97.
Cooke R & Pate E (1985). The effects of ADP and phosphate on the contraction of muscle fibers. Biophysical
Journal 48, 789-798.
Curtin NA (1987). Intracellular pH and buffer power of type I and 2 fibres from skeletal muscle of Xenopus
laevis. Pfliigers Archiv 408,386-389.
Role ofIntracellular Acidosis in Muscle Fatigue 67

Curtin NA & Edman KAP (1994). Force-velocity relation for frog muscle fibres: effects of moderate fatigue
and of intracellular acidification. Journal ofPhysiology (London) 475, 483-494.
DeHaan A, Jones DA & Sargeant AJ (1989). Changes in velocity of shortening, power output and relaxation
rate during fatigue of rat medial gastrocnemius muscle. Pflugers Archiv 413,422-428.
Eberstein A & Sandow A (1963). Fatigue mechanisms in muscle fibers. In: Gutman E, Hink P (eds.), The Effect
of Use and Disuse on Neuromuscular Functions, pp. 515-526. Amsterdam: Elsevier.
Edman KAP & Lou F (1990). Changes in force and stiffness induced by fatigue and intracellular acidification
in frog muscle fibres. Journal ofPhysiology (London) 424,133-149.
Edman KAP & Mattiazzi AR (1981). Effects of fatigue and altered pH on isometric force and velocity of
shortening at zero load in frog muscle fibres. Journal ofMuscle Research and Cell Motility 2,321-334.
Fabiato A & Fabiato F (1978). Effects of pH on the myofilaments and the sarcoplasmic reticulum of skinned
cells from cardaic and skeletal muscles. Journal of Physiology (London) 276, 233-255.
Fitts RH (1994). Cellular mechanisms of muscle fatigue. Physiological Reviews 74, 49-94.
Fryer MW, Owen VJ, Lamb GD & Stephenson DG (1995). Effects of creatine phosphate and Pi on Ca2+
movements and tension development in rat skinned skeletal muscle fibres. Journal of Physiology
(London) 482, 123-140.
Godt RE & Nosek TM (1989). Changes of intracellular milieu with fatigue or hypoxia depress contraction of
skinned rabbit skeletal and cardiac muscle. Journal of PhYSiology (London) 412, 155-180.
Giith K & Potter JD (1987). Effect of rigor and cycling cross-bridges on the structure oftroponin C and on the
Ca2+ affmity of the Ca2+-specific regulatory sites in skinned rabbit psoas fibers. Journal ofBiological
Chemistry 262, 13627-13635.
Hill AV & Kupalov P (1929). Anaerobic and aerobic activity in isolated muscles. Proceedings of the Royal
Society (Series B) 105, 313-322.
Juel C (1988). Intracellular pH recovery and lactate efflux in mouse soleus muscles stimulated in vitro: the
involvement of sodium/proton exchange and a lactate carrier. Acta Physiologica Scandinavica 132,
363-371.
Kentish JC (1991). Combined inhibitory actions of acidosis and phosphate on maximum force production in
rat skinned cardiac muscle. Pflugers Archiv 419, 310-318.
Kentish JC & Palmer S (1989). Calcium binding to isolated bovine cardiac and rabbit skeletal troponin-C is
affected by pH but not by caffeine or inorganic phosphate. Journal ofPhysiology (London) 417, 160P.
Lamb GD, Recupero E & Stephenson DG (1992). Effect of myoplasmic pH on excitation-contraction coupling
in skeletal muscle fibres of the toad. Journal of Physiology (London) 448, 211-224.
Liinnergren J & Westerblad H (1991). Force decline due to fatigue and intracellular acidification in isolated
fibres from mouse skeletal muscle. Journal of Physiology (London) 434, 307-322.
Lee JA, Westerblad H & Allen DG (1991). Changes in tetanic and resting [Ca2+li during fatigue and recovery
of single muscle fibres from Xenopus laevis. Journal ofPhysiology (London) 433, 307-326.
Ma J, Fill M, Knudson M, Campbell KP & Coronado R (1988). Ryanodine receptor of skeletal muscle is a gap
junction-type channel. Science 242,99-102.
Mason MJ, Mainwood GW & Thoden JS (1986). The influence of extracellular buffer concentration and
propionate on lactate efflux from frog muscle. Pflugers Archiv 406,472-479.
Metzger 1M & Moss RL (1987). Greater hydrogen ion induced depression of tension and velocity in skinned
single fibres of rat fast than slow muscles. Journal of Physiology (London) 393, 727-742.
Millar NC & Homsher E (1990). The effect of phosphate and calcium on force generation in glycerinated
rabbit skeletal muscle fibers. Journal of Biological Chemistry 265, 20234-20240.
Nagesser AS, Van Der Laarse WJ & Elzinga G (1993). ATP formation and ATP hydrolysis during fatiguing,
intermittent stimulation of different types of single muscle fibres from Xenopus laevis. Journal of
Muscle Research and Cell Motility 14, 608-618.
Nakamura T & Yamada K (1992). Effects of carbon dioxide on tetanic contraction offrog skeletal muscles
studied by phosphorus nuclear magnetic resonance. Journal of PhYSiology (London) 453, 247-259.
Nassar-Gentina V, Passonneau IV & Rapoport SI (1981). Fatigue and metabolism of frog muscle fibers during
stimulation and in response to caffeine. American Journal ofPhysiology 241, CI60-CI66.
Needham DM (1971). Machina Carnis: The Biochemistry of Muscular Contraction and Its Historical
Development. Cambridge: Cambridge University Press.
Orchard CH & Kentish JC (1990). Effects of changes of pH on the contractile function of cardiac muscle.
American Journal of Physiology 258, C967-C981.
Parkhouse WS (1992). The effects of ATP, inorganic phosphate, protons, and lactate on isolated myofibrillar
ATPase activity. Canadian Journal of Physiology and Pharmacology 70, 1175-1181.
Persechini A, Stull JT & Cooke R (1985). The effect of myosin phosphorylation on the contractile properties
of skinned rabbit skeletal muscle fibers. Journal of Biological Chemistry 260, 7951-7954.
68 D. G. Allen et al.

Roos A & Boron WF (1981). Intracellular pH. Physiological Reviews 61,297-434.

Sahlin K, Edstrom L, Sjoholm H & Hultman E (1981). Effects of lactic acid accumulation and ATP decrease
on muscle tension and relaxation. American Journal ofPhysiology 240, CI21-CI26.
Smith OL, Donoso P, Bauer CJ & Eisner DA (1993). Relationship between intracellular pH and metabolic
concentrations during metabolic inhibition in isolated ferret heart. Journal of Physiology (London)
472, 11-22.
Van Der Laarse WJ, Liinnergren J & Diegenbach PC (1991). Resistance to fatigue of single muscle fibres from
Xenopus related to succinate dehydrogenase and myofibrillar ATPase activities. Experimental Physi-
ology 76, 589-596.
Vl'Jllestad NK, Sejersted OM, Bahr R, Woods JJ & Bigland-Ritchie B (1988). Motor drive and metabolic
responses during repeated submaximal contractions in humans. Journal of Applied Physiology 64,
1421-1427.
Westerblad H & Allen DO (1991). Changes in myoplasmic calcium concentration during fatigue in single
mouse muscle fibres. Journal of General Physiology 98,615-635.
Westerblad H & Allen DO (1992a). Changes of intracellular pH during repeated tetani in single mouse skeletal
muscle fibres. Journal ofPhysiology (London) 449, 49-71.
Westerblad H & Allen DO (1992b). Myoplasmic free Mg2+ concentration during repetitive stimulation of
single fibres from mouse skeletal muscle. Journal of Physiology (London) 453: 413-434.
Westerblad H & Allen DO (1992c). Myoplasmic [Mg2+l iconcentration inXenopus muscle fibres at rest, during
fatigue and during metabolic blockade. Experimental Physiology 77,733-740.
Westerblad H & Allen DO (1993a). The influence of pH on contraction, relaxation and [Ca2+li in intact single
fibres from mouse muscle. Journal ofPhysiology (London) 466, 611-628.
Westerblad H & Allen DO (1993b). The role of [Ca2+li in the slowing of relaxation in fatigued single fibres
from mouse skeletal muscle. Journal of Physiology (London) 468, 729-740.
Westerblad H & Liinnergren J (1988). The relation between force and intracellular pH in fatigued. single
Xenopus muscle fibres. Acta Physiologica Scandinavica 133, 83-89.
Westerblad H & Liinnergren J (1991). Slowing of relaxation during fatigue in single mouse muscle fibres.
Journal of Physiology (London) 434. 323-336.
Westerblad H & Liinnergren J (1994). Changes of the force-velocity relation, isometric tension and relaxation
rate during fatigue in intact, single fibres of Xenopus skeletal muscle. Journal of Muscle Research
and Cell Motility 15, 287-298.
Westerblad H, Lee JA, Liinnergren J & Allen DO (1991). Cellular mechanisms of fatigue in skeletal muscle.
American Journal ofPhysiology 261, CI95-C209.
 4

ROLE OF INTERSTITIAL POTASSIUM

G. SjegaardI and A. J. McComas 2

IDepartment of Physiology
National Institute of Occupational Health
Copenhagen, Denmark
2Department of Medicine, McMaster University
Hamilton, Ontario

ABSTRACT

Interstitial potassium concentration, [K+], is modulated during muscle activity due to a

number of different mechanisms: diffusion and active transport of K+ in combination with
water fluxes. The relative significance of the various mechanisms for muscle function is
quantified. The effect of interstitial [K+] locally on the single muscle fiber is discussed along
with its effect on the cardiovascular and respiratory systems and its role in motor control. It
is concluded that K+ may playa significant role in the prevention as well as the development
of fatigue.

INTRODUCTION

Normal function of muscle fibers is critically dependent on the electrolyte balance

within the muscles, and especially on Ca2+ and Mg2+ in the different compartments within
the cell, and on the concentration gradients of Na+ and K+ across the sarcolemma (plas-
malemma). The ultimate force regulator is cytosolic Ca 2+which in tum is controlled by fluxes
of Na+ and K+ across the membrane in relation to the action potential. Quantitatively the
fluxes of Na+ and K+ per action potential are small but nevertheless, large changes in
interstitial [K+] have been reported in many studies.

Muscle Activity Induces K+ Loss from Muscle

Fenn first showed that muscle activity induces a K+ loss (Fenn & Cobb, 1936; Fenn,
1937), and that this was proportional to the magnitude and frequency of muscle contraction
(Fenn, 1938). De Lanne and colleagues were among the first to monitor plasma [K+] with
time during and following muscle work (De Lanne et aI., 1959). K+ is highly permeable
across the capillary membrane (Crone et aI., 1978) and thus plasma [K+] in the venous
effluent reflects interstitial [K+] in a well perfused muscle. The net loss from the muscle to
the interstitial space can then be estimated as (plasma flow) x (arterial- venous)[K+]. When

69
70 G. Sjogaard and A. J. McComas

muscle blood flow is occluded (e.g., due to high tissue pressure during a contraction) no net
K+ loss from the muscle can occur. Nevertheless, intracellular K+ will diffuse down its
electrochemical gradient into the interstitial space, and using microelectrodes placed in the
interstitial space dramatic increases in [K+] have been reported (Hirche et aI., 1980; Vyskocil
et aI., 1983).

Quantitative Data on K+ Fluxes in Humans

Data have been obtained from whole body exercise in humans during which the subjects
perceived fatigue or even became exhausted (Bergstrom et aI., 1971; Bergstrom et aI., 1973;
Sejersted et aI., 1984; Sejersted & Hallen, 1987; Medb0 & Sejersted, 1985). The highest plasma
[K+] reported was 9 mmol1- 1during intense running (Medb0 & Sejersted, 1990). An attempt to
calculate the maximal rate ofK+ loss from the leg muscles per kg wet weight showed a value
of 1.5 mmol kg ww- I min- I during running (Sejersted & Hallen, 1987) and 1.8 mmol kg ww- I
min- I during bicycling (V0llestad et aI., 1994). However, for such calculations many assump-
tions have to be made and a better experimental design is needed to obtain more reliable data.
One such set-up is the knee-extensor model in which a single muscle group (mass of
approx. 2.5 kg ww) can be studied during voluntary contractions. The K+ balance was studied
during various modes of exercise: dynamic and static from low to high intensities. The largest
total K+ loss occurred during submaximal dynamic exercise which could be maintained for
2 hr or more before exhaustion occurred, and amounted to around 20 mmol kg ww- I
(Sj0gaard et aI., 1985). However, the highest rate of net K+ loss occurred during high
intensity dynamic contractions, which the subjects could only continue for around 5 min and
was 1 mmol kg ww- I min- I (Juel et aI., 1990; Sj0gaard et aI., 1985). Interestingly, this value
is similar to that estimated above for whole body exercise. The dynamic contractions were
performed with a frequency of 1 Hz, and thus a K+ loss of 15-17 [Link] kg ww- I per contraction
was calculated. The venous effluent [K+] was around 6.3-6.8 mmol 1-1 and the arterial [K+]
was 5.5-5.8 mmol 1-1. These values are lower than those during whole body exercise and
the difference is probably due to the smaller muscle mass in the knee-extensor model,
allowing for a better clearance ofK+ into other tissues.
The largest increase in interstitial [K+] probably occurs during static contractions
where the tissue pressure impedes blood flow (Barcroft & Millen, 1939). The contraction-
induced K+ release is then trapped in the interstitial space where it accumulates throughout
the contraction. This interpretation is in agreement with the highest interstitial [K+] in man
(15 mmol 1-1) being reported during intensive isometric contractions which could be
sustained for only a few seconds (Vyskocil et aI., 1983).

Quantitative Data on K+ Fluxes in Animals

The most precise estimates of the magnitude ofK+ losses relative to muscle contrac-
tions are obtained by studying stimulation of animal muscles. The recalculation of data from
a number of studies shows K+ losses of 2.3-4.8 mmol kg ww- I min- I during continuous
subtetanic stimulation (5 Hz), 0.9-1.5 mmol kg ww- I min- I during continuous tetanic
stimulation, and 0.5-2.0 mmol kg ww- I min- I during intermittent tetanic stimulation (for
references see Sj0gaard, 1991). If calculated per contraction during intermittent stimulation,
the values ranged from 11-17 [Link] kg ww- I in 0.75-1.5 Hz trains, which comes close to the
value for humans during 1 Hz voluntary contraction (see above). A more detailed analysis
allows the K+ loss per stimulation to be calculated in the electrical stimulation studies. At
stimulation frequencies ofl-5 Hz the K+ loss per stimulus was 3.9-7.5 [Link] kg ww- I , values
which are in close agreement with those predicted from a model ofK+ efflux (Hazeyama &
Sparks, 1979). During voluntary contractions in man the actual number of firings is not
Role of Interstitial Potassium 71

known but may be estimated: each contraction period during the above intense dynamic
exercise lasted for less than 0.5 s and, with maximal firing rates of 30 Hz (Thomas et aI.,
1991), a reasonable assumption is approximately 10 discharges per contraction (Sj0gaard et
aI., 1985). On average the K+ loss per stimuli is then 1.7 /-lmol kg ww- 1, a value which is
relatively low compared to the animal data. The main explanation for this discrepancy may
be that the K+ efflux should be expressed relative to cell surface (cm2) and not to cell mass
(kg) in order to obtain data comparable between species.

Focus on Interstitial K+ Concentration

The activity-induced K+ flux from the intra- to the extracellular space of muscle
affects the concentrations in both compartments and the transmembrane concentration
gradients, which are of major functional significance. The reason for focusing on the
interstitial [K+] is that this compartment shows the largest changes. This characteristic is due
to the interstitial space having a small volume and a low resting [K+], relative to the
surrounding muscle fibers. The fibers can therefore be envisaged as a large sink ofK+ which
may readily leak into the interstitium.
We shall demonstrate that K+ fluxes may be causally related to fatigue but at the same
time point out that regulatory mechanism exist which are sufficiently potent to prevent a
K+-induced catastrophe from occurring (myotonia, paralysis). However, with extreme fatigue
(i.e., exhaustion) the preventive mechanisms may fail as both muscle excitation and force
decrease. This final stage may be considered as a safety mechanism which protects the muscle
fiber against further energy depletion and thus metabolic exhaustion and cell destruction.

FACTORS AFFECTING INTERSTITIAL [K+]

The interstitial compartment is surrounded by two sets of membranes, the sar-
colemma (plasmalemma) and the capillary membrane. Estimates of changes in the amount
of interstitial K+ must be based on the net fluxes across both membranes (Fig. 1). Further,
the interstitial [K+] is affected by water fluxes across these membranes as well as by
lymphatic drainage (Aukland & Reed, 1993). During exercise there is a net efflux of water
from the intramuscular capillaries into the interstitial space, and subsequently into the muscle
fibers as well. These water fluxes are induced primarily by changes in small-molecule
osmotic pressure, although K+ per se does not contribute significantly (Watson et aI., 1993).
It is this transfer of water which is responsible for the swelling and firmness of a muscle
belly after intense or prolonged effort. In quadriceps muscles exercised to fatigue, the
extracellular water may double per kg dry weight tissue, from 0.33 to 0.64 I kg dw- 1
(Sj0gaard, 1986). Even if no net change took place in electrolyte flux, the increased
interstitial volume would reduce [K+] by half.

Efflux of K+ Across Sarcolemma Into the Interstitial Space

The resting membrane potential (RMP) comes close to the K+ equilibrium potential,
E K , for resting values of intracellular and extracellular [K+] but does not completely balance
the concentration dependent K+ efflux. However, homeostasis is achieved by the Na+/K+-
pump which, at rest, is activated to 2-6% of its maximal capacity (Clausen et aI., 1987).
During muscle activity the increase in interstitial [K+], reported in the above studies,
is traditionally assumed to be caused by the action potential. More specifically, an outward
flux ofK+ occurs through the delayed rectifier K+ channels during the repolarization phase.
Each action potential is associated with a net loss ofK+ which, in single frog muscle fibers,
72 G. Sjegaard and A. J. McComas

INTERSTITIAL SPACE

K+ -channels

interstitial [ K~

Inactive
fibres

"Remote"
Tissues
e.g. muscle

Figure 1. Possible pathways for K+ fluxes.

was 9.6 pmol cm-3 at 21 ° C or about 5 llIIlol kg ww l (Hodgkin & Horowicz, 1959; see also above).
Calculations of unidirectional fluxes from isotopic flux have been somewhat larger: 10.7 pmol
cm-3 (or about l71lIIlol kgww- I) in the rat diaphragm at 38° C (Creese etal., 1958) and 9.4llIIlol
kg ww l in the rat soleus at 30° C (Clausen & Everts, 1987). Recent evidence has been presented
that, in addition to the delayed rectifier K+ channels, the ATP-sensitive K+ channels (K+ATP-Chan-
nels) are involved in fatigue development and recovery, and that these last channels may have dual
effects on K+ fluxes (Renaud et aI., 1994). During repetitive contractions the K+ATP-channels
increase K+ efflux, by contributing to the repolarization phase of the action potential. After the
contractions cease, however, recovery offorce is markedly delayed if the channels are blocked by
glibenclamide, a finding which would be consistent with the channels normally allowing a net
influx ofK+ to occur (see below). It is relevant to this interpretation that the K+ conductance is
increased 5-fold in single frog fibers which have been stimulated to exhaustion (Fink & Liittgau,
1976) and that this increase is prevented by glibenclamide (Castle & Haylett, 1987). Also,
although the K+ATP-channels have been reported not to open until ATP concentrations fall to 2
mmol I-I (Spruce et aI., 1985), a value well below those found in fatigue, the sensitivity of these
channels would be greatly increased by a fatigue-induced rise in [H+]. A further consideration is
whether ATP is comparlmentalized within the muscle fiber, such that the concentration close to the
sarcolemma is lower than that in the remainder of the cytosol.
Role of Interstitial Potassium 73

There are also Ca2+-activated K+ channels (K+ca-channels) in the sarcolemma, and

although some controversy still exists, recent studies have demonstrated that [Ca2+] tran-
sients occur in the muscle large enough to activate K+ efflux (Juel, 1988b). Juel has suggested
that, in fatigue, these may contribute to K+ efflux (Juel, 1988a; Juel, 1986). If so, this
mechanism could account for the K+ loss from the fibers being larger than the increase in
Na+ content, as found in a number of studies (Juel, 1986; Juel, 1988b; Hirche et aI., 1980;
Sj0gaard et aI., 1985; Sj0gaard, 1983).
Another factor which may be important for the magnitude of changes in interstitial [K+],
and one which does not appear to have been considered previously, is the motor unit architecture
of the muscle. Since the pioneering glycogen-depletion studies (Kugelberg & Edstrom, 1968;
Brandstater & Lambert, 1973), it has been accepted that the muscle fibers of any one motor unit
are normally intermingled with those of 20 or more others. One functional advantage of this
arrangement is that, in submaximal contractions, it reduces interstitial [K+] by allowing K+ to
diffuse from the vicinity ofthe active fibers into the spaces surrounding quiescent fibers. When the
intermingling of motor units is disrupted, as in the fiber type grouping of chronic denervation, it is
likely that [K+] may rise to unusually high levels in the centers of fiber clusters.

Enhanced Pumping of K+ Back into Fibers

Even in a quiescent muscle fiber, Na+/K+-pumping is required to compensate for the
passive inward flux ofNa+ and associated loss ofK+ which are consequences of the permeabil-
ity of the muscle plasmalemma. During exercise, however, the Na+/K+-pumps can increase their
outputs significantly, possibly 20-fold (Clausen et aI., 1987), particularly in the slow-twitch
fibers (Everts et aI., 1988). A potent stimulus for enhanced pumping is the rise in intracellular
[Na+] which accompanies impulse activity (Hodgkin & Horowicz, 1959; Sejersted & Hallen,
1987). A second factor is B-adrenergic stimulation of the pump, since the enhanced activity is
abolished by propranolol (Kuiack & McComas, 1992; cf. Everts et aI., 1988). It is possible that
the unmyelinated sympathetic axons which terminate directly on the muscle fibers (Barker &
Saito, 1981) are especially important sources of norepinephrine. The varicosities ofthese fibers
could release transmitter following both increased sympathetic nerve activity in exercise and
passive depolarization by interstitial [K+] and extracellular currents generated by the extrafusal
muscle fibers. In addition, there is evidence to indicate that CGRP (calcitonin gene-related
peptide), released from the motor nerve terminals, is another pump stimulant (Andersen &
Clausen, 1993). Finally, the pump would be affected by epinephrine and norepinephrine
released from the adrenal glands during exercise. Despite its enhanced activity, however, the
Na+/K+-pump does not keep pace with the K+ efflux. Some findings support the concept ofa
regulatory limitation of its activity (Medb0 & Sejersted, 1990; Hallen et aI., 1994) while others
speak in favor of an insufficient capacity of the Na+/K+-pump when stimulation frequencies
around 30 Hz are attained (Clausen & Everts, 1989).
A important consideration is that the enhanced pumping affects not only contracting
fibers but quiescent ones also. Thus, in half-maximal contractions pump-induced hyperpo-
larizations were found which were as prominent in the relaxed fibers as in the active ones
(Kuiack & McComas, 1992). There is evidence to indicate that fibers in quiescent muscles
may also participate in the extra pumping since, in non-exercising limbs, the venous [K+]
may be lower than the arterial [K+] (Sj0gaard, 1986; Lindinger et aI., 1990).

Diffusion of K+ across the Capillary Membrane

If the K+ released from the muscle fibers was distributed only to the extracellular fluid, the
plasma [K+] would be expected to reach values close to 10 mmoll- i , based on calculations ofK+
loss during several hours of intense exercise (Sj0gaard & Saltin, 1985). However, in such studies
74 G. Sjogaard and A. J. McComas

the venous effluent [K+] never attained values higher than 5.5 mmol- 1 indicating a limited increase
in interstitial [K+], and (v-a)[K+] was maintained in the order of 0.2 mmol- 1• This finding is
explained by an uptake of K+ into inactive tissues e.g., resting muscle as mentioned above
(Sj0gaardetaI., 1988; LindingeretaI., 1990). Thus, a large muscle blood flow in combination with
a relatively small exercising muscle mass (knee-extensors) allowed the transport to have sufficient
capacity to clear K+ from the interstitial space in the exercising muscle. During bicycle exercise, a
redistribution ofK+ from exercising muscle to other compartments outside the vascular bed has
also been reported (V011estad et aI., 1994). One consequence ofK+ being accumulated remote
from the exercising muscle is that it is no longer available for immediate reuptake into the
fatiguing muscle fibers. Further, the recovery process may be delayed for this reason (Sj0gaard,
1990; Bystrom & Sj0gaard, 1991).
The significance of the K+ wash-out by blood flow for muscle function is emphasized
under conditions in which blood flow through the muscle had ceased as in steady contractions
during which the intramuscular pressure exceeded the systolic blood pressure, or in the
application of a tourniquet. One indication of this is the sudden recovery of the M-wave when
a tourniquet is released from a previously exercised limb (McComas et aI., 1993), attributable
to the rapid increases in EK and E RMP • The removal ofK+ is facilitated by the hyperaemia which
accompanies exercise and which may, in part, be mediated by K+ (see below).

EFFECTS OF INCREASED INTERSTITIAL [K+]

Changes in the interstitial [K+] will have local effects, involving muscle fiber
contractility and membrane excitability. However, reflex mechanisms may play an important
role in development of fatigue (Fig. 2).

interstitial [ K +]

AP(?
Surface T -tubule
•
vasodilation in
exercising muscle
t

motor
t t
1/mre1
1 E-C neuron muscle
respiration blood flow
coupling firing ,

prom~
'\ preventing/
postponing

FATIGUE ~
Figure 2. Mechanisms playing a role in interstitial [K+] promoting or postponing fatigue.
Role oflnterstitial Potassium 75

Muscle Fiber Surface Plasmalemma

Increased concentrations of extracellular K+ reduce contractile force (e.g., Clausen
& Everts, 1991) and one mechanism for this effect is undoubtedly the depression of muscle
fiber excitability. Thus in exercised human intercostal muscle fibers, a linear relationship
has been demonstrated between the logarithm of the external [K+] and the RMP (resting
membrane potential), except at low values ofK+ (Ludin, 1970). As the RMP falls, the fiber
first enters a phase in which spontaneous action potentials may occur (myotonia); depolari-
zations greater than this cause impulse block, due to persisting inactivation ofNa+ channels.
However, impulse block at the surface membrane may be prevented in muscle fibers during
voluntary contractions due to the electrogenic component of the Na+JK+-pump.

Increased Electrogenic Component ofthe Na+JK+-Pump

As discussed above the Na+JK+-pump transports Na+ and K+ actively against their
concentration gradients and is ultimately responsible for sustaining the entire RMP, by
offsetting the ionic fluxes secondary to Na+ leakage. However, it is also directly responsible for
a small but significant fraction of the RMP. Thus, the observed RMP exceeds that calculated
from the Goldman-Hodgkin-Katz equation by some -5 mV to -8 mV (Hicks & McComas,
1989). During maximal stimulated contractions the electrogenic contribution of the Na+JK+-
pump can be as much as -30 m V and may even increase the membrane potential above its
quiescent value. The hyperpolarization can be detected with intracellular microelectrodes in
animal muscles (Hicks & McComas, 1989) and, in intact human muscles, is responsible for the
enlargement of the M-wave (muscle compound action potential) evoked by maximal motor
nerve stimulation, both during and following exercise (McComas et aI., 1994). The extent of
the hyperpolarization, as reflected in the size of the M-wave, is greater in some muscles than
others; of the muscles examined, it is especially prominent in the biceps brachii, while animal
studies suggest that fast-twitch motor units are more affected than slow-twitch (Enoka et aI.,
1992). Regardless of whether or not the M-wave is enlarged, the enhanced Na+JK+ -pumping is
capable of maintaining muscle fiber excitability for at least one minute of maximal contrac-
tions, even if the circulation is arrested (Fitch & McComas, 1985). The level of pumping, as
reflected in the potentiation of the M-wave, appears to be proportional to the frequency of
muscle excitation (Cupido, Galea & McComas, unpublished observations).
The electrogenic contribution ofthe Na/K+-pump to the RMP may cause the RMP to
become larger than EK (the K+ equilibrium potential), in which case a net K+ flux back into
the fiber will occur. Since the size of the flux will be proportional to the difference between
RMP and E K, the passive inward flux ofK+ will be especially large during exercise, because
of the large electrogenic component of the Na/K+-pump (see above). Early in exercise, when
the membrane is hyperpolarized, the flux will be increased by opening of the inward rectifier
channels (Adrian et aI., 1969).
As fatigue sets in and the RMP begins to fall, those channels which are normally
inhibited by ATP (K+ATP-channels) may start to open, again helping to keep the inward K+ flux
high in between action potentials and during recovery (see above and Renaud et aI., 1994). Of
particular importance is the sensitivity of these channels to H+, such that a fall in intracellular
pH, from 7.2 to 6.3, decreases the affinity ofthe channels for ATP 15-fold (Davies et aI., 1992).
Likewise, the K+ca-channels would be expected to become increasingly important in fatigue,
aiding in the net inward K+ flux as long as the RMP exceeds E K. When the hydrolysis of ATP
is no longer able to drive the Na+/K+-pump sufficiently, or if the latter is inhibited by oxidative
stress, there would be a net outward K+ flux; this would occur because the RMP would fall
below E K, secondary to the small permeability of the plasmalemma for Na+. It is questionable
whether or not this stage is reached during voluntary contractions; the fact that M-wave is quite
76 G. Sjogaard and A. J. McComas

well preserved in some muscles during exercise (Bigland-Ritchie et ai., 1982) would suggest
that the Na+/K+-pump can maintain the RMP with only a modest reduction.
In contrast to the possible effects ofK+ on the surface plasmalemma, it is likely that
the function in the T-tubules may be seriously jeopardized by a rising [K+] (see below).

T -Tubule Plasmalemma
The situation regarding the T-tubules may be rather different to that of the inter-fiber
space, since the very narrow lumens of the tubules restrict diffusion of K+ (Hodgkin &
Horowicz, 1959) and the K+ permeability of the T-tubular membrane is approximately twice
that of the surface membrane (Eisenberg & Gage, 1969). If the inter-fiber [K+] can rise to 10-15
mM in an isometric contraction (Vyskocil et ai., 1983), then the T-tubular [K+] must be higher
still. Further, the density of Na+/K+-pumps is lower in the T-tubules than at the fiber surface
(Fambrough et ai., 1987), so that the compensatory effects of pumping (see above) will be less
evident. These factors could combine to block impulse propagation down the T-tubules from
the surface plasmalemma, so that the myofibrils would become increasingly dependent on
electrotonic spread of the surface action potential down the low-resistance T-tubular pathway.
In unfatigued fibers of amphibians the size of the surface signal is just sufficient to activate the
innermost myofibrils in the fiber by this mechanism (Adrian et ai., 1969). However, were the
surface potential to decline, as it does later in fatigue, the excitation of the central myofibrils
would become increasingly precarious. Such a conclusion would be consistent with the results
ofWesterblad et ai. in which progressively weaker Ca2+transients occurred towards the centers
of fatigued single fibers (Westerblad et ai., 1990). In conclusion, then, the more important
effects of a raised extracellular [K+] are likely to involve excitation-contraction coupling via
the T-tubules, rather than the excitability of the surface plasmalemma.

Cardiovascular and Respiratory System

When a subject begins to exercise, a number of physiological events take place in the
cardiovascular system - the pulse quickens, blood pressure rises and, if the contractions are
intermittent, blood flow increases through the muscles (Kiens et ai., 1989; Saltin et ai., 1987).
Changes in heart rate and blood pressure can be mediated by central command as well as
reflexes, the latter of which are elicited from afferent fibers originating in the active muscles
(Goodwin et aI., 1972). These fibers are either thinly myelinated (group III) or unmyelinated
(group IV), and their free nerve endings must be responsive to some property of the contracting
muscle - either deformation of the investing connective tissue or a change in the metabolic
milieu of the muscle fibers (Kniftki et aI., 1981). K+ is one of several candidates for exciting
the metabolically-sensitive fibers, and has been shown capable of activating group III and group
IV muscle afferents, albeit transiently (Rybicki et ai., 1985). Evidence in favor of such a role
for K+ in exercise includes the close parallelism between the relative magnitudes and time-
courses of blood pressure and [K+] respectively (Fallentin et ai., 1992; Saltin et ai., 1981).
Similar arguments to the above suggest that there is also a casual relationship between
[K+] and increased blood flow. For example, if Doppler ultrasonography is used to measure
arterial blood velocity, and hence flow, then the changes in [K+] and blood flow vary together
during exercise (Kiens et ai., 1989) although they may be dissociated in the recovery period
(Jensen et aI., 1993). In addition, the arterial perfusion ofhyperkalaemic solutions produces
relaxation of the arterial smooth muscle (Dawes, 1941). These observations do not, of course,
deny the possibility that other factors may contribute to the hyperaemia of exercise (e.g.,
reduced p02, ATP, inorganic phosphate, and prostaglandins).
A third effect ofK+ on the cardiovascular system involves the heart. It is well known
that intravascular injections ofK+ can produce cardiac arrest and use of this is made in cardiac
Role of Interstitial Potassium 77

surgery. In intensive exercise, such as treadmill running or cycle ergometry, arterial [K+]
concentrations as high as 8.2 mM have been found (Medb0 & Sejersted, 1990) and would
be expected to affect cardiac excitability, as indicated by the ECG. In patients with
hyperkalaemia, plasma [K+] values of 8 mmoll or greater produce atrial depression and
o
]

AV block, while 10 mmoll [K+] can cause ventricular arrhythmias, including fibrillation
o
]

(Lipman et aI., 1984). Taken together, these observations draw attention to the dangers of
sudden intensive exercise in untrained subjects and may explain some cases of exercise-in-
duced cardiac arrest, in the absence of occlusive arteriopathy.
Finally, it is known that K+ stimulates the respiratory system also, increasing the
ventilation rate (see, for example, Paterson et aI., 1990). This effect ofK+ depends on a reflex
in which the afferent fibers are excited by K+ -sensitive cells in the carotid body (Band &
Linton, 1986).

Motor Control
There is now circumstantial evidence that the reduction in motoneuron firing rate,
which occurs during fatiguing contractions, may have a reflex origin (see Bigland-Ritchie et
aI., Chapter 27; Bigland-Ritchie et aI., 1986; Woods et aI., 1987). Although the afferent limb of
the reflex is thought to involve Group III and Group IV axons, the identities of the exciting
stimuli remain uncertain; it is quite likely, however, that interstitial [K+] is one factor, possibly
activating both metaboreceptor and nociceptor fibers. This reduction in motoneuron activation
is a process for blunting the rise in interstitial [K+], such that the discharge rate declines and
there are fewer releases ofK+ per unit time, which will prevent or postpone fatigue. In contrast
to this an hypothesis has been suggested (Johansson & Sojka, 1991) that increased interstitial
[K+] may elicit a vicious circle via the gamma-loop. Evidence has been presented that activity
in the group III but also group IV sensory afferents may have excitatory effects, especially on
static but also on dynamic gamma-motoneurons. These effects are strong enough to increase
firing in the muscle spindle afferents, which in tum will raise the activation level in the pool of
alpha-motoneurons projecting to the extrafusal muscle fibers. This reflex mediated muscle
activity leads to further K+ fluxes and increased metabolism.
For as long as interstitial [K+] is elevated an activity-induced increased cytosolic [Ca2+]
will persist according to the above hypothesis. This will be enhanced by the interstitial [K+] per
se inducing a Ca2+ uptake into the muscle fiber (Everts et aI., 1993; Barnes, 1993). Elevated
cytosolic [Ca2+] will increase the respiration of the cell causing an even faster metabolic
exhaustion (Barnes, 1993). Most critical to the muscle cell is that a maintained increased
intracellular [Ca2+] will induce an overload of the mitochondria causing further disruption of
metabolism, and will also breakdown the cell membrane (Jackson et aI., 1984). To prevent this
catastrophe the inactivation ofthe cell membrane comprises a perfect mechanism. Thus muscle
fatigue and safety mechanisms may be interrelated via interstitial [K+].

ACKNOWLEDGMENTS

The authors acknowledge the research support of the Medical Research Council of
Denmark (G.s.), and the Natural Sciences and Engineering Research Council (NSERC) of
Canada ([Link].). Attendance of the authors at the 1994 Bigland-Ritchie conference was
supported, in part, by the National Institute of Occupational Health, Denmark (G.s.), NSERC
([Link].), the United States Public Health Service (National Center for Medical Rehabili-
tation Research, National Institute of Child Health and Human Development;G.s.), and the
Muscular Dystrophy Association (USA; G.s.).
78 G. Sjegaard and A. J. McComas

REFERENCES
Adrian RH, Costantin LL & Peachey LD (1969). Radial spread of contraction in frog muscle fibres. Journal
ofPhysiology (London) 204, 231-257.
Andersen SLY & Clausen T (1993). Calcitonin gene-related peptide stimulates active Na+-K+ transport in rat
soleus muscle. American Journal ofPhysiology 264, C419-C429.
Aukland K & Reed RK (1993). Interstitial-lymphatic mechanisms in the control of extracellular fluid volume.
PhYSiological Reviews 73, 1-77.
Band DM & Linton RAF (1986). The effect of potassium on carotid body chemoreceptor discharge in the
anaesthetized cat. Journal of PhYSiology (London) 381, 39-47.
Barcroft H & Millen JLE (1939). The blood flow through muscle during sustained contraction. Journal of
Physiology (London) 97, 17-31.
Barker D & Saito M (1981). Autonomic innervation of receptors and muscle fibres in cat skeletal muscle.
Proceedings of the Royal Society ofLondon, Series B-Biological Sciences B212, 317-332.
Barnes WS (1993). Effects ofCa2+-channel drugs on K+-induced respiration in skeletal muscle. Medicine and
Science in Sports and Exercise 25, 473-478.
Bergstrom J, Guarnieri G & Hultman E (1971). Carbohydrate metabolism and electrolyte changes in human
muscle tissue during heavy work. Journal ofApplied Physiology 30, 122-125.
Bergstrom J, Guarnieri G & Hultman E (1973). Changes in muscle water and electrolytes during exercise. In:
Keul J. (ed.), Limiting Factors of Physical Performance, pp. 173-178. Stuttgart: Georg Thieme.
Bigland-Ritchie BR, Dawson NJ, Johansson RS & Lippold OCJ (1986). Reflex origin for the slowing of
motoneurone firing rates in fatigue of human voluntary contractions. Journal ofPhysiology (London)
379,451-459.
Bigland-Ritchie B, Kukulka CG, Lippold OCJ & Woods 11 (1982). The absence of neuromuscular transmission
failure in sustained maximal voluntary contractions. Journal of Physiology (London) 330, 265-278.
Brandstater ME & Lambert EH (1973). Motor unit anatomy. In Desmedt JE (ed.), New Developments in
Electromyography and Clinical Neurophysiology, pp. 14-22. Basel: Karger
Bystrom S & Sjegaard G (1991). Potassium homeostasis during and following exhaustive submaximal static
handgrip contractions. Acta Physiologica Scandinavica 142, 59-66.
Castle NA & Haylett DG (1987). Effect of channel blockers on potassium efflux from metabolically exhausted
frog skeletal muscle. Journal ofPhysiology (London) 383, 31-43.
Clausen T & Everts ME (1987). Is the Na, K-pump capacity in skeletal muscle inadequate during sustained
work? In: Proceedings of the Vth International Conference on Na, K-ATPase. Arhus, Denmark. June
14-19, New York: Alan Liss Inc.
Clausen T & Everts ME (1989). Regulation of the Na, K-pump in skeletal muscle. Kidney International 35,
1-13.
Clausen T & Everts ME (1991). K+-induced inhibition of contractile force in rat skeletal muscle: role of active
Na+-K+ transport. American Journal ofPhysiology 261, C799-C807.
Clausen T, Everts ME & Kjeldsen K (1987). Quantification of the maximum capacity for active sodium-po-
tassium transport in rat skeletal muscle. Journal ofPhysiology (London) 388,163-181.
Creese R, Hashish S & Scholes NW (1958). Potassium movements in contracting diaphragm muscle. Journal
of Physiology (London) 143, 307-324.
Crone C, Frekjrer-Jensen J, Friedman 11 & Christensen 0 (1978). The permeability of single capillaries to
potassium ions. Journal of General Physiology 71, 195-220.
Davies NW, Standen NB & Stanfield PR (1992). The effect of intracellular pH on ATP-dependent potassium
channels of frog skeletal muscle. Journal of Physiology (London) 445, 549-568.
Dawes GS (1941). The vaso-dilator action of potassium. Journal of Physiology (London) 99, 224-238.
De Lanne R, Barnes JR & Brouha L. (1959). Changes in osmotic pressure and ionic concentrations of plasma
during muscular work and recovery. Journal ofApplied Physiology 14, 804-808.
Eisenberg RS & Gage PW (1969). Ionic conductances of the surface and transverse tubular membranes of frog
sartorius fibres. Journal of General Physiology 53,279-297.
Enoka RM, Trayanova N, Laouris Y, Bevan L, Reinking RM & Stuart DG (1992). Fatigue-related changes in
motor unit action potentials of adult cats. Muscle & Nerve 14, 138-150. ,
Everts ME, Lemo T & Clausen T (1993). Changes in K+, Na+ and calcium contents during in vivo stimulation
of rat skeletal muscle. Acta Physiologica Scandinavica 147, 357-368.
Everts ME, Retterstel K & Clausen T (1988). Effects of adrenaline on excitation-induced stimulation of the
sodium-potassium pump in rat skeletal muscle. Acta Physiologica Scandinavica 134, 189-198.
Fallentin N, Jensen BR, Bystrom S & Sjegaard G (1992). Role of potassium in the reflex regulation of blood
pressure during static exercise in the human. Journal ofPhysiology (London) 451, 643-651.
Role of Interstitial Potassium 79

Fambrough DM, Wolitzky BA, Tamkim MM & Takeyasu K (1987). Regulation of the sodium pump in
excitable cells. Kidney International 32 (suppl. 23), S97-S112.
Fenn WO (1937). Loss of potassium in voluntary contraction. American Journal ofPhysiology 120, 675-680.
Fenn WO (1938). Factors affecting the loss of potassium from stimulated muscles. American Journal of
Physiology 124, 213-229.
Fenn WO & Cobb DM (1936). Electrolyte changes in muscle during activity. American Journal ofPhysiology
115, 345-356.
Fink R & Liittgau HC (1976). An evaluation of the membrane constants and the potassium conductance in
metabolically exhausted muscle fibres. Journal of PhYSiology (London) 263, 215-238.
Fitch S & McComas A (1985). Influence of human muscle length on fatigue. Journal ofPhysiology (London)
362,205-213.
Goodwin GM, McCloskey DI & Mitchell JH (1972). Cardiovascular and respiratory responses to changes in
central command during isometric exercise at constant muscle tension. Journal of PhYSiology
(London) 226, 173-190.
Hallen J, Gullestad L & Sejersted OM (1994). K+ shifts of skeletal muscle during stepwise bicycle exercise
with and without B-adrenoceptor blockade. Journal ofPhysiology (London) 477, 149-159.
Hazeyama Y & Sparks HV (1979). A model of potassium ion eftlux during exercise of skeletal muscle.
American Journal of Physiology 236, R83-R90.
Hicks A & McComas AJ (1989). Increased sodium pump activity following repetitive stimulation of rat soleus
muscles. Journal of Physiology (London) 414, 337-349.
Hirche H, Schumacher E & Hagemann H (1980). Extracellular K+ concentration and K+ balance of the
gastrocnemius muscle of the dog during exercise. Pjlugers Archiv 387,231-237.
Hodgkin AL & Horowicz P (1959). The influence of potassium and chloride ions on the membrane potential
of single muscle fibres. Journal of Physiology (London) 148, 127-160.
Jackson MJ, Jones DA & Edwards RHT (1984). Experimental skeletal muscle damage: The nature of the
calcium-activated degenerative processes. European Journal of Clinical Investigation 14, 369-374.
Jensen BR, Fallentin N, Sjl2lgaard G & Bystrom S (1993). Plasma potassium concentration and Doppler blood
flow during and following submaximal handgrip contractions. Acta Physiologica Scandinavica 147,
203-211.
Johansson H & Sojka P (1991). Pathophysiological mechanisms involved in genesis and spread of muscular
tension in occupational muscle pain and in chronic musculoskeletal pain syndromes: A hypothesis.
Medical Hypotheses 35, 196-203.
Juel C (1986). Potassium and sodium shifts during in vitro isometric muscle contraction, and the time course
of the ion-gradient recovery. Pjlugers Archiv 406, 458-463.
Juel C (1988a). Is a Ca2+-dependent K+ channel involved in the K+ loss from active muscles? Acta Physiologica
Scandinavica 132, P26.
Juel C. (1988b). The effect ofbeta2-adrenoceptor activation on ion-shifts and fatigue in mouse soleus muscles
stimmulated in vitro. Acta Physi%gica Scandinavica 134, 209-216.
Juel C, Bangsbo J, Graham T & Saltin B (1990). Lactate and potassium fluxes from human skeletal muscle
during and after intense, dynamic, knee extensor exercise. Acta Physiologica Scandinavica 140,
147-159.
Kiens B, Saltin B, Walll2le L & Wesche J (1989). Temporal relationship between blood flow changes and release
of ions and metabolites from muscle upon single weak contractions. Acta Physiologica Scandinavica
136,551-559.
Kniffki KD, Mense S & Schmidt RF (1981). Muscle receptors with fine afferent fibres which may evoke
circulatory reflexes. Circulation Research 48 (suppl I), 25-31.
Kugelberg E & Edstrom L (1968). Differential histochemical effects of muscle contraction on phosphorylase
and glycogen in various types of fibres: relation to fatigue. Journal ofNeurology, Neurosurgery and
Psychiatry 31, 415-423.
Kuiack S & McComas AJ (1992). Transient hyperpolarization of non-contracting muscle fibres in anaesthe-
tized rats. Journal ofPhysiology (London) 454, 609-618.
Lindinger MI, Heigenhauser GJF, McKelvie RS & Jones NL (1990). Role of nonworking muscle on blood
metabolites and ions with intense intennittent exercise. American Journal of Physiology 258,
RI486-RI494.
Lipman BS, Dunn M & Massie E (1984). Clinical Electrocardiography, pp. 268-271. Chicago: Year Book
Medical Publishers Inc.
Ludin HP (1970). Microelectrode study of dystrophic human skeletal muscle. European Neurology 3, 116-121.
McComas AJ, Galea V & Einhorn RW (1994). Pseudofacilitation: a misleading tenn. Muscle & Nerve 17,
599-607.
80 G. Sjogaard and A. J. McComas

McComas AJ, Galea V, EinhornRW, Hicks AL& Kuiack S (1993). The role of the Na+, K+-pump in delaying
muscle fatigue. In: Sargeant AJ, Kernell D (eds.), Neuromuscular Fatigue, pp. 35-43. Amsterdam:
Royal Netherlands Academy of Arts and Sciences.
Medbo JI & Sejersted OM (1985). Acid-base and electrolyte balance after exhausting exercise in endurance-
trained and sprint-trained subjects. Acta Physiologica Scandinavica 125, 97-109.
Medbo JI & Sejersted OM (1990). Plasma potassium changes with high intensity exercise. Journal of
Physiology (London) 421, 105-122.
Paterson DJ, Friedland JS, Bascom DA, Clement ID, Cunningham DA, Painter R & Robbins PA (I 990).
Changes in arterial K+ and ventilation during exercise in normal subjects and subjects with McArdle's
syndrome. Journal of Physiology (London) 429, 339-348.
Renaud JM, Light PE & Comtois AS (1994). The effect of glibenclamide on frog skeletal muscle: evidence
for K+(ATP) channel activation during fatigue. Abstracts. Ontario Exercise Physiology Meeting,
(Toronto, Canada, February 1994).
Rybicki KJ, Waldrop TG & Kaufman MP (1985). Increasing gracilis muscle interstitial potassium concentra-
tions stimulate group III and IV afferents. Journal ofApplied Physiology 58, 936-941.
Saltin B, Sjogaard G, Gaffney FA & Rowell LB (1981). Potassium, lactate, and water fluxes in human
quadriceps muscle during static contractions. Circulation Research 48 (suppl I), 1-18-1-24.
Saltin B, Sjogaard G, Strange S & Juel C (1987). Redistribution ofK+ in the human body during muscular
exercise; its role to maintain whole body homeostasis. In: Shiraki K, Yousef MK (eds.), Man in
Stressful Environments. Thermal and Work Physiology, pp. 247-267. Springfield, IL: CC Thomas.
Sejersted OM & Hallen J (1987). Na, K homeostasis of skeletal muscle during activation. In: Marconnet P,
Komi P (eds.), Muscle Function in Exercise and Training. Medicine and Science in Sports, vol 26,
pp. 1-11. Basel: Karger.
Sejersted OM, Medbo n, Orheim A & Hermansen L (1984). Relationship between acid-base status and
electrolyte balance after maximal work of short duration. In: Marconnet P, Poortmans JR, Hermansen
L. (eds.), Physiological Chemistry of Training and Detraining. Medicine and Science in Sports, vol
17, pp. 40-55. Basel: Karger.
Sjogaard G (1983). Electrolytes in slow and fast muscle fibers of humans at rest and with dynamic exercise.
American Journal of PhYSiology 245, R25-R31.
Sjogaard G (I986). Water and electrolyte fluxes during exercise and their relation to muscle fatigue. Acta
Physiologica Scandinavica 128 (suppI556), 129-136.
Sjogaard G (1990). Exercise-induced muscle fatigue: The significance of potassium. Acta Physiologica
Scandinavica 140 (supp!. 593), 1-64.
Sjogaard G (1991). Role of exercise-induced potassium fluxes underlying muscle fatigue: a brief review.
Canadian Journal of Physiology and Pharmacology 69, 238-245.
Sjogaard G, Adams RP & Saltin B (1985). Water and ion shifts in skeletal muscle of humans with intense
dynamic knee extension. American Journal of Physiology 248, RI90-RI96.
Sjogaard G & Saltin B (1985). Potassium redistribution within the body during exercise. Clinical Physiology
5 (suppl 4), 150.
Sjogaard G, Savard G & Juel C (1988). Muscle blood flow during isometric activity and its relation to muscle
fatigue. European Journal ofApplied Physiology and Occupational Physiology 57,327-335.
Spruce AE, Standen NB & Stanfield PR (1985). Voltage-dependent ATP-sensitive potassium channels of
skeletal muscle membrane. Nature 316, 736-738.
Thomas CK, Bigland-Ritchie B & Johansson RS (1991). Force-frequency relationships of human thenar motor
units. Journal of Neurophysiology 65, 1509-1516.
Vyskocil F, Hnik P, Rehfeldt H, Vejsada R & Ujec E (1983). The measurement ofKe+ concentration changes
in human muscles during volitional contractions. Pjliigers Archiv 399,235-237.
Vollestad NK, Hallen J & Sejersted OM (1994). Effect of exercise intensity on potassium balance in muscle
and blood of man. Journal of Physiology (London) 475, 359-368.
Watson PD, Gamer RP & Ward DS (1993). Water uptake in stimulated cat skeletal muscle. American Journal
of Physiology 264, R790-R 796.
Westerblad H, Lee JA, Lamb AG, Bolsover SR & Allen DG (1990). Spatial gradients of intracellular calcium
in skeletal muscle during fatigue. Pjliigers Archiv 415, 734-740.
Woods JJ, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal of Neurophysiology 58,125-137.
SECTION II
Fatigue at the Neuromuscular Junction

In this section, axonal branch points, the neuromuscular junction, and the sarcolemma
are considered as possible limiting sites in fatigue. Usually, the safety margins for propaga-
tion and transmission are sufficiently high at all three sites to ensure that adequate excitation
reaches the t-tubules. Some controversy exists about the extent to which significant failure
of neuromuscular transmission occurs during normal human muscle fatigue However, no
doubt exists about the involvement of this site in pathological conditions such as myasthenia
gravis. Three approaches to the study of neuromuscular transmission during fatiguing
contractions are presented.
In Chapter 5, Sieck and Prakash review both the pre- and postsynaptic mechanisms
involved in normal neuromuscular transmission. In addition, they present data on the
adaptations at these sites. Presynaptically, neuromuscular transmission failure may result
from the inability of the axonal action potential to propagate into all its terminal branches,
and from reductions in the frequency and size of quantal release. Failure of branch-point
propagation is less well studied experimentally, but there is indirect evidence that it occurs,
with a rise in interstitial potassium as a contributing cause. Postsynaptically, failure may
occur because of a reduction in the number of acetylcholine receptors or because of an
impairment of sarcolemmal excitability.
The likelihood that the sarcolemmal action potential fails to activate the muscle fiber
fully is tackled in Chapter 6 (Fuglevand). It is argued that for mammalian fast-twitch fibers
in particular, the reduction in action potential amplitude during fatigue may mean that
membrane depolarization is insufficient for full activation and that this mechanism would
contribute to the decline in force. Given the known changes in action potential size (and
area) produced by nerve stimulation in some studies of human fatigue, this remains a
possibility. However, because it is difficult to maintain a supramaximal stimulus intensity
for the requisite motor axons in human studies during vigorous exercise, a reduction in
evoked action potential amplitude must be considered critically.
Chapter 7 (Trontelj & Stalberg) introduces the technique of single-fiber EMG which
has been a powerful clinical tool in diagnosis of disorders at the neuromuscular junction.
Rather than studying only motor units that can be recruited voluntarily at low forces, it is
now possible to examine the responses of a wide range of single muscle fibers to axonal
stimulation. This has increased the scope of the method as it permits the behavior of muscle
fibers to be studied over a range of stimulus frequencies. Most end-plates in normal subjects
have relatively stable jitter (i.e. the time difference between muscle action potentials in fibers
from the same motor unit), although the jitter increases in some at stimulus frequencies
greater than 15-20 Hz. This method allows the effective safety factor at the neuromuscular
82 Fatigue at the Neuromuscular Junction

junction to be assessed and compared with that observed in pathological conditions. Under
ischemic conditions, which occur during strong isometric contractions, jitter increases and,
after 2000-4000 discharges some end-plates will eventually develop a conduction block and
fatigue will increase.
These chapters highlight different aspects of neuromuscular transmission: force
production can be jeopardized at one or more of the steps required for effective activation
of muscle fibers. However, under most physiological conditions, these steps are designed
with comfortable safely margins. Under some in vivo circumstances, perhaps for some motor
units when motoneuron discharge rates are high and blood flow is impaired, fatigue might
be enhanced through failure of neuromuscular transmission. Such circumstances are inter-
esting because they provide insight into the design limits of the neuromuscular junction.
 5

FATIGUE AT THE NEUROMUSCULAR

JUNCTION
Branch Point vs. Presynaptic vs. Postsynaptic Mechanisms

G. C. Sieck and Y. S. Prakash

Departments of Anesthesiology and Physiology and Biophysics

Mayo Clinic and Foundation
Rochester, Minnesota 55905

ABSTRACT

There are several pre- and postsynaptic sites where neuromuscular transmission failure
(NTF) can occur, leading to peripheral muscle fatigue. Presynaptic sites of NTF include:
axonal branch point conduction block; a failure of excitation-secretion coupling at the
presynaptic terminal; reductions in quantal release of ACh; and reductions in quantal size.
Postsynaptic sites of NTF include: cholinergic receptor desensitization; and reduced sar-
colemmal excitability. Susceptibility to NTF increases with stimulation frequency and is
most prevalent in fatigable fast-twitch motor units. In addition, susceptibility to NTF varies
with age and with conditions of altered use.

INTRODUCTION
Muscle fatigue can be either central (failure of excitation of motoneurons) or peripheral
(failure in the transmission of the neural signal, or a failure of the muscle to respond to neural
excitation) in origin (Bigland-Ritchie et aI., 1978, 1982). At the peripheral level, several pre- and
postsynaptic mechanisms and sites are potentially implicated, including: a failure of action
potential propagation along the axon; inadequate presynaptic release of acetylcholine (ACh);
insufficient depolarization of the postsynaptic membrane; failure of action potential propagation
along the sarcolemma; and, failure of excitation-contraction coupling. This review addresses
possible mechanisms of peripheral fatigue that are proximal to a failure of excitation-contraction
coupling and that can be generally categorized as neuromuscular transmission failure (NTF).

NEUROMUSCULAR TRANSMISSION FAILURE

The fact that NTF can occur under certain conditions is indisputable, but its contri-
bution to muscle fatigue during normal motor behaviors remains controversial (e.g., Merton,

83
84 G. C. Sieck and Y. S. Prakash

1954; Stephens & Taylor, 1972; Bigland-Ritchie et al., 1982). NTF failure has been detected
using two gener~l techniques: I) comparison of the forces elicited by indirect nerve versus
direct muscle stimulation; and 2) assessment of evoked endplate potentials (EPPS) or
compound muscle action potentials (e.g., M waves). Both measurements have been used
extensively to assess the incidence ofNTF.

Comparison of Nerve versus Muscle Stimulation

The relative contribution of NTF to muscle fatigue can be estimated by comparing
force loss during nerve stimulation to that elicited by direct muscle stimulation (thus
bypassing the neural sites offailure) (Fig. I). Using this procedure, the relative contributions
ofNTF and contractile failure to muscle fatigue can be partitioned, and estimated based on
the following formula (Aldrich et al., 1986):

NTF = (F - MF)/(l - MF)

where NTF is the relative contribution of NTF to fatigue, F is force loss during repetitive
nerve stimulation and MF is force loss during direct muscle stimulation.
The relative contribution of NTF to muscle fatigue depends on the frequency of
stimulation (Thesleff, 1959; Aldrich et al., 1986; Kuei et al., 1990; Johnson & Sieck, 1993),
and this frequency dependence ofNTF may explain the decrement in motoneuron discharge
rate that occurs with sustained voluntary contractions (Bigland-Ritchie et al., 1979; Bigland-
Ritchie & Lippold, 1979; Bigland-Ritchie, 1984; Woods et al., 1987). By decreasing
motoneuron discharge rate, susceptibility to NTF can be reduced. Such a reduction in
motoneuron discharge rate may be reflexive, emanating from small-diameter afferents within
the muscle (Woods et al., 1987), and/or intrinsic to the motoneuron (Kernell, 1965; Kernell
& Monster, 1982a; 1982b)(for further discussion of this issue, see Binder-Macleod, Chapters
16; Windhorst & Boorman, Chapter 17; Hagbarth & Macefield, Chapter 18; and Garland &
Kaufman, Chapter 19).
Fiber type composition ofa muscle is also an important determinant of the extent of
NTF. Kugelberg and Lindegren (1979) suggested that muscles with a greater proportion of
fast-twitch, glycolytic fibers may be more susceptible to NTF. In support, we used an in vitro
nerve-diaphragm muscle preparation and the technique of glycogen depletion to demonstrate
that type IIblIIx fibers were the most susceptible to NTF, especially at higher rates of

rndirect erve Stimulation Direct Muscle Stimulation

/1

t...I.."w"."'&", ,wv,JL"rw"'t'ltr'+ww,,,AtI!M'Wtl'u'M'"MMIIILII!MMlIt"~'HMWMl
135 150 165 180 I!Is 210 225 240 255 270 285 300
Time (5)

Figure 1. The relative contribution ofNTF to muscle fatigue can be estimated by periodically superimposing
direct muscle stimulation (ms) onto repetitive nerve stimulation (ns). In this case, the adult rat diaphragm was
activated by stimulation of the phrenic nerve at 40 Hz in 330 ms duration trains repeated each s for 5 min.
Every IS s, direct muscle stimulation was superimposed. The difference between the forces generated by
indirect nerve vs. direct muscle stimulation provides an index ofNTF.
Fatigue at the Neuromuscular Junction 8S

Table 1. Assessment ofNTF in type-identified fibers of rat diaphragm by

comparison of the extent of glycogen depletion
10 Hz 7S Hz
Nerve Muscle Nerve Muscle
% # % # % # % #
[Glycogen] Depleted [Glycogen] Depleted [Glycogen] Depleted [Glycogen] Depleted
.J.. fibers .J.. fibers .J.. fibers .J.. fibers
38±S 86±S S3±6 9S±2 38±6 80±9 67±St 98±1
IIa 46±S 84±S S9±6 89±7 36±4 t 67±10t 67±S 96±2
lIb Sl±6* 9S±3*t 70±4*t 97±3 30±6 t SO±IO*t 80±4*tt 99±l
%[Glycogen].J..: Decrement in glycogen (% change in optical density from control); #Depleted
Fibers: Proportion (% of total) of "depleted" fibers with glycogen levels 2 SD lower than control
values. All values are means ± SEM. * indicates significant difference (p< [Link]) from type I fibers;
t indicates significant (p<[Link]) difference from type IIa fibers; and t indicates significant
difference (p<[Link]) from 10 Hz stimulation.

stimulation (Johnson & Sieck, 1993). The underlying premise was that during repetitive
nerve stimulation, those fibers that are more susceptible to NTF will not be activated and
therefore, will not display the depletion of glycogen stores resulting from exhaustive
activation. During repetitive direct muscle stimulation, type IIblIIx fibers displayed substan-
tial and rapid depletion of their glycogen stores. In contrast, during repetitive nerve
stimulation, type IIblIIx fibers showed far less glycogen depletion. Differences in the
depletion of glycogen stores between direct muscle and indirect nerve stimulation for type
I and IIa fibers were much less pronounced (Table 1).
Susceptibility to NTF can also adapt under a variety of conditions of postnatal
development and altered muscle use. For example, using an in vitro nerve-diaphragm muscle
preparation in the rat, we demonstrated that adaptations in the extent ofNTF occur during
early postnatal development (Fournier et aI., 1991). At all stimulation frequencies, NTF in
the neonatal diaphragm muscle is much more pronounced than in the adult. As a result, in
the neonate, diaphragm muscle fatigue induced by repetitive nerve stimulation (at 40 Hz) is
more pronounced than that induced by direct muscle stimulation (Fig. 2). Using a similar
procedure, Feldman and colleagues (1991) also found that the neonatal rat diaphragm is more
susceptible to NTF than the adult.

Figure 2. Fatigue of the rat diaphragm was

o Direct Muscle Stirn
induced by either repetitive direct muscle * _ Indirect Nerve Stirn
simulation or indirect phrenic nerve stimu-
lation at 40 Hz in 330 ms duration trains
*
repeated each second for a 2-min period. A
fatigue index was calculated as a ratio of the
force generated after 2 min to the initial
force. In neonates (0-7 days), fatigue in-
duced by nerve stimulation was far more
pronounced than that induced by direct
muscle stimulation, indicating a greater
contribution of NTF. This difference was
less striking in the 21 day old and absent in
the adult. (* indicates significant difference
(p<[Link]) between nerve and muscle stimu- o~ __
lation). 0-7 Days
86 G. C. Sieck and Y. S. Prakash

100 .
c
o
'in -
(J) Q)
90
.- :>
E o>
f/):'::
Figure 3. The relative contribution ofNTF to total fatigue
OCTL
. .
c.j!! of the rat diaphragm muscle was estimated using the tech-
!!!- 80
OSI
I-~ nique shown in Fig. I. Following 2 weeks of inactivity
... 0

r
<11- . nx induced by cervical (C 2) spinal cord hemisection, the rela-
:;~
tive contribution of NTF decreased. In contrast, after 2

~~
u~ 70 '
Q) f'J LOAD

i
(J)
:> ... I. weeks of inactivity induced by TTX blockade, NTF in-
E.2
o .- creased. In both cases, the intact contralateral hemidia-
I
... <11
::::lIJ.. 60 phragm increased its activity by 50% (compensatory
Q)
Z loading), and after 2 weeks, the relative contribution of

50
40 Hz 75 Hz
~ NTF to total fatigue decreased on the loaded side. (*
indicates significant difference (p<0.05) from control val-
ues).

Recently, we have also observed that the extent of NTF in the rat diaphragm
muscle is affected by prolonged periods of altered use (Miyata et aI., 1994). For example,
following two weeks of inactivity of the right hemidiaphragm, induced by cervical spinal
cord transection to block descending inspiratory drive to the phrenic motoneuron pool
(spinal isolation), the relative contribution of NTF to diaphragm fatigue induced by
repetitive nerve stimulation at 40 and 75 Hz was diminished by 22% and 15%, respectively
(Fig. 3).
In another model of diaphragm hemiparalysis, where axonal propagation of action
potentials was blocked by continuous superfusion of the phrenic nerve with tetrodotoxin
(TTX), the extent ofNTF was found to be markedly increased (Fig. 3). The underlying basis
for these discrepant effects of prolonged inactivity on the susceptibility of the diaphragm
muscle to NTF remains unclear but may relate to the effects of mismatching motoneuronal
and muscle activities. With TTX blockade of axonal propagation, the diaphragm muscle is
paralyzed, but phrenic motoneuronal activity increases by approximately 50%. It is possible
that, as a result, normal neurotrophic influences emanating from the motoneuron are altered
with TTX blockade of axonal propagation. In contrast, with spinal isolation, diaphragm
muscle paralysis is the result of inactivation of phrenic motoneurons. In this case, normal
neurotrophic influences may persist because motoneuronal and muscle activities remain
matched. This concept is supported by the fact that the muscle contractile and morphological
adaptations that occur with TTX blockade are similar to those associated with phrenic
denervation. In both cases, type IIb and IIx muscle fibers atrophy, specific force is reduced,
and maximum unloaded shortening velocity slows. In contrast, only minimal muscle
adaptations are found following two weeks of inactivity induced by spinal isolation. Since
hemidiaphragm paralysis and its consequent mechanical effects are present in both the TTX
and spinal isolation models, yet these two conditions are associated with divergent effects
on susceptibility to NTF, it is doubtful that muscle inactivity per se causes the improvement
in neuromuscular transmission following spinal isolation. This is supported by the fact that
the compensatory overloading of the intact, contralateral hemidiaphragm was also associated
with improved neuromuscular transmission (Fig. 3). It is more likely that an alteration in
muscle use, whether an increase or a decrease in activity, is associated with an increase in
neurotrophic influences that promote improvements in synaptic efficacy. Putative neurotro-
phic factors that might be involved in such adaptive responses would include calcitonin gene
related peptide (CGRP) (Tsujimoto & Kuno, 1988) and/or ciliary neurotrophic factor
(CNTF) (Helgren et aI., 1994).
Fatigue at the Neuromuscular Junction 87

Assessment of Evoked Muscle Action Potentials - M Wave

The contribution of NTF to fatigue induced by voluntary contractions remains
controversial. Merton (1954) utilized an assessment of evoked M waves as a means of
estimating the extent of NTF during maximal voluntary contractions. He observed no
reduction in M wave amplitude of the adductor pollicis muscle during more than 3 min of
maximum voluntary contractions, in spite of a substantial loss in force. From these results,
he concluded that NTF does not contribute to fatigue during maximal voluntary contractions.
Subsequently, Bigland-Ritchie and colleagues (Bigland-Ritchie & Lippold, 1979; Bigland-
Ritchie et al., 1979, 1982; Bellemare & Bigland-Ritchie, 1987; McKenzie et al., 1992) have
conducted a number of similar studies assessing changes in M wave amplitude during
voluntary contractions, and have concluded that NTF does not contribute to fatigue induced
by sustained maximal voluntary contractions of human muscles. In contrast, other investi-
gators have reported that M wave amplitude decreases during fatiguing maximal voluntary
contractions of the human adductor pollicis (Naess & Storm-Mathisen, 1955) and first dorsal
interosseus (Stephens & Taylor, 1972) muscles, suggesting to them that significant NTF does
occur and does contribute to fatigue. Grob (1961) also found that repeated electrical
stimulation of the ulnar nerve at frequencies greater than 25 Hz resulted in a reduction in M
wave amplitude which paralleled the reduction in tetanic force ofthe first dorsal interosseus
muscle. During inspiratory loaded breathing in unanesthetized sheep, Bazzy and Donnelly
(1993) have also reported a reduction in M wave amplitude that paralleled force decline,
suggesting a significant contribution of NTF to diaphragm fatigue. The reasons for these
discrepant results are unclear but may relate to the experimental conditions and the type of
voluntary or involuntary behaviors studied. However, it must be emphasized that changes
in the M wave do not necessarily indicate a failure of neuromuscular transmission as they
may be induced by changes in the sarcolemmal action potential (see Sj0gaard & McComas,
Chapter 4).
Changes in evoked muscle action potentials have also been used to assess suscepti-
bility of single muscle fibers and motor units to NTF. At the single fiber level, NTF can
present itself as a failure to evoke an EPP in response to nerve stimulation and/or a reduction
in EPP amplitude (Fournier et aI., 1991). Using an in vitro nerve-diaphragm muscle
preparation in rats, we recorded EPPS in response to repetitive stimulation at 10, 20, 40 and
75 Hz. In adult muscle, the incidence of a failure to evoke EPPS was found to be dependent
on stimulation frequency, and was not appreciable until evoked at a stimulation rate of 75
Hz (Table 2). In contrast, in neonates the incidence ofEPP failure was much more appreciable
at all stimulation frequencies, and far more prevalent as compared to adults (Table 2).
Presumably, the failure to evoke EPPS reflected a presynaptic axonal propagation failure,

Table 2. Postnatal changes in axonal propagation failure and EPP amplitude in rat diaphragm
muscle fibers during repetitive stimulation at different rates
Propagation failure rate (%) % Decrement in EPP amplitude

Age 10Hz 20Hz 40Hz 75 Hz 10Hz 20Hz 40Hz 75 Hz

0-7 days 4.4±1.7* l4.5±3.5* 40.5±3.8* 65.6±2.2* 52.2±1O.8 70.0±8.2 87.2±7.7* 91.3±4.4*
14-21 days O.O±O.O 2.0±1.0 16.7±4.2** 37.5±4.8** 63.7±9.7 61.2±5.2 78.5±3.l 77.4±2.8
Adult O.O±O.O 2.5±0.5 4.0±2.4 26.0±5.l 61.3±7.8 62.8±6.2 73.2±2.7 75.1±1.3
Propagation failure rate was calculated as the proportion of absent evoked EPPS in the first 10 pulses of each
stimulus train. The % decrement in EPP amplitude was that noted in the lOth stimulus pulse. Values are means
± SD. * indicates significant difference (p<0.05) from older animals. ** indicates significant difference
(p<0.05) from adults.
88 G. C. Sieck and Y. S. Prakash

most likely occurring at axonal branch points (see below). We also concluded that the higher
incidence of axonal branch point failure in neonates reflected the increased axonal branching
associated with polyneuronal innervation of muscle fibers. In both the adult and neonatal
diaphragm, repetitive nerve stimulation was also associated with a decrement in EPP
amplitude (Table 2). At both ages, this decrement was dependent on stimulation frequency.
In contrast to the incidence of EPP failure, the decrement in EPP amplitude was greater in
neonates only at higher rates of stimulation. The decrement in amplitude with repetitive
stimulation could reflect either pre- or postsynaptic mechanisms (see below), but certainly
synaptic efficacy would be reduced as a result.
As mentioned above, muscle fiber type is an important determinant of susceptibility
to NTF. Fast-twitch glycolytic and/or type IIblIIx fibers, which appear to be most susceptible
to NTF (Kugelberg & Lindegren, 1979; Johnson & Sieck, 1993), comprise motor units that
are more fatigable (Burke et aI., 1973; Sieck et aI., 1989). Accordingly, in motor unit studies,
it has been demonstrated that fast-twitch fatigable (type FF) motor units are more susceptible
to NTF than more fatigue-resistant fast-twitch (type FR) and slow-twitch (type S) units
(Clamann & Robinson, 1985; Sandercock et aI., 1985; Sieck & Fournier, 1990).
In the cat diaphragm, we estimated the extent of NTF in different motor unit types
by assessing changes in the evoked motor unit action potential amplitude (M wave) during
repetitive stimulation (Figs. 4 and 5).
Motor units were isolated by microdissection of cervical (C 4-C6) ventral root fila-
ments and classified as type F or S based on the presence or absence of "sag" in unfused
tetani, respectively (Burke et aI., 1973). Motor unit fatigue was assessed by repetitive
stimulation at 40 Hz in 330 ms duration trains repeated each second. A fatigue index was

A Unit #1 120

80

~
MWave

40
MWave A A 1\
rms
...J " 0 60 120
B Unil#2
120 • Force
Cl MWave
rms
MWave
80

MWave
rms
JI A f\I \ 40

-i ~

Train
Number I I I I I 0 60 120
2 3 118 119 120 Time (s)

Figure 4. Evoked motor unit M waves were recorded in the adult cat diaphragm during repetitive stimulation
at 40 Hz. The rms amplitude of the 13 evoked M waves in each train was calculated. Some motor units displayed
very little change in evoked M waves during the 2-min stimulation period (A: unit #1) while other units
displayed a marked reduction in M wave amplitude (B: unit #2).
Fatigue at the Neuromuscular Junction 89

FF
., •••
Flnt
1.
FRand S

,
100 000

:a
0 0
I· 0 00.

• •
:~
••
I
cP

.. •
...... 75
Figure 5. Changes in the ratio of
~
'-'
Cl)
•
motor unit M wave rms after 2 min
~ •
stimulation to the initial M wave
rms value (M wave index). A fa- ]-
50
•• ~ • • •
o Slow
tigue index was also calculated as < I
•• I.
•• • Fast
the ratio of force generated after 2 Cl)

min stimulation to the initial force. 1; 25 • .1

Fatigue resistant type Sand FR ~ I
~
units displayed very little NTF (M
wave index> 0.75), while type FF
I
•
and FInt units displayed variable ~.oo 0.25 0.50 0.75 1.00
extents of NTF. Fatigue Index

then calculated as the ratio force generated after 2 min stimulation to the initial force (Burke
et aI., 1973). During repetitive stimulation, evoked unit M waves were recorded using
fine-wire electrodes. The electromyographic (EMG) signals were amplified, band-pass
filtered (20 Hz to 1 kHz), and then root mean square (rms) ofthe 13 M waves in each stimulus
train was calculated based on the following transfer equation

Vrms = ...JAVG(V2 EMG)

where Vrms was the voltage output of the rms circuit, V EMG was the voltage input of the EMG
signal, and AVG was the averaging time constant (55 ms) of the rms circuit. Since the duration
of the EMG signal (i.e., train duration of330 ms) far exceeded the AVG duration (i.e., 55
ms), the rms value reached 99% of maximum output (i.e., rms plateau) during each stimulus
train. Therefore, this rms calculation yielded an output signal that was directly related to the
total power of the evoked EMG signal during each stimulus train. Changes in rms of motor
unit M waves were used to assess the presence of NTF. In some motor units, a force
decrement was dissociated from any NTF (Fig. 4A), while in other units the force decrement
paralleled the decline in M wave amplitude (Fig. 4B).
In a popUlation of motor units in the cat diaphragm, the extent ofNTF was estimated
by calculating the ratio of M wave rms after 2 min of stimulation to the initial M wave rms
value. This M wave index was then compared to a fatigue index calculated as the ratio of
motor unit force after 2 min stimulation to the initial force (Fig. 5). Type FR and S motor
units (fatigue index>0.75) displayed little change in M wave index, indicating very little
NTF. In contrast, type FF motor units (fatigue index < 0.25) and fatigue intermediate
fast-twitch units (type FInt; fatigue index 0.25 to 0.75) displayed a range of decrement in M
wave amplitudes.

PRESYNAPTIC SITES OF NEUROMUSCULAR TRANSMISSION

FAILURE
Presynaptic sites where NTF might occur can be broadly categorized as being
localized either to axons or to the presynaptic terminal. At the axon level, NTF can result
from a failure of axonal propagation of action potentials. At the muscle fiber, axonal
propagation failure would present itself as an absence of evoked EPPS. NTF can also result
90 G. C. Sieck and Y. S. Prakash

from mechanisms localized to the presynaptic terminal leading to inadequate neurotransmit-

ter release. At the muscle fiber, failure localized to the presynaptic terminal would appear as
a reduction in EPP amplitude and ineffective synaptic transmission, although such a
depression ofEPP amplitude could also reflect postsynaptic mechanisms ofNTF (see below).

Failure of Axonal Propagation of Action Potentials

It has been recognized for a some time that action potentials may fail to propagate
along each branch of a motor axon (Barron & Matthews, 1935). In experiments in the rat
diaphragm, Krnjevic and Miledi (1958, 1959) observed that prolonged repetitive stimulation
of the phrenic nerve resulted in a significant incidence of action potential propagation failure
which they attributed to a failure of propagation at axonal branch points. They also noted
that the incidence of presynaptic NTF was greater at higher rates of stimulation (Krnjevic
& Miledi, 1959). Subsequently, a number of studies have also reported such axonal
conduction blocks in the peripheral nervous system in a variety of species (Bittner, 1968;
Parnas, 1972; Grossman et aI., 1973, 1979; Hatt & Smith, 1976; Smith, 1980, 1983; Schiller
& Rahamimoff, 1989). In a detailed study of the excitor axon innervating the abductor muscle
of the crayfish, Smith (1980) reported that prolonged stimulation of the axon led to action
potential propagation failure at axonal bifurcations, with blockage occurring first in the more
peripheral terminal arborization of the axon and then progressively spreading centrally to
regions where axonal caliber was larger (i.e., centripetal blocking pattern; Fig. 6). Therefore,
points of axonal bifurcation are considered to be sites where failure of action potential
generation and neurotransmission may occur (Raymond & Lettvin, 1978).
The potential mechanisms underlying the failure of axonal action potential propaga-
tion are summarized in Fig. 6. Accompanying action potential failure, there is also a
prolonged depolarization of the axonal membrane and a decrease in the inward Na+ current
(Hatt & Smith, 1976). Axonal propagation failure can be reversed by hyperpolarization of
the axon and/or by placing the preparation in a solution with low external K+ levels. These
observations suggested that a prolonged refractory period of the axon may be responsible
for the failure of action potential generation and the consequent propagation failure at these
axonal branch points. Alterations in perineural Na+ and K+ concentrations with repetitive
axonal stimulation may therefore be responsible for axonal conduction block. In support, it
has been shown that during repetitive stimulation, an accumulation of K+ in the space
surrounding the axon is associated with the development of conduction failure (Adelman et

Figure 6. Potential mechanisms un-

derlying axonal branch point failure in-
clude: changes in axonal geometry

\ (higher axial resistivity, ra and lower

membrane capacitance em of daughter
branches); and increased membrane re-
fractoriness at smaller branches due to
perineural accumulation of K+ result-
ing from repetitive stimulation. As a
result, axonal propagation failure dis-
Centripetal Blocking Pattern plays a centripetal pattern.
Fatigue at the Neuromuscular Junction 91

ai., 1973; Grossman et aI, 1979; Smith, 1983). Shifts in ionic concentrations are more likely
to occur in regions with smaller axonal sizes, since the surface-to-volume ratio is increased
in these smaller axons, resulting in more pronounced changes in membrane potential due to
ionic concentration changes (Smith, 1980).
Experimental observations of axonal conduction failure have led to a number of
theoretical models that have focused on the contributions of axonal geometry on impulse
propagation (Khodorov et ai., 1969; Goldstein & Rall, 1974; Waxman, 1975; Swad10w et
ai., 1980; Stockbridge, 1988; Stockbridge & Stockbridge, 1988). Even if periaxona1 accu-
mulation of K+ ions is ignored, changes in the excitability of the axonal membrane caused
by differences in axonal geometry can lead to propagation failure in smaller axons, depend-
ing on impulse frequency in the parent axon (Stockbridge, 1988; Stockbridge & Stockbridge,
1988). For example, in a bifurcating axon, a short region of higher axial resistivity and lower
capacitance in one of the daughter branches, compared to the rest of the adjacent axonal
regions, can produce a frequency-dependent differential conduction along that daughter
branch, in comparison to the other, normal branch. Even a very small change in resistivity
and/or capacitance may result in axonal conduction being limited only to a band of
stimulation frequencies, with varying probabilities of conduction block at different stimula-
tion frequencies. Further, in shorter daughter branches, axonal conduction may be facilitated
at higher frequencies of stimulation, compared to longer daughter branches. These complex
interactions between the susceptibility to action potential propagation failure and axonal
geometry may lead to NTF occurring in only a portion of muscle fibers within a motor unit.
The probability of blockade of action potentials at axonal branch points may be an
important factor influencing the differences in neuromuscular transmission properties of
different types of motor units. For example, type FF motor units, which are most susceptible
to NTF (Fig. 5), tend to have more muscle fibers within the unit, i.e., a larger innervation
ratio. Therefore, axons innervating type FF motor units with larger innervation ratios will
have a greater number of axonal branches, and possibly, a greater probability of axonal
conduction block as a result. Sandercock and colleagues (1985) recorded muscle fiber action
potentials in fibers comprising motor units in the cat medial gastrocnemius muscle using
microelectrodes, and reported that occasionally there was a failure to evoke action potentials
in some muscle unit fibers. The incidence of such failures was more prevalent in type FF
motor units as compared to other motor unit types. They also noted that the incidence of
presynaptic NTF was dependent on the frequency of stimulation. We recorded evoked motor
unit M waves in the cat diaphragm, and observed that in type FF and FInt motor units, abrupt
transient changes in the waveform ofM waves occurred (Sieck & Fournier, 1990). Presum-
ably, the presynaptic failure of neuromuscular transmission occurred in only some muscle
unit fibers. A presynaptic action potential conduction block at axonal branch points would
explain these observations of abrupt changes in M wave waveform. Moreover, such a
conduction block of only some muscle unit fibers during repetitive stimulation would reduce
the effective innervation ratio ofthese motor units and lead to reductions in mechanical force
(fatigue).
The higher incidence ofNTF during early postnatal development in the rat diaphragm
(Table 2) may also be influenced by axonal conduction block. During early postnatal
development, muscle fibers are innervated by more than one motoneuron (polyneuronal
innervation; Redfern, 1970; Bagust et ai., 1974). Therefore, motor unit innervation ratios
are greater during early postnatal development, and there are more axonal branch points
where conduction block might occur. Subsequently, as a result of synapse elimination, motor
unit innervation ratio decreases, and the number of axonal branch points is reduced accord-
ingly. In the rat diaphragm, synapse elimination is complete by postnatal day 14 (Sieck &
Fournier, 1991). A greater number of axonal branch points would increase the probability of
axonal propagation failure in neonatal motor units (Table 2). Furthermore, the smaller axonal
92 G. C. Sieck and Y. S. Prakash

calibers in the neonate would increase axial resistivity and lower membrane capacitance.
These intrinsic electrophysiological properties combined with possible heterogeneity of
axonal calibers may be conducive to an increased incidence of axonal branch point conduc-
tion block. In addition, as mentioned above, shifts in ionic concentrations are more likely to
occur around smaller axons.

Mechanisms of Failure at the Presynaptic Terminal

At the presynaptic terminal, NTF can result from a reduction in the release of ACh.
Fig. 7 summarizes several potential mechanisms that might underlie a reduction in presy-
naptic ACh release.
By reducing extracellular Ca2+ at the frog neuromuscular junction, del Castillo and
Katz (1954) demonstrated that Ca2+ influx is critical for neuromuscular transmission.
Subsequently, a number of investigators have confirmed this dependence of neurotransmitter
release on Ca2+ influx at the presynaptic terminal. Using a voltage clamp technique, Llinas
and Heuser (1977) further demonstrated that Ca2+ influx at the presynaptic terminal is via
voltage-dependent Ca2+ channels, and that there is very little delay between Ca2+ influx and
transmitter release, suggesting that these Ca2+ channels were located in close proximity to
the active zones of the nerve terminal. The entry ofCa2+ via voltage-dependent channels into
the nerve terminal leads to a fusion of synaptic vesicles to the nerve terminal membrane, and
the release of ACh into the synaptic cleft via a second messenger cascade. While the exact
cascade of events is not entirely clear, studies using isolated, highly enriched preparations
of synaptic vesicles and nerve terminal cytoplasm have suggested that calmodulin, a Ca2+
binding protein, plays a major role in regulating neurotransmitter release (DeLorenzo, 1981).
Whether calmodulin or other second messengers involved in excitation-secretion coupling
at the presynaptic terminal are affected during repetitive stimulation remains unclear.
However, quantal (vesicular) release of ACh at the neuromuscular junction may be reduced
by either a decrease in Ca2+ influx through voltage-gated Ca2+ channels or by a reduction in
the Ca2+ sensitivity of processes within the nerve tenninal that lead to vesicular release.
Since the study of del Castillo and Katz (1954), it has been recognized that during
repetitive nerve stimulation of curarized nerve-muscle preparations, the change in EPP
amplitude is dynamic. There is an initial facilitation of EPP amplitude, followed by a
depression. Intracellular Ca2+ at presynaptic terminals can accumulate with repetitive stimu-

Presynaptic
I ACh Quantal Release I ACh Quantal Size
I Terminal depolarization Altered vesicular
(C;'+ entry) recycling?
I Calcium sensilivity
I # of vesicles

AChR I Sarcolemmal Figure 7. Possible pre- and postsynaptic

Desensitization Excitability sites where NTF might occur. At the presy-
ACh accumulation I Extracellular K+ naptic terminal, transmission failure might
at receptors Na+ channel
inactivation
arise from a reduction in ACh quantal release
and/or a decrease in quantal size. At the post-
synaptic endplate, ACh receptors (AChR)
may become desensitized and/or sarcolem-
Postsynaptic mal excitability may be reduced.
Fatigue at the Neuromuscular Junction 93

lation, and this may form the basis for the initial facilitation ofEPP amplitude, i.e., increased
probability of quantal release and increased quantal content. The decrement in EPP amplitude
with repetitive nerve stimulation has been attributed to either a reduction in the number of
quanta of ACh being released by the presynaptic nerve terminal, or to a reduction in quantal
size (i.e., the amount of ACh per vesicle as indicated by the average miniature endplate
potential, MEPP, amplitude) (del Castillo & Katz, 1954; Jones & Kwanbunbumpen, 1970;
Kurihara & Brooks, 1975; Smith, 1984) . A reduction in the number of vesicles released by
the presynaptic terminal during repetitive nerve stimulation may be attributed to a reduction
in Ca2+ influx, a decrease in Ca2+ sensitivity of excitation-secretion coupling or a depletion
of the number of vesicles available for release. Repetitive stimulation of the nerve terminal
may lead to an accumulation of extracellular K+ due to limited diffusion possibilities
(Frankenhaeuser & Hodgkin, 1956; Meech, 1974; Hatt & Smith, 1976; Smith, 1980). As a
result, the terminal membrane would depolarize thereby initially increasing Ca2+ influx.
However, the increase in intracellular Ca2+ may also increase K+ conductance via Ca2 +-de-
pendent K+ channels, making the terminal membrane refractory to further depolarization
(Meech, 1974). The net effect would be a reduction ofCa2+ influx and reduced neurotrans-
mitter release.
Ca2+ sensitivity of excitation-secretion coupling may also be reduced by repetitive
nerve stimulation. Volle and Branisteanu (1976) reported that during repetitive stimulation
of the frog neuromuscular junction, there was a reduction in the number of quanta released,
despite an apparent increase in intracellular Ca2+ as affected, experimentally, by increasing
extracellular Ca2+. This observation suggests that repetitive stimulation decreases the Ca2 +
sensitivity of ACh release. In melanotrophs of the rat pituitary gland, Okano and colleagues
(1993) demonstrated that vesicular release was reduced following repetitive exposure to
extracellular Ca2 +, although intracellular Ca2+ was unaffected. They attributed the reduced
Ca2+ sensitivity of vesicular release of histamine in mast cells to a depletion of guanosine
triphosphatase (GTP). It is also possible that at the neuromuscular junction, depletion of
GTP, or some other factor in a second messenger cascade, may also play a role in the
reduction of Ca2+ sensitivity of ACh release following repetitive stimulation.
At the frog neuromuscular junction, Volle and Branisteanu (1976) also found that
quantal release was reduced during repetitive nerve stimulation despite an increase in the
probability of release (MEPP frequency). They concluded that the decrease in quantal release
was due, at least in part, to a depletion of synaptic vesicles. There is morphological evidence
of vesicular depletion at the presynaptic terminal following repetitive nerve stimulation. For
example, based on electron microscopic (EM) evidence in the electric organ of the narcine
brasiliensis, repetitive exhaustive stimulation of the nerve was reported to result in nearly a
50% reduction in the total number of synaptic vesicles (Boyne et aI., 1975). Lentz and
Chester (1982) used the uptake of horseradish peroxidase at the frog neuromuscular junction
to follow presynaptic events during nerve stimulation, and found depletion of synaptic
vesicles along with increased axolemmal infoldings and membranous cisternae. This sug-
gests that vesicular recycling is impaired following repetitive stimulation.
In addition to a reduction in the number of quanta released, it is also possible that
following repetitive stimulation there is a reduction in quantal size. At the neuromuscular
junction of the rat diaphragm, Jones and Kwanbunbumpen (1970) reported that repetitive
nerve stimulation caused both a reduction in the number of quanta released, and a reduction
in quantal size. However, in this study, curare was used to abolish muscle contractions, which
causes difficulties in the calculation of quantal size. In the rat diaphragm muscle, Kurihara
and Brooks (1975) used a cut muscle preparation to abolish muscle contractions. They
showed that the initial decrease in EPP amplitude during repetitive nerve stimulation (at 30
Hz) was entirely due to a reduction in the number of quanta released, without any decrement
of quantal size. They concluded that the decrement in EPP amplitude during this initial period
94 G. C. Sieck and Y. S. Prakash

was not sufficient to cause a failure in neuromuscular transmission. However, following this
initial period, these authors found that while quantal release continued to be reduced, there
was also a linear decline in quantal size (i.e., average MEPP amplitude). They further
concluded that during this latter stage of repetitive nerve stimulation, the reduction in EPP
amplitude was sufficient to cause NTF and muscle fatigue. However, the underlying cause
of a reduction in quantal size is unclear.
Quantal content and release may be affected in a variety of conditions where there
are adaptations in the neuromuscular junction. For example, during early postnatal develop-
ment there is a rapid growth of the neuromuscular junction in parallel with an increase in
muscle fiber size. We recently observed that the postnatal growth of endplates on type I
muscle fibers (comprising type S motor units), exceeded that dictated by muscle fiber
growth, while the growth of type II fiber (comprising fast-twitch motor units) endplates
matched the postnatal change in fiber size. During early postnatal development, there is a
progressive decrease in MEPP amplitude (Diamond & Miledi, 1962, Kelly 1978), which
may be attributed to a change in muscle fiber size and/or a decrease in quantal size. In an in
vitro nerve diaphragm muscle preparation, Kelly (1978) reported that the reduction in MEPP
amplitude during early postnatal development was accompanied by an increase in the quantal
content ofEPPS. Kelly (1978) also reported that the quantal content ofEPPS decreased by
over two-fold during prolonged repetitive stimulation at 10 Hz. The safety factor for
neuromuscular transmission was also reported to be lower in the neonatal diaphragm as
compared to adults, and at all ages, quantal content decreased with continued repetitive
stimulation (Kelly, 1978). A reduction in MEPP amplitude occurs during aging (Gutmann et
aI., 1971; Kelly & Roberts, 1977; Kelly, 1978; Smith, 1984). In general, quantal content has
been reported to be higher in older animals (Kelly & Roberts, 1977; Smith, 1984; Alshuaib
& Fahim, 1990), suggesting an increase in quantal release, since MEPP amplitude is reduced.
During early postnatal development and in older animals, it has also been reported
that MEPP frequency is lower (Kelly & Zacks, 1969; Gutmann et aI., 1971; Kelly, 1978;
Smith, 1984). It is likely that in neonates, the lower MEPP frequency reflects the smaller
surface area of the neuromuscular junction and the associated reduction in active sites for
ACh quantal release. Whether such a morphological basis underlies the lower MEPP
frequency with aging remains unknown.

POSTSYNAPTIC SITES OF NEUROMUSCULAR TRANSMISSION

FAILURE

There are two major postsynaptic sites where NTF can occur: at the motor endplate
due to a desensitization of the cholinergic receptor (AChR); and at the sarcolemma due to a
reduction in excitability. Krnjevic and Miledi (1959) provided evidence that NTF at the
neuromuscular junction ofthe rat diaphragm muscle was due to a combination of a decrement
in EPP amplitude and a reduction in sarcolemmal excitability, leading to a failure in the
generation of action potentials in muscle fibers.

Desensitization of the Cholinergic Receptor

Katz and Thesleff (1957) found that prolonged exposures of the neuromuscular
junction to ACh resulted in a reduction in endplate conductance due to a desensitization of
the AChR. Desensitization involves a slow transition of AChR channels to non-conducting
states due to the continued presence of ACh, with recovery only upon removal of stimulation
(Katz & Thesleff, 1957; Sakmann et aI., 1980; Feltz & Trautmann, 1982). As mentioned
Fatigue at the Neuromuscular Junction 95

above, the decrement in EPP amplitude during repetitive stimulation could reflect a decrease
in quantal release, but it could also reflect a desensitization of the AChR due to continued
accumulation of ACh in the synaptic cleft, compounded by inadequate diffusion of the
transmitter out of the synaptic cleft, and incomplete hydrolysis of the ACh by acetylcholi-
nesterase (Katz & Thesleff, 1957), especially at higher frequencies of stimulation. At the
frog neuromuscular junction in cut muscle fibers in the absence of acetylcholinesterase,
Giniatullin and colleagues (1986) observed a 23% reduction in MEPP amplitude following
repetitive stimulation at 10Hz over a period of 60 s, indicating AChR desensitization.
Albuquerque and colleagues (1986) have suggested that cyclic AMP may be involved in the
phosphorylation of the AChR. During repetitive stimulation, changes in cAMP level may
lead to AChR desensitization, and reduced EPP amplitUdes.

Reduction in Sarcolemmal Excitability

Repetitive activation of the neuromuscular junction may reduce sarcolemmal excitabil-
ity (Gruber, 1914; Rosenblueth, 1940; Krnjevic & Miledi, 1958; Metzger & Fitts, 1986; Roed,
1988; see also Fuglevand, Chapter 6). Edwards (1984) proposed that fatigue could be
categorized as either high-frequency fatigue, characterized by a transient, rapidly recovering
loss offorce after high frequencies of stimulation, and low-frequency fatigue, characterized by
prolonged loss in force. It was suggested that high-frequency fatigue was due, at least in part,
to a failure of action potential propagation by the sarcolemma, possibly due to an accumulation
of extracellular K+, membrane depolarization and Na+ channel inactivation leading to a
reduction in Na+ conductance (Bigland-Ritchie et aI., 1979; Edwards, 1984). However, the
influence of K+ on sarcolemmal excitability during repetitive stimulation of muscle remains
controversial. Juel (1988) showed that high-frequency fatigue of the mouse soleus and extensor
digitorum longus muscles was associated with a decrease in action potential propagation
velocity which was induced by increased extracellular K+, but was nearly independent of
moderate changes in the Na+ gradient. Sjogaard (1991) also suggested that an accumulation of
K+ occurred following repetitive stimulation. She attributed the increase in extracellular K+ to
the opening of Ca2+-dependent K+ channels. The likelihood that the increase in interstial K+
contributes to fatigue is considered in Chapter 4, Sjogaard and McComas.
While reduced sarcolemmal excitability may result from an accumulation of extracel-
lular K+, a number of studies have suggested that changes in extracellular K+ concentration are
not large enough to change sarcolemmal excitability. Using intracellular recordings in an
isolated rat phrenic nerve-diaphragm muscle preparation, Metzger and Fitts (1986) found no
change in resting membrane potential of muscle fibers following either high- or low-frequency
stimulation, thus arguing against a significant accumulation of extracellular K+. Instead, they
suggested that the decline in amplitude of sarcolemmal action potentials during repetitive
stimulation was most likely due to changes in extracellular Na+. Renaud and Light (1992) also
investigated the role of extracellular K+ in muscle fatigue by testing whether increasing K+
concentration in unfatigued frog sartorius muscle fibers caused a decrease in force similar to
that observed during fatigue. However, accumulation of K+ at the surface of the sarcolemma
was not sufficiently high to reduce force during repetitive stimulation via blockade of muscle
action potential propagation. However, they suggested that K+ may playa role in causing failure
at the level of the T-tubule; thus affecting excitation-contraction coupling.
In addition to changes in Na+ and K+ ionic concentrations at the sarcolemma, pH may
also playa role in muscle fatigue. Renaud (1989) used ion-selective microelectrodes on frog
sartorius muscle to measure pH during tetanic contractions elicited by field stimulation, and
found that extracellular pH inhibits tetanic force recovery by acting on the outer surface of
the sarcolemma. In the isolated hamster diaphragm, Brody and colleagues (1991) reported
that changing extracellular pH initially results in decreased conduction velocity along the
96 G. C. Sieck and Y. S. Prakash

muscle membrane and a leftward shift in the median frequency of the EMG. However, with
stimulation, the rate of decay of conduction velocity was less than the rate of decay of the
median frequency. This suggested a complex interaction between pH and sarcolemmal
excitability.
Sarcolemmal excitability is influenced by the intrinsic electrophysiological properties
of muscle fibers. For example, larger muscle fibers will have a lower axial resistance and higher
membrane capacitance, and this will affect the extent of depolarization for any given level of
synaptic current. For this reason, the size of the motor endplate varies proportionately with fiber
diameter (Nystrom, 1968; Wemig et ai., 1986). Motor endplates on type I muscle fibers are
larger than those on type II fibers (Ellisman et ai. , 1976; Nystrom, 1968; Fahim et ai. , 1984).
Furthermore, the relationship between fiber diameter and endplate size is true only within a
given fiber type since type I fibers are generally smaller than type II fibers . The matching of
endplate size with fiber size is important in that it will ensure adequate synaptic current for
action potential generation in muscle fibers (safety factor; ratio ofEPP amplitude to threshold
for action potential generation). The safety factor for type II muscle fibers is greater than that
for type I fibers and, in both fiber types, sufficiently large to ensure reliable generation of action
potentials. With repetitive stimulation, the safety factor for type I fibers shows little change
while that for type II fibers displays a rapid decline (Gertler & Robbins, 1978).
The neuromuscular junction can adapt during postnatal development and under
conditions of altered muscle use (Balice-Gordon & Lichtman, 1990, Balice-Gordon et ai.,
1990). For example, during early postnatal development in the rat diaphragm, motor
endplates on both type I and II fibers become larger as muscle fiber size increases. However,
the growth of motor endplates on type I fibers exceeds that predicted by the increase in fiber
size alone while the growth of endplates on type II fibers matches the change in fiber
diameter. Perhaps this disproportionate growth of motor endplates on type I and II fibers
contributes to the increased susceptibility of type II fibers to NTF.
We have observed that following two weeks of diaphragm muscle inactivity induced by
cervical spinal cord hemisection (spinal isolation) motor endplates on type II fibers become
elongated and increase in area relative to fiber diameter (Fig. 8). This morphological change is
associated with an improvement in neuromuscular transmission. Following two weeks of
inactivity in the spinal isolated animals, type II fiber cross-sectional area was unaffected
suggesting that sarcolemmal excitability was most likely unchanged. The improved neuromus-
cular transmission in the spinal isolated animals was therefore most likely attributable to an
increase in synaptic current due to increased quantal release. In contrast, following two weeks
of diaphragm muscle inactivity induced by TTX blockade of phrenic axonal propagation, type
II fibers atrophied while motor endplate size remained unaffected. As a result, the relative size

150
.1
o CTL

CJ SI
Figure 8. Surface area of motor endplates on type· identified
• TTX
,.I. rat diaphragm muscle fibers was normalized for fiber diame-
:f, ter. In controls, the normalized endplate surface area on type
.. ': I fibers was larger than that on type II fibers. Following two
weeks of inactivity induced by spinal isolation (SI) or TTX

o I
Type I Type II
-
blockade, the normalized endplate surface area on type II
fibers increased whereas that on type I fibers remained un-
changed.
Fatigue at the Neuromuscular Junction 97

of motor endplates on type II fibers increased, as in the spinal isolated animals (Fig. 8).
However, this increase in the relative size of motor endplates on type II fibers in the TTX
animals was associated with an increased susceptibility to NTF. Since in the TTX animals, there
was a reduction in cross-sectional area of type II muscle fibers, the sarcolemma should be more
excitable if anything. Therefore, the increased susceptibility to NTF in the TTX animals most
likely resulted from reduced quantal release of ACh.

SUMMARY AND CONCLUSIONS

A number of potential sites of transmission failure have been identified, although the
relative contributions of these to muscle fatigue remain controversial. Blockade of axonal
propagation of action potential, due to changes in axonal excitability and/or axonal geometry,
may partly explain differences in the excitability of different motor unit types to NTF, and
the greater incidence of transmission failure in neonatal muscles. Failure of ACh release at
the presynaptic terminal may be due to reduced Ca2+ influx, decreased Ca 2+ sensitivity of
excitation-secretion coupling and/or a depletion of synaptic vesicles. At the postsynaptic
membrane, AChR desensitization may be an important cause of NTF during repetitive
stimulation. In addition, reduced sarcolemmal excitability may also playa role in NTF.
Clearly, each of these potential mechanisms may adapt under a variety of conditions, and
the relative contribution of each might vary depending on fiber type.

ACKNOWLEDGMENTS

The authors would like to acknowledge the invaluable contributions of Drs. Mario
Fournier and Hirofumi Miyata. This work was supported by United States Public Health
Service grants HL 34817 and HL 37680. Attendance of [Link]. at the 1994 Bigland-Ritchie
conference was supported, in part, by the University of Miami.

REFERENCES
Adelman WJ, Plati Y & Senft JP (1973). Potassium ion accumulation in a periaxonal space and its effect on
the measurements of membrane potassium ion conductance. Journal of Membrane Biology 12,
387-410.
Albuquerque EX, Deshpande SS, Aracava Y, Alkondon M & Daly JW (1986). Possible involvement of cyclic
AMP in the expression of desensitization of the nicotinic acetylcholine receptor. A study with
forskolin and its analogs. FEBS Letters 199, 113-20.
Aldrich TK, Shander A, Chaudhry I & Nagashima H (1986). Fatigue of isolated rat diaphragm: role of impaired
neuromuscular transmission. Journal ofApplied Physiology 61,1077-1083.
Alshuaib WB & Fahim MA (1990). Aging increases calcium influx at motor nerve terminal. International
Journal of Developmental Neuroscience 8, 655-666.
Bagust J, Lewis DM & Westerman RA (1974). The properties of motor units in a fast and slow twitch muscle
during postnatal development in the kitten. Journal of Physiology (London) 237, 75-90.
Balice-Gordon RJ, Breedlove SM, Bernstein S & Lichtman JW (1990). Neuromuscular junctions shrink and
expand as muscle fiber size is manipulated: in vivo observations in the androgen-sensitive bulbocav-
ernosus muscle of mice. Journal of Neuroscience 10, 2660-2671.
Balice-Gordon RJ & Lichtman JW (1990). In vivo visualization ofthe growth of pre- and postsynaptic elements
of neuromuscular junctions in the mouse. Journal of Neuroscience 10, 894-908.
Barron DH & Matthews BHC (1935). Intermittent conduction in the spinal cord. Journal of Physiology
(London) 85, 73-103.
Bazzy AR & Donnelly DF (1993). Diaphragmatic failure during loaded breathing: role of neuromuscular
transmission. Journal ofApplied Physiology 74, 1679-1683.
98 G. C. Sieck and Y. S. Prakash

Bellemare F & Bigland-Ritchie B (1987). Central components of diaphragm fatigue assessed by phrenic nerve
stimulation. Journal ofApplied Physiology 62,1307-1316.
Bigland-Ritchie B (1984). Muscle fatigue and the influence of changing neural drive. In: Loke J (ed.),
Symposium on Exercise: Physiology and Clinical Applications. Clinics in Chest Medicine 5,21-34.
Bigland-Ritchie B, Jones DA, Hosking GP & Edwards RHT (1978). Central and peripheral fatigue in sustained
maximum voluntary contractions of human quadriceps muscle. Clinical Science and Molecular
Medicine 54,609-614.
Bigland-Ritchie B, Jones DA & Woods JA (1979). Excitation frequency and muscle fatigue: Electrical
responses during human voluntary and stimulated contractions. Experimental Neurology 64, 414-427.
Bigland-Ritchie B, Kukulka CG, Lippold OCJ & Woods JJ (1982). The absence of neuromuscular transmission
failure in sustained maximal voluntary contractions. Journal of Physiology (London) 330, 265-278.
Bigland-Ritchie B & Lippold OCJ (1979). Changes in muscle activation during prolonged maximal voluntary
contractions. Journal ofPhysiology (London) 292, 14-15P.
Bittner GD (1968). Differentiation of nerve terminals in crayfish opener muscle and its functional significance.
Journal of General Physiology 51, 731-758.
Boyne AF, Bohan TP & Williams TH (1975). Changes in cholinergic synaptic vesicle populations and the
ultrastructure of the nerve terminal membranes of N arcine brasiliensis electric organ stimulated to
fatigue in vivo. Journal of Cell Biology 67,814-25.
Brody LR, Pollock MT, Roy SH, DeLuca CJ & Celli B (1991). pH-induced effects on median frequency and
conduction velocity of the myoelectric signal. Journal ofApplied PhYSiology 71,1878-85.
Burke RE, Levine DN, Psairis P & Zajac FE (1973). PhYSiological types and histochemical profiles of motor
units of cat gastrocnemius. Journal ofPhysiology (London) 234, 723-748.
Clamann HP & Robinson AJ (1985). A comparison of electromyographic and mechanical fatigue properties
in motor units of the cat hindlimb. Brain Research 327, 203-219.
del Castillo J & Katz B (1954). The effect of magnesium on the activity of motor nerve endings. Journal of
Physiology (London) 124, 553-559.
DeLorenzo RJ (1981). The calmodulin hypothesis ofneurotransmission. Cell Calcium 2, 365-385.
Diamond J & Miledi R (1962). A study of foetal and new-born rat muscle fibers. Journal of Physiology
(London) 162, 393-408.
Edwards RHT (1984). New techniques for studying human muscle function, metabolism and fatigue. Muscle
& Nerve 7, 599-609.
Ellisman MH, Rash JE, Staehelin LA & Porter KR (1976). Studies of excitable membranes II. A comparison
of specializations at neuromuscular junctions and nonjunctional sarcolemmas of mammalian fast and
slow twitch muscle fibers. Journal of Cell Biology 68,752-774.
Fahim MA, Holley JA & Robbins N (1984). Topographic comparison of neuromuscular junctions in mouse
slow and fast twitch muscles. Neuroscience 13, 227-235.
Feldman JD, Bazzy AR, Cummins TR & Haddad GG (1991). Developmental changes in neuromuscular
transmission in the rat diaphragm. Journal ofApplied Physiology 71, 280-286.
Feltz A& Trautmann A (1982). Desensitization at the frog neuromuscular junction: A biphasic process. Journal
of Physiology (London) 332, 257-272.
Fournier M, Alula M & Sieck GC (1991). Neuromuscular transmission failure during postnatal development.
Neuroscience Letters 125, 34-36.
Frankenhaeuser B & Hodgkin AL (1956). The after-effects of impulses in the giant nerve fibres of Loligo.
Journal of Physiology (London) 131,341-376.
Gertler RA & Robbins N (1978). Differences in neuromuscular transmission in red and white muscles. Brain
Research 142, 160-164.
Goldstein SS & Rail W (1974). Changes of action potential shape and velocity for changing core conductor
geometry. Biophysical Journal 14, 731-757.
Grob D (1961). Muscular disease. Bulletin ofthe New York Academy ofMedicine 37,809-834.
Grossman Y, Parnas I & Spira ME (1979). Differential conduction block in branches ofa bifurcating axon.
Journal of Physiology (London) 295, 283-305.
Grossman Y, Spira ME & Parnas I (1973). Differential flow of information into branches of a single axon.
Brain Research 64, 379-386.
Gruber CM (1914). Studies in fatigue. IV. The relation of adrenalin to curare and fatigue in normal and
denervated muscles. American Journal ofPhysiology 34, 89-96.
Giniatullin RA, Baltser SK, Nikolskii EE & Magazanik LG (1986). Postsynaptic potentiation and desensiti-
zation of the myoneural synapse of the frog induced by rhythmic stimulation of a motor nerve.
Neirofizi%giia 18, 645-54.
Fatigue at the Neuromuscular Junction 99

Gutmann E, Hanlikova V & Vyskocil F (1971). Age changes in cross-striated muscle of the rat. Journal of
Physiology (London) 216, 331-343.
Hatt H & Smith DO (1976). Synaptic depression related to presynaptic axon conduction block. Journal of
Physiology (London) 259, 367-393.
Helgren ME, Squinto SP, Davis HL, Parry DJ, Boulton TG, Heck CS, Zhu Y, Yancopoulos GD, Lindsay RM
& DiStefano PS (1994). Trophic effect of ciliary neurotrophic factor on denervated skeletal muscle.
Cell 76, 493-504.
Johnson BD & Sieck GC (1993). Differential susceptibility of diaphragm muscle fibers to neuromuscular
transmission failure. Journal ofApplied Physiology 75, 341-348.
Jones SF & Kwanbunbumpen S (1970). Some effects of nerve stimulation and hemicholinium on quantal
transmitter release at the mammalian neuromuscular junction. Journal of Physiology (London) 207,
51-61.
Juel C (1988). Muscle action potential propagation velocity changes during activity. Muscle Nerve 11, 714-719.
Kandel ER (1981). Calcium and the control of synaptic strength by learning. Nature 293, 697-700.
Katz B & Thesleff S (1957). On the factors which determine the amplitude of the "miniature end-plate
potential". Journal ofPhysiology (London) 137,267-278.
Kelly AM & Zacks SI (1969). The fine structure of motor endplate histogenesis. Journal of Cell Biology 42,
154-169.
Kelly SS (1978). The effect of age on neuromuscular transmission. Journal ofPhysiology (London) 274, 51-62.
Kelly SS & Roberts DV (1977). The effect of age on the safety factor in neuromuscular transmission in the
isolated diaphragm of the rat. British Journal ofAnaesthesiology 49,217-22.
Kernell D (1965). The adaptation and the relation between discharge frequencies and current strength of cat
lumbosacral motoneurons stimulated by long-lasting injected current. Acta Physiologica Scandi-
navica 65, 65-73.
Kernell D & Monster AW (1982a). Time course and properties oflate adaptation in spinal motoneurons of the
cat. Experimental Brain Research 46, 191-196.
Kernell D & Monster AW (1982b). Motoneuron properties and motor fatigue. An intracellular study of
gastrocnemius motoneurons of the cat. Experimental Brain Research 46, 197-204.
Khodorov BI, Timin YN, Vilenkin Y & Gul'ko FB (1969). Theoretical analysis of the mechanisms of
conduction of a nerve pulse over an inhomogeneous axon. I. [Link] through a portion with
increased diameter. Biojizika 14, 304-315.
Kmjevic K & Miledi R (1958). Failure of neuromuscular transmission in rats. Journal ofPhYSiology (London)
140,440-461.
Kmjevic K & Miledi R (1959). Presynaptic failure of neuromuscular propagation in rats. Journal ofPhysiology
(London) 149, 1-22.
Kuei JH, Shadmehr R & Sieck GC (1990). Relative contribution of neuromuscular transmission failure to
diaphragm fatigue. Journal ofApplied Physiology 68, 174-180.
Kugelberg E & Lindegren B (1979). Transmission and contraction fatigue of rat motor units in relation to
succinate dehydrogenase activity of motor unit fibers. Journal ofPhysiology (London) 288, 285-300.
Kurihara T & Brooks IE (1975). The mechanism of neuromuscular fatigue: A study of mammalian muscle
using excitation-contraction coupling. Archives ofNeurology 32, 168-174.
Lentz TL & Chester J (1982). Synaptic vesicle recycling at the neuromuscular junction in the presence of a
presynaptic membrane marker. Neuroscience 7, 9-20.
Llinas RR & Heuser IE (1977). Depolarization-release coupling systems in neurons. Neuroscience Research
Program Bulletin 15, 555-687.
McKenzie DK, Bigland-Ritchie B, Gorman RB & Gandevia SC (1992). Central and peripheral fatigue of
human diaphragm and limb muscles assessed by twitch interpolation. Journal ofPhysiology (London)
454, 643-656.
Meech RW (1974). The sensitivity of Helix aspersa neurones to injected calcium ions. Journal of Physiology
(London) 237, 259-278.
Merton PA (1954). Voluntary strength and fatigue. Journal of Physiology (London) 123, 553-564.
Metzger JM & Fitts RH (1986). Fatigue from high- and low-frequency muscle stimulation: role of sarcolemma
action potentials. Experimental Neurology 93, 320-333.
Miyata H, Zhan WZ, Prakash YS & Sieck GC (1994). Influence of inactivity on contribution of neuromuscular
transmission failure to diaphragm fatigue. Medicine and Science in Sports and Exercise 26, S167.
Naess K & Storm-Mathisen A (1955). Fatigue of sustained tetanic contractions. Acta Physiologica Scandi-
navica 34, 351-366.
Nystrom B (1968). Postnatal development ofmotornerve terminals in "slow-red" and "fast-white" cat muscles.
Acta Neurologica Scandinavica 44, 363-383.
100 G. C. Sieck and Y. S. Prakash

Okano K, Monck JR & Fernandez JM (1993). GTP gamma S stimulates exocytosis in patch-clamped rat
melanotrophs. Neuron 11, 165-72.
Parnas I (1972). Differential block at high frequency of branches of a single axon innervating two muscles.
Journal of Neurophysiology 35, 903-914.
Raymond SA & Lettvin JA (1978). Aftereffects of activity in peripheral axons as a clue to nervous encoding.
In: Waxman SG (ed.), Physiology and Pathobiology ofAxons, pp. 203-225. New York: Raven Press.
Redfern PA (1970). Neuromuscular transmission in newborn rats. Journal of Physiology (London) 209,
701-709.
Renaud JM (1989). The effect of lactate on intracellular pH and force recovery of fatigued sartorius muscles
of the frog, Rana pipiens. Journal of Physiology (London) 416, 21-47.
Renaud JM & Light P (1992). Effects of K+ on the twitch and tetanic contraction in the sartorius muscle of
the frog, Rana pipiens. Implication for fatigue in vivo. Canadian Journal of Physiology and
Pharmacology 70,1236-1246.
Roed A (1988). Fatigue during continuous 20 Hz stimulation of the rat phrenic nerve diaphragm preparation.
Acta Physiologica Scandinavica 134, 217-21.
Rosenblueth A (1940). The electrical excitability of mammalian striated muscle. American Journal of
Physiology 129, 22-38.
Sakmann B, Patlak J & Neher E (1980). Single acetylcholine-activated channels show burst-kinetics in
presence of desensitizing concentrations of agonist. Nature 286, 71-73.
Sandercock TG, Faulkner JA, Albers JW & Abbrecht PH (1985). Single motor unit and fiber action potentials
during fatigue. Journal ofApplied Physiology 58, 1073-1079.
Schiller Y & Rahamimoff R (1989). Neuromuscular transmission in diabetes: response to high frequency
activation. Journal of Neuroscience 9,3709-3719.
Sieck GC & Fournier M (1990). Changes in diaphragm motor unit EMG during fatigue. Journal of Applied
Physiology 68, 1917-1926.
Sieck GC & Fournier M (1991). Developmental aspects of diaphragm muscle cells. In: Haddad GG, Farber
JP (eds.), Developmental Neurobiology of Breathing, pp. 375-478. New York: Dekker.
Sieck GC, Fournier M & Enad JG (1989). Fiber type composition of muscle units in the cat diaphragm.
Neuroscience Letters 97, 29-34.
Sj0gaard G (1991). Role of exercise-induced potassium fluxes underlying muscle fatigue: a brief review.
Canadian Journal of Physiology and Pharmacology 69, 238-45.
Smith DO (1980). Mechanisms of action potential propagation failure at sites of axon branching in the crayfish.
Journal of Physiology (London) 301, 243-259.
Smith DO (1983). Axon conduction failure under in vivo conditions in crayfish. Journal of Physiology
(London) 344, 327-333.
Smith DO (1984). Acetylcholine storage, release and leakage at the neuromuscular junction of mature adult
and aged rats. Journal of Physiology (London) 347, 161-76.
Stephens JA & Taylor A (1972). Fatigue of maintained voluntary muscle contractions in man. Journal of
Physiology (London) 220, 1-18.
Stockbridge N (1988). Differential conduction at axonal bifurcations. II. Theoretical basis. Journal of
Neurophysiology 59, 1286-1295.
Stockbridge N & Stockbridge LL (1988). Differential conduction at axonal bifurcations. I. Effect of electro-
tonic length. Journal of Neurophysiology 59, 1277-1285.
Swadlow HA, Kocsis JD & Waxman SG (1980). Modulation of impulse conduction along the axonal tree.
Annual Review of Biophysics and Bioengineering 9, 143-179.
Thesleff S (1959). Motor end-plate "desensitization" by repetitive nerve stimuli. Journal of Physiology
(London) 148, 659-664.
Tsujimoto T & Kuno M (1988). Calcitonin gene-related peptide prevents disuse-induced sprouting of rat motor
nerve terminals. Journal of Neuroscience 8, 3951-3957.
Volle RL & Branisteanu DD (1976). Quantal parameters of transmitter release at the frog neuromuscular
junction. Naunyn Schmiedebergs Archives of Pharmacology 295, 103-8.
Waxman SG (1975). Integrative properties and design principles ofaxons. International Review of Neurobi-
ology 18, 1-40.
Wernig A, Jans H & Zucker H (1986). A parametric study of the neuromuscular junction during ontogenesis
and under different external conditions. In: Katz B, Rahamimoff R (eds.), Calcium, Neuronal
Function and Transmitter Release, pp. 413-430. Boston: Martinus NijhoffPublishers.
Woods JJ, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal ofNeurophysiology 58,125-137.
 6

THE ROLE OF THE SARCOLEMMA ACTION

POTENTIAL IN FATIGUE

A. J. Fuglevand

The John B. Pierce Laboratory

New Haven, Connecticut 06519

ABSTRACT

A prevalent feature of neuromuscular fatigue is a decline in the extracellularly recorded

myoelectric signal. One factor that could underlie this change is a decrease in the amplitude
of the sarcolemmal action potential. Based on observed reductions in action potential
amplitude without effect on force, it has been argued that changes in the action potential
during sustained activity would be unlikely to contribute to fatigue. However, those obser-
vations were primarily from experiments in which 1) high frequency stimulation may have
caused signal cancellation due to action potential overlap; or 2) sustained membrane
depolarization may have directly activated excitation-contraction coupling. The relatively
low and narrow range of membrane depolarization required for full activation of amphibian
and slow-twitch mammalian fibers makes them resistant to incomplete activation if action
potentials are depressed during fatigue. Mammalian fast-twitch fibers, on the other hand,
require greater depolarization for full activation and also exhibit a greater decrease in action
potential amplitude with fatigue. Therefore, it seems probable that fatigue-related decline in
action potential amplitude in these fibers leads to incomplete activation and loss of force.

INTRODUCTION

The loss of force of skeletal muscle during sustained activity is often accompanied
by a decline in the extracellular or surface-detected electromyographic signal. This occurs
under a variety of experimental situations: during maximal voluntary contraction of human
quadriceps (Bigland-Ritchie et aI., 1983); in electrically evoked responses of human first
dorsal interosseous following submaximal voluntary contractions (Fuglevand et aI., 1993);
during intermittent stimulation (brief 40 Hz trains) of cat extensor digitorum longus (Enoka
et aI., 1989); and during continuous 80 Hz stimulation offast-twitch motor units of cat medial
gastrocnemius (Clamann & Robinson, 1985). The fatigue-related fall in EMG could repre-
sent diminished excitation of muscle, or it might reflect ancillary adaptations that have little
impact on force development. The interpretation of the EMG decline, therefore, seems
crucial for understanding the mechanisms of neuromuscular fatigue.

101
102 A. J. Fuglevand

The decline in EMG responses to nerve stimulation could be due to impairment of

transmission at axonal branch points or at the neuromuscular junction (see Sieck, Chapter
5). In addition, the decrease in EMG during sustained voluntary contraction could indicate
a diminished output from the motomeuron pool. These possibilities clearly represent a
decrease in muscle excitation. Whether fatigue-related diminution of muscle excitation
directly promotes force loss, however, is an unresolved question which depends on how the
input-output properties of motor units adapt during different fatigue tasks (Enoka & Stuart,
1992).
Another factor that may underlie the decline in EMG is a reduction in the amplitude
of the sarcolemmal action potential. Intra- and extracellular recordings of single muscle
fibers have shown the sarcolemmal action potential to decrease during prolonged activity
(Sandercock et aI., 1985; Uinnergren & Westerblad, 1986; Metzger & Fitts, 1986; Radicheva
et aI., 1986; Balog et aI., 1994). Activity-associated alteration in the transmembrane distri-
bution of electrolytes (Hirche et aI., 1980; Vyskocil et ai., 1983; Sj0gaard, 1990, 1991;
Fujimoto & Nishizono, 1993) can cause progressive depolarization of the resting membrane
potential (Kwiecinski et aI., 1984; Juel; 1986; Lindinger & Heigenhauser, 1991) which slows
(Juel, 1988) and decreases the amplitude of the action potential (Jones, 1981; Jones &
Bigland-Ritchie, 1986; Uinnergren & Westerblad, 1986). If membrane depolarization is
large enough, action potentials may fail to propagate altogether (Sandercock et aI., 1985;
Juel, 1988; Balog et aI., 1994). Whether this occurs in voluntary contraction, however, has
been extremely difficult to establish. Also, it is unclear if a diminished sarcolemmal action
potential per se could fail to fully actuate the voltage-sensor calcium-release system of the
t-tubule. This review explores the possibility that fatigue-related changes in the sarcolemmal
action potential may contribute to diminished muscle excitation.

SIGNAL CANCELLATION OF EXTRACELLULAR POTENTIALS

WITH HIGH-FREQUENCY STIMULATION

Some evidence indicates that reduction in action potential amplitude is not associated
with a loss in muscle force. Action potentials recorded extracellularly in frog muscle fibers
decline significantly with stimulation frequencies ~ 50 Hz without a concomitant decrease
in force (Uittgau, 1965). Similarly, in slow-twitch motor units of the cat hindlimb, there is
a substantial depression ofEMG responses with continuous 80 Hz stimulation prior to any
reduction in force (Clamann & Robinson, 1985). These observations suggest that action
potential size can vary over a large range with little effect on excitation-contraction coupling
and force. Other studies, however, have shown that during intermittent stimulation with 40
Hz trains, reduction of action potential amplitude is accompanied with decline in force in
mammalian fast-twitch fatigable motor units and muscle with a high proportion offast-twitch
fibers (Reinking et ai., 1975; Gardiner & Olha, 1987; Enoka et ai., 1989). This discrepancy
may be partly related to species and fiber-type differences in the relation between membrane
potential and activation (described in a later section) and may also be an effect of extracellular
recording of action potentials when stimulating at high rates.
Action potential amplitude decays steeply with increased distance between active
fiber and extracellular electrode (Gath & StiHberg, 1978; Albers et ai., 1989; Fuglevand et
aI., 1992). In addition, the duration of the biphasic action potential, as detected by extracel-
lular electrodes, increases with electrode-fiber distance (Fuglevand et aI., 1992). Therefore,
when stimulating at high frequencies and recording EMG responses with extracellular or
surface electrodes, it is possible for the positive phase of one action potential to overlap in
Role of the Sarcolemma Action Potential in Fatigue 103

Stimulus Rate ~1 40HZ

.tv
Stimulus I O.25V

EMG 1 0.05 mV

20ms

Force 1 40mN

1s

Figure 1. Example of electromyographic signal cancellation due to action potential overlap. Inset: five
superimposed surface-detected EMG responses to intraneural stimulation (-1 Hz) ofa single motor unit in
human flexor digitorum superficialis muscle. The same unit was activated by a train of stimuli in which
frequency increased from 5 to 80 Hz and then returned to 5 Hz. For frequencies above about 45 Hz, there was
a progressive decay in EMG amplitude which was restored when frequency decreased. This pattern of change
in the EMG was likely caused by progressive signal cancellation due to increasing overlap in biphasic action
potentials with a duration of about 22 ms (from unpublished data of Macefield, Fuglevand and Bigland-
Ritchie).

time with the negative phase of an adjacent potential. This causes signal cancellation and
reduces the magnitude of the detected EMG response.
An example of this effect is presented in Fig. 1. A single motor unit of human flexor
digitorum superficialis was stimulated via an intraneural microelectrode with a train in which
frequency continuously increased from 5 to 80 Hz and then decreased back to 5 Hz. As would
be anticipated for a biphasic motor unit potential with a duration of 22 ms (inset, Fig. 1)
there was a progressive decline in the action potential amplitude due to overlap as frequency
increased above 45 Hz. The immediate and essentially symmetric recovery of action
potential amplitude as frequency declined indicates that the depressed action potential was
an artifact of signal cancellation associated with high-frequency stimulation.
A consistent feature of fatigue is a broadening of the action potential associated
with slowing of propagation velocity (Juel, 1988; Lindstrom et aI., 1977). An increase
in the duration of the action potential usually precedes reduction in amplitude or can
occur in the absence of change in amplitude (Bigland-Ritchie et aI., 1982, Duchateau &
Hainaut, 1987, Kranz et aI., 1983). Therefore, the progressive decline in action potential
amplitude seen when stimulating at relatively high frequencies (Clamann & Robinson,
1985; Liittgau, 1965) could be due partly to a gradual widening of the action potential
leading to greater temporal overlap and signal cancellation of the EMG. A step decrease
in amplitude between the first and second action potentials in high-frequency trains
(Liittgau, 1965; his Figs. 2, 3, 7, 8, 9, 12) and a nearly instantaneous recovery of the
action potential upon cessation of high frequency stimulation (Liittgau, 1965, his Fig. 5;
Clamann & Robinson, 1985, their Fig. 6) are indicative of some degree of overlap
cancellation. Therefore, to identify the physiological association between EMG and force
in fatigue, it is important to ensure that interstimulus intervals are longer than the duration
of the extracellularly detected action potential.
104 A. J. Fuglevand

NON-FATIGUING MANIPULATION OF ACTION POTENTIAL AND

ITS EFFECT ON FORCE

The resting membrane potential can be experimentally controlled by varying the ionic
composition of the extracellular medium in which isolated muscle or muscle fibers are bathed
(Hodgkin & Horowicz, 1960; Dulhunty, 1980; Kwiecinski et aI., 1984). It is then possible
to investigate the relation between action potential magnitude and mechanical response
independent of muscle activity. A significant decline in action potential amplitude has been
described for amphibian and mammalian muscle fibers in low sodium or elevated potassium
solutions (Sandow, 1952; Jones, 1981; Jones & Bigland-Ritchie, 1986; Liinnergren &
Westerblad, 1986) with no decrement in twitch force (Sandow, 1952; Uinnergren & Wester-
blad, 1986). It was proposed, therefore, that the sarcolemmal action potential operates as a
trigger with a wide margin of safety (Sandow, 1952).
The interpretation of these experiments, however, is complicated by the direct effect
of sustained membrane depolarization on excitation-contraction coupling. Muscle fibers
exposed to high potassium develop force in the absence of propagated action potentials
(Hodgkin & Horowicz, 1960; Liittgau, 1963; Caputo, 1972) and the magnitude of the force
contracture can be as large or larger than that obtained with tetanic stimulation (Hodgkin &
Horowicz, 1960; Dulhunty & Gage, 1985). High extracellular potassium depolarizes the
t-tubular membrane which directly activates the voltage sensor and induces sustained
calcium release from the sarcoplasmic reticulum (Dulhunty, 1992). Thus, although experi-
mental depolarization of the membrane diminishes the magnitude of the action potential
transient, it can at the same time activate excitation-contraction coupling directly. Mainte-
nance, indeed potentiation (Sandow, 1952; Chapman, 1969; Uinnergren & Westerblad,
1986) of the force response under these conditions, therefore, may be less a demonstration
of an action potential safety factor and more a reflection of direct activation of the t-tubular
membrane.

MUSCLE ACTIVATION BY STEADY STATE DEPOLARIZATION

The difficulty associated with experimental control of the transient action potential
has prompted investigations in which the membrane potential is held constant by adjusting
the concentration of extracellular potassium or by voltage clamp (Caputo & de Bolanos,
1979; Dulhunty, 1992). A significant property of muscle activation revealed in these
experiments is the dependence of developed force on membrane potential. Membrane
depolarization by 30 - 40 mV from the resting potential is required to initiate a just-detectable
mechanical response in amphibian and slow twitch mammalian fibers (Dulhunty, 1982;
Hodgkin & Horowicz, 1960). To elicit a mechanical response in mammalian fast twitch fibers
requires an additional 10-15 mV depolarization (Dulhunty, 1980; Dulhunty & Gage, 1983).
Also, the threshold for mechanical response increases as the duration of the depolarization
decreases (Dulhunty, 1980, 1982). Thus, further depolarization of about 10mV is needed to
attain mechanical threshold for brief pulses of duration similar to that of an action potential.
Beyond threshold depolarization, force increases as a sigmoid function of membrane
potential (Hodgkin & Horowicz, 1960; Luttgau & Oetliker, 1968; Dulhunty, 1980; Garcia
et aI., 1991). The membrane potential required to achieve half-maximum force under
steady-state depolarization is about -48 mV for amphibian muscle (frog semitendinosus,
Hodgkin & Horowicz, 1960), -25 mV for mammalian slow twitch (rat soleus, Dulhunty &
Gage, 1983, 1985), and -10mV for mammalian fast-twitch muscle fibers (rat extensor
digitorum longus, Dulhunty & Gage, 1983; 1985). Maximal force is attained at steady-state
Role of the Sarcolemma Action Potential in Fatigue 105

peak of action potential

fatigue fresh
I I
1.0
~ +

0.8
Q)
~
0 0.6
U.
Q)
.2:
10 0.4
CD
a:
0.2

0.0
-80 -60 -40 -20 o 20 40
Membrane Potential (mV)

Figure 2. Relation between steady-state membrane potential and force for different muscles. The solid arrow
indicates the approximate overshoot amplitude of action potentials in unfatigued muscle fibers. With fatigue,
the action potential amplitude may decline (dashed arrow) which would likely have minimal effect on frog or
mammalian slow-twitch fibers but could cause a significant decrement in force in mammalian fast-twitch
fibers (frog semitendinosus curve estimated from data of Hodgkin and Horowicz, 1960; rat soleus and extensor
digitorum longus (edl) curves from Dulhunty and Gage, 1985).

membrane potentials of about -25 mV for amphibian, -10m V for mammalian slow twitch,
and +15 mV for mammalian fast twitch fibers (Fig. 2). For brief, action potential-like
depolarization it is probable that these curves would be shifted toward slightly more positive
potentials (Heistracher & Hunt, 1969).
Although not yet fully identified, the mechanisms that underlie the marked depend-
ence of force on membrane potential seem related to the activation of the voltage-sensors in
the t-tubule. The movement of charged elements within the t-tubu1ar membrane appears to
provoke calcium release from the sarcoplasmic reticulum (Schneider & Chandler, 1973;
Kovacs et aI., 1979). The magnitude of calcium release depends on the number of charged
elements transposed to activating positions in the membrane (Chandler et aI., 1975). The
movement of this intramembranous charge has been quantified and, like force, is sigmoidally
related to membrane potential (Schneider & Chandler, 1973; Kovacs et aI., 1979; Dulhunty
& Gage, 1983).
For present purposes, two features of the data represented in Fig. 2 warrant particular
attention. First, activation of muscle depends critically on the magnitude (and duration) of
the membrane depolarization; and, therefore, does not operate as a simple trigger or
all-or-nothing phenomenon. Second, the relation between developed force and membrane
potential varies with fiber type, muscle, and species. In fresh muscle fibers, the overshoot
of the intracellular action potential is to about + 20 mV (solid arrow, Fig. 2) and does not
differ substantially across fiber types or muscles (19.5 mV in frog semitendinosus, Balog et
aI., 1994; 25.8 mV in frog sartorius, Light et aI., 1994; 20.1 mV in rat soleus, 25.8 mV in
rat extensor carpi radialis, Hanson, 1974; 23.0 mVin rat diaphragm, Metzger & Fitts, 1986;
16.8 mV in human intercostal, Kwiecinski et aI., 1984). Action potentials of this magnitude
should be more than adequate to fully activate amphibian and slow-twitch mammalian fibers
while just sufficient to completely activate mammalian fast-twitch fibers (Fig. 2).
Many studies have shown the intracellular action potential to decline as a conse-
quence of sustained activity (Hanson & Persson, 1971; Benzanilla et aI., 1972; Hanson, 1974;
Metzger & Fitts, 1986; Radicheva et aI., 1986; Westerblad & Liinnegren, 1986; Liinnergren
106 A. J. Fuglevand

& Westerblad, 1986, 1987; Balog et aI., 1994). The reduction can be large enough so as to
eliminate the overshoot and thus the peak of the action potential is negative (Benzanilla et
aI., 1972; Uinnergren & Westerblad, 1986, 1987). The magnitude of the depression appears
to depend on the pattern of stimulation; continuous high-frequency stimulation causes a rapid
and marked decline in the intracellular action potential (e.g., peak amplitude to -28 mV in
14 s with 70 Hz stimulation, Liinnergren & Westerblad, 1986) while less of a change in
amplitude occurs with intermittent high-frequency stimulation (e.g., peak amplitude to +7.3
m V with 100 ms trains at 150 Hz every 1 s for 5 min, Balog et aI., 1994).
Although less well documented, there is a greater decline in the intracellular action
potential in fast-twitch as compared to slow-twitch fibers (Hanson, 1974). Extracellular record-
ings of muscle and motor unit action potentials provide additional indirect support for fiber-type
differences in susceptibility for action potential decline. Fast-twitch fatigable motor units and
muscles comprised of a high proportion offast-twitch fibers exhibit a substantial decline in action
potential amplitude during prolonged activity whereas little change is seen in fatigue-resistant
motor units and muscles (Reinking et al., 1975; Kugelberg & Lindegren, 1979; Sandercock et aI.,
1985; Gardiner & Olha, 1987; EnokaetaI., 1989; Hanun et aI., 1989; Larsson etaI., 1991).
The crucial issue is whether a reduction in the action potential would impair fiber
activation. If the peak of the intracellular action potential declined to -10mV during sustained
activity (dashed arrow, Fig. 2) this would have little or no effect on force production in amphibian
or manunalian slow-twitch fibers. It would, however, significantly reduce force for manunalian
fast-twitch fibers. Fast-twitch fibers, therefore, appear susceptible to incomplete activation, not
only because of the more positive position of their activation curves, but also because of their
greater propensity for depressed action potentials (as described above). Furthermore, fatigue
shifts the steady-state activation curves toward more positive potentials (by about 10mV, Liittgau
& Oetliker, 1968; Garcia et aI., 1991) thereby enhancing the probability of diminished activation.
Finally, Pagala and colleagues (1994) compared changes in electrically evoked
tetanic force to potassium contracture force in mouse extensor digitorum longus following
3 min of intermittent stimulation with 30 Hz trains. Tetanic force declined by 50% whereas
there was no reduction in potassium contracture force. These findings suggest that the
reduction in force with electrical stimulation was due to an inadequate activation of the
t-tubular membrane because direct depolarization of that membrane with high potassium
revealed no deficit in force. Those observations are consistent with the proposal that
fatigue-related depression of the sarcolemmal action potential in fast twitch muscle fibers
may contribute to incomplete activation and force loss.

ACKNOWLEDGMENTS

The author wishes to thank Dr. Michael Walsh for helpful comments on the manu-
script. Some of the work was supported by United States Public Health Service grants NS
14576 and HL 30062 (to Brenda Bigland-Ritchie). Attendance of the author at the 1994
Bigland-Ritchie conference was supported, in part, by the University of Miami.

REFERENCES
Albers BA, Put JHM, Wallinga W & Wirtz P (1989). Quantitative analysis of single muscle fibre action
potentials recorded at known distances. Electroencephalography and Clinical Neurophysiology 73,
245-253.
Balog EM, Thompson LV & Fitts RH (1994). Role of sarcolemma action potentials and excitability in muscle
fatigue. Journal ofApplied Physiology 76, 2157-2162.
Role of the Sarcolemma Action Potential in Fatigue 107

Benzanilla F, Caputo C, Gonzalez-Serratos H & Venosa RA (1972). Sodium dependence of the inward spread of
activation in isolated twitch muscle fibres of the frog. Journal ofPhysiology (London) 223, 507-523.
Bigland-Ritchie B, Johansson R, Lippold OCJ & Woods JJ (1983). Contractile speed and EMG changes during
fatigue of sustained maximal voluntary contractions. Journal ofNeurophysiology 50, 313-324.
Bigland-Ritchie B, Kukulka CG, Lippold OCJ & Woods JJ (1982). The absence of neuromuscular transmission
failure in sustained maximal voluntary contractions. Journal of Physiology (London) 330, 265-278.
Caputo C (1972). The time course of potassium contractures of single muscle fibres. Journal of Physiology
(London) 223, 483-505.
Caputo C & de Bolanos PF (1979). Membrane potential, contractile activation and relaxation rates in voltage
clamped short muscle fibres of the frog. Journal of Physiology (London) 289, 175-189.
Chandler WK, Schneider MF, Rakowski RF & Adrian RH (1975). Charge movements in skeletal muscle.
Philosophical Transactions of the Royal Society of London. Series B 270, 501-505.
Chapman JB (1969). Potentiating effect of potassium on skeletal muscle twitch. American Journal of
Physiology 217, 898-902.
Clamann HP & Robinson AJ (1985). A comparison of electromyographic and mechanical fatigue properties
in motor units of the cat hindlimb. Brain Research 327, 203-219.
Duchateau J & Hainaut K (1987). Electrical and mechanical failure during sustained and intermittent
contractions in humans. Journal of Applied Physiology 58,942-947.
Dulhunty AF (1980). Potassium contractures and mechanical activation in mammalian skeletal muscles.
Journal of Membrane Biology 57, 223-233.
Dulhunty AF (1982). Effects of membrane potential on mechanical activation in skeletal muscle. Journal of
General Physiology 79, 233-251.
Dulhunty AF (1992). The voltage-activation of contraction in skeletal muscle. Progress in Biophysics and
Molecular Biology 57,181-223.
Dulhunty AF & Gage PW (1983). Asymmetrical charge movement in slow- and fast-twitch mammalian muscle
fibres in normal and paraplegic rats. Journal of Physiology (London) 341, 213-231.
Dulhunty AF & Gage PW (1985). Excitation-contraction coupling and charge movement in denervated rat
extensor digitomm longus and soleus muscles. Journal of Physiology (London) 358, 75-89.
Enoka RM, Rankin LL, Stuart DG & Volz KA (1989). Fatigability of rat hindlimb muscle: associations between
electromyogram and force during a fatigue test. Journal of Physiology (London) 408, 251-270.
Enoka RM & Stuart DG (1992). Neurobiology of muscle fatigue. Journal of Applied Physiology 72,
1631-1648.
Fuglevand AJ, Winter DA, Patla AE & StashukD (1992). Detection of motor unit action potentials with surface
electrodes: influence of electrode size and spacing. Biological Cybernetics 67,143-153.
Fuglevand AJ, Zackowski KM, Huey KA & Enoka RM (1993). Impairment of neuromuscular propagation during
human fatiguing contractions at submaximal forces. Journal ofPhysiology (London) 460, 549-572.
Fujimoto T & Nishizono H (1993). Involvement of membrane excitation failure in fatigue induced by
intermittent submaximal voluntary contraction of the first dorsal interosseous muscle. Journal of
Sports Medicine and Physical Fitness 33, 107-117.
Garcia MDC, Gonzalez-Serratos H, Morgan JP, Perreault CL & Rozycka M (1991). Differential activation of
myofibrils during fatigue in phasic skeletal muscle cells. Journal ofMuscle Research and Cell Motility
12,412-424.
Gardiner PF & 01ha AE (1987). Contractile and electromyographic characteristics of rat plantaris motor unit
types during fatigue in situ. Journal of Physiology (London) 385, 13-34.
Gath I & Stalberg E (1978). The calculated radial decline of the extracellular action potential compared with
in situ measurements in the human brachial biceps. Electroencephalography and Clinical Neurophysi-
ology 44,547-552.
Hamm TM, Reinking RM & Stuart DG (1989). Electromyographic responses of mammalian motor units to a
fatigue test. Electromyography and Clinical Neurophysiology 29, 485-494.
Hanson J (1974). The effects of repetitive stimulation on the action potential and the twitch of rat muscle. Acta
Physiologica Scandinavica 90, 387-400.
Hanson J & Persson A (1971). Changes in the action potential and contraction of isolated frog muscle after
receptive stimulation. Acta Physiologica Scandinavica 81, 340-348.
Heistracher P & Hunt CC (1969). The relation of membrane changes to contraction in twitch muscle fibres.
Journal of Physiology (London) 201, 589-611.
Hirche H, Schumacher E & Hagemann H (1980). Extracellular K+ concentration and K+ balance of the
gastrocnemius of the dog during exercise. Pjliigers Archiv 387,231-237.
Hodgkin AL & Horowicz P (1960). Potassium contractures in single muscle fibres. Journal of Physiology
(London) 153, 386-403.
108 A. J. Fuglevand

Jones DA (1981). Muscle fatigue due to changes beyond the neuromuscular junction. In: Edwards RHT (ed.),
Human Muscle Fatigue: Physiological Mechanisms. Ciba Foundation Symposium, pp. 178-196.
London: Pitman Medical.
Jones DA & Bigland-Ritchie B (1986). Electrical and contractile changes in muscle fatigue. In: Saltin B (ed.),
International Series on Sport Sciences. Biochemistry of Exercise VI, pp. 377-392. Champaign IL:
Human Kinetics.
Juel C (1986). Potassium and sodium shifts during in vitro isometric muscle contraction, and the time course
of the ion-gradient recovery. Pjliigers Archives 406, 458-463.
Juel C (1988). Muscle action potential propagation velocity changes during activity. Muscle & Nerve 11, 714-719.
Kovacs L, Rios E & Schneider MF (1979). Calcium transients and intramembrane charge movement in skeletal
muscle fibres. Nature 279, 391-396.
Kranz H, Williams AW, Cassel J, Caddy DJ & Silberstein RB (1983). Factors determining the frequency content
of the electromyogram. Journal ofApplied Physiology 55,392-399.
Kugelberg E & Lindegren B (1979). Transmission and contraction fatigue of rat motor units in relation to
succinate dehydrogenase activity of motor unit fibers. Journal ofPhysiology (London) 288, 285-300.
Kwiecinski H, Lehmann-Hom F & Rudel R (1984). The resting membrane parameters of human intercostal
muscle at low, normal, and high extracellular potassium. Muscle & Nerve 7,60-65.
Liinnergren J & Westerblad H (1986). Force and membrane potential during and after fatiguing, continuous
high-frequency stimulation of single Xenopus muscle fibers. Acta Physiologica Scandinavica 128,
359-368.
Liinnergren J & Westerblad H (1987). Action potential fatigue in single skeletal muscle fibres of Xenopus.
Acta Physiologica Scandinavica 129, 311-318.
Larsson L, Edstrom L, Lindegren B, Gorza L & Schiaffino, S (1991). MHC composition and enzyme-histo-
chemical and physiological properties of a novel fast-twitch motor unit type. American Journal of
PhYSiology 261, C93-CI01.
Light PE, Comtois AS & Renaud 1M (1994). The effect of glibenclamide on frog skeletal muscle: evidence
for K+ ATP channel activation during fatigue. Journal of Physiology (London) 475, 495-507.
Lindinger MI & Heigenhauser GJF (1991). The roles of ion fluxes in skeletal muscle fatigue. Canadian Journal
ofPhysiology and Pharmacology 69, 246-253.
Lindstrom L, Kadefors R & Petersen I (1977). An electromyographic index for localized muscle fatigue.
Journal ofApplied Physiology 43, 750-754.
Ltittgau HC (1963). The action of calcium ions on potassium contractures of single muscle fibres. Journal of
PhYSiology (London) 168, 679-697.
Ltittgau HC (1965). The effect of metabolic inhibitors on the fatigue of the action potential in single muscle
fibres. Journal of Physiology (London) 178, 45-67.
Ltittgau HC & Oetliker H (1968). The action of caffeine on the activation of the contractile mechanism in
striated muscle fibres. Journal of Physiology (London) 194, 51-74.
Metzger JM & Fitts RH (1986). Fatigue from high- and low-frequency muscle stimulation: role of sarcolemma
action potentials. Experimental Neurology 93, 320-333.
Pagala M, Ravindran K, Amaladevi B, Namba T & Grob D (1994). Potassium and caffeine contractures of
mouse muscles before and after fatiguing stimulation. Muscle & Nerve 17, 852-859.
Radicheva N, Gerilovsky L & Gydikov A (1986). Changes in the muscle fibre extracellular action potentials
in long-lasting (fatiguing) activity. European Journal ofApplied Physiology 55, 545-552.
Reinking RM, Stephens JA & Stuart DG (1975). The motor units of cat medial gastrocnemius: problem of
their categorisation on the basis of mechanical properties. Experimental Brain Research 23, 301-313.
Sandercock TG, Faulkner JA, Albers JW & Abbrecht PH (1985). Single motor unit and fiber action potentials
during fatigue. Journal of Applied Physiology 58,1073-1079.
Sandow A (1952). Excitation-contraction coupling in muscular response. Yale Journal ofBiology and Medicine
25, 176-201.
Schneider MF & Chandler WK (1973). Voltage dependent charge movement in skeletal muscle: a possible
step in excitation-contraction coupling. Nature 242,244-246.
Sj0gaard G (1990). Exercise-induced muscle fatigue: the significance of potassium. Acta Physiologica
Scandinavica Supplement 593, 1-63.
Sj0gaard G (1991). Role of exercise-induced potassium fluxes underlying muscle fatigue: a brief review.
Canadian Journal of Physiology and Pharmacology 69, 238-245.
Vyskocil F, Hnik P, Rehfeldt H, Vejsada R & Ujec E (1983). The measurement ofK/ concentration changes
in human muscles during volitional contractions. Pjliigers Archiv 399,235-237.
Westerblad H & Liinnergren J (1986). Force and membrane potential during and after fatiguing, intermittent
tetanic stimulation of single Xenopus muscle fibers. Acta [Link] Scandinavica 128, 369-378.
 7

SINGLE FIBER ELECTROMYOGRAPHY IN

STUDIES OF NEUROMUSCULAR FUNCTION

J. V. Tronteljl and E. Stalberg2

1 University Institute of Clinical Neurophysiology

University Medical Center of Ljubljana, Ljubljana, Slovenia
2 Department of Clinical Neurophysiology
University Hospital
Uppsala, Sweden

ABSTRACT

Single-fiber electromyography (SFEMG) allows precise study of the microphysiology of the

human motor unit under normal conditions. The physiological parameters that can be
quantified include impulse transmission along the intramuscular axon collaterals, pre- and
post synaptic events at the neuromuscular junction, and muscle fiber membrane properties.
This chapter illustrates some of the advantages of SFEMG in studies of neuromuscular
fatigue in normal muscle, as well as in disorders of neuromuscular transmission, and
conditions associated with disturbed muscle fiber depolarization-repolarization.

INTRODUCTION

Single-fiber electromyography (SFEMG) is based on extracellular recordings of

single muscle fiber action potentials with a small electrode, 25 !Jm in diameter, contained in
a side port of a steel cannula which is inserted into muscle. The average uptake radius of the
electrode is about 300 !Jm. Owing to this high spatial resolution, single and multiple muscle
fiber action potentials of individual motor units can be reliably recognized and differentiated
from others, allowing recordings over prolonged periods of time and in different experimen-
tal circumstances. During repetitive discharges, the single fiber action potentials have a
well-defined and reproducible shape that justifies time measurements with an accuracy as
high as 0.1 !JS.
These recording advantages have been exploited to study morphological and func-
tional details of the motor unit, ranging from action potential parameters of single muscle
fibers to discharge characteristics of ventral hom cells and cortico-spinal tract axons.
Microstimulation of individual muscle fibers and motor axons, either within the muscle or
nerve trunk, has conferred additional advantages and enlarged the scope of research problems
that can be studied with this method. A detailed description of the SFEMG technique, as well

109
110 J. V. Trontelj and E. StlUberg

as a comprehensive review of its uses and findings in health and diseased muscle, has been
published (Stalberg & Trontelj, 1994)
In this chapter, we discuss some SFEMG findings pertinent to studies of muscle
fatigue. These include analysis of the safety of transmission in the axonal tree and across the
neuromuscular junction (NMJ) in ischemia and certain neuromuscular disorders, changes in
muscle fiber propagation velocity, and muscle fiber de- and repolarization disturbances.

TRANSMISSION SAFETY IN THE INTRAMUSCULAR NERVE

TREE

When recordings are made from three or more muscle fibers during voluntary
activation, SFEMG offers the opportunity to study transmission in the intramuscular axon-
collateral tree. In studies of axon reflexes in normal subjects, no failure of transmission is
detectable at the nodes of Ranvier. In conditions associated with peripheral reinnervation,
two or more components in a multi-unit recording may show simultaneous intermittent
blocking. These potentials also show a large concomitant jitter relative to other parts of the
action potential complex. This is usually due to an intermittent block in the common axonal
branch supplying those muscle fibers from which the blocked action potentials are recorded.
The concomitant jitter in relation to the non-blocking spikes in the multi-unit recording thus
results from unreliable propagation of the axonal impulse in this common branch (Stalberg
& Thiele, 1972).
As with the neuromuscular junction (NMJ) blocking encountered in myasthenia
gravis, such axonal blocking may increase during repetitive discharge, and with an increase
in stimulation rate. Both may produce a decrement in the surface-recorded EMG responses
to repetitive stimulation. Axonal blocking may also show a response to edrophonium; that
is, decreased jitter and less frequent blocking (Stalberg & Thiele. 1972) from which it follows
that the presence of decrement and a positive edrophonium effect are not absolute proof of
a synaptic transmission defect. Also, the blocking of even a single axonal impulse may result
in a prolonged reduction of tension produced by the axon's motor unit, due to an inverse
catch-like effect (Burke et al.. 1970).

TRANSMISSION SAFETY AT THE NORMAL NEUROMUSCULAR

JUNCTION

When a nerve fiber is stimulated repetitively at suprathreshold strength, and re-

sponses are recorded from a single muscle fiber, there is a latency variability in the order of
tens of microseconds. This is calledjitter (Ekstedt, 1964). Its main source in normal muscle
is at the NMJ. As discussed elsewhere (Stalberg & Trontelj, 1994), it is due largely to small
fluctuations in the firing threshold of single muscle fibers, resulting in a variable neuromus-
cular transmission time. Minor variations in the amplitude and slope of the end-plate
potential are also contributing factors.
In pathology, additional mechanisms may contribute to the jitter. For example, there
may be an abnormally low amplitude of the EPP due to decreased postsynaptic sensitivity,
exaggerated variability ofEPP amplitude due to impaired mechanism of transmitter release
and/or disintegrated EPP shape due to asynchronous transmitter release. Thus, a combination
of factors determines the amount of both normal and abnormal jitter.
Single Fiber Electromyography in Studies of Neuromuscular Function 111

Jitter lusl
150,-------------------------- - - - - -- - - - - - - ,
87 'high safety' NMJs
125

100

75

50

GJ [J GJ GJ
25

0
GJ
0.5 Hz 1 Hz 2 Hz 5 Hz 10 Hz 20 Hz
Stimulation rate

Jitter lusl
400 ~--~~------------------------------,

Figure 1. litter measurements at dif-

ferent stimulation rates in normal mus-
cle. Median (asterisk) and range (bar) 300
of values provided. Top: low jitter, high
safety NMJs; n = 87; jitter :0:; 20/-ls).
Bottom: highjitter,lowsafetyNMls; n 200
= 33; jitter> 20/-ls). Note increase in
jitter as the stimulation rate is raised
from 0.5 to 5 Hz, and a decrease bet- 100
ween 10 to 20 Hz (reprinted with per-
mission from Stalberg & Trontelj,
1994; their Fig. 7.14). 0 -'---r------r-----.,.------r----~-------,,..-~
0.5 Hz 1 Hz 2 Hz 5 Hz 10 Hz 20 Hz
Stimulation rate

Jitter can be studied in voluntarily activated muscle. In this case, the action potential
of another muscle fiber of the same motor unit serves as the time reference for its measure-
ment.

Changes in Jitter Related to Stimulation Rate

The effect of the degree of activity on jitter is of interest in normal muscle, and even
more so in cases of disturbed neuromuscular transmission. Most normal end-plates display
a rather uniform jitter (::; 20 I1s) at different stimulation rates within the physiological range.
We denote these as high safety factor endplates. In a recent study of 120 NMJ s in the extensor
digitorum communis muscle (EDC) of 10 normal adults, the mean jitter values showed
relatively little change when tested at different stimulation rates (Fig. 1). Jitter was slightly
but significantly smaller at the lowest stimulation rate (0.5 Hz), as compared with a range
of rates from 2 to 50 Hz. The majority of endplates exhibited this behavior. However, some
maintained remarkably uniform, rather low jitter throughout the whole range of stimulation
frequencies (2 to 50 Hz). Not all muscle fibers could follow stimulus frequencies> 15-20
Hz. In such instances, the amplitude of the single fiber action potential (SFAP) rapidly
diminished, and its rise time and latency increased. These changes progressed until the SFAP
could no longer be recognized. When the stimulation rate was subsequently reduced, the
SFAP reverted to its previous state. This reversible state is attributed to a progressive failure
and subsequent recovery of the electrogenic Na+ -K+ pump (McComas & Einhorn, 1990).
Failure of the this pump is probably responsible for much of the decrement of the electrical
112 J. V. Trontelj and E. Stilberg

response of a muscle to supramaximal stimulation of its nerve at rates ~ 50 Hz (Desmedt,

1973). However, in contrast to the situation with micro stimulation, supramaximal tetanic
activation of a muscle tends to produce ischemia and progressive hypoxia. These may
eventually compromise neuromuscular transmission.
Approximately 25% of randomly sampled end-plates in the Fig. I study showed
changes in jitter related to stimulation (i.e., activation) rate. Jitter at most stimulation rates
exceeded 20 ~s (low safety factor end-plates). The typical pattern was a progressive increase
in jitter to values> 30 ~s as the stimulation rate was increased from 0.5 to 5 Hz. The range
of jitter values decreased significantly at stimulation rates of 10, and particularly 20 Hz.
Conversely, end-plates with moderate jitter values at low rates only rarely showed a
significant increase (depression) at 20 Hz. Both patterns were more commonly seen in the
low safety normal (and moderately abnormal) NMJs, as well as in the blocked NMJs of
patients with myasthenia. The pattern exhibited by endplates with moderate jitter (no change)
was much more common than that which decreased at 20 Hz. At 50 Hz (not shown in Fig.
I), a substantial increase in jitter occurred in over two thirds of both low and high safety
end-plates of normal subjects. However, jitter still did not exceed the upper limit of normal
for 10 Hz (40 ~s) in any of the tested NMJs. This suggests that at these high rates, the
presynaptic terminals have a remarkable ability to cope with the increased demand for
acetylcholine (ACh) synthesis and mobilization.
A study of the safety factor (Schiller et aI., 1975) of individual NMJs using regional
curare showed that NMJs with higher jitter were more sensitive to curare than those with
lower jitter. From this and the present study, it can be concluded that the jitter value is a
reliable index of the safety factor of neuromuscular transmission.

Long-Term Recording
Continuous jitter recordings from a SFAP pair is possible for up to three hours at
stimulation rates between 10 and 15 Hz. In normal muscle there is little change in jitter
throughout this period. For stimulation rates> 30 Hz, jitter does not change appreciably for
at least 10 min, by which time the test muscle becomes fatigued.

NEUROMUSCULAR TRANSMISSION DURING ISCHEMIA

To study the effects of ischemia, Dahlbiick and colleagues (1970) applied a sphyg-
momanometer to the subject's arm above the elbow, and the cuff was inflated to 200 mm
Hg. The EDC muscle contracted weakly and SFEMG recordings were performed before,
during and after (release of cuff) ischemia. The cuff was deflated when total blocking of
one or the other action potential was evident. Following a few minutes of continuous activity
during ischemia, the jitter increased rapidly, with one or the other action potential blocked
intermittently; first rarely, and then more frequently until total block had occurred. The time
to blocking was shorter with high activation rates. Approximately 2000-4000 discharges
were required before the onset or blocking. After ischemia, the blocked action potentials
recovered quickly, and jitter became close to normal within a few minutes. If a second period
of ischemia was applied within I min, the time to impulse blocking was shorter than on the
first occasion. Thus, in ischemia, block of neuromuscular transmission precedes the failure
of depolarization/repolarization. This is in contrast to the rather normal jitter associated with
progressive SFAP decrement during high frequency microstimulation in the normally
perfused muscle.
Single Fiber Electromyography in Studies of Neuromuscular Function 113

NEUROMUSCULAR TRANSMISSION IN DISEASE

Myasthenia Gravis
The electrophysiological mechanisms underlying disturbed neuromuscular transmis-
sion in this disease have been well elucidated in in vitro studies (Elmqvist et ai., 1964;
Lambert et ai., 1976, Engel et ai., 1976). While the number of ACh quanta released from the
presynaptic terminal per nerve impulse (quantum content of the end-plate potential (EPP))
is normal, the EPP amplitude is reduced due to a deficiency of postsynaptic ACh receptors,
as well as to blocking of the receptors by IgG antibodies.
During activity, the deficiency and blocking of ACh receptors are partially counter-
acted by an increased mobilization rate of ACh vesicles in the presynaptic nerve terminai.
This results in a short lived potentiation (Magleby & Zengel, 1976), followed within seconds
to minutes by an even more pronounced postactivation depression of the EPP. In a repetitive
nerve stimulation test, the size of the EPP response represents the net result of these two
opposing processes.
The typical SFEMG findings in a patient with myasthenia (StiHberg et ai., 1971;
Stalberg et ai., 1974; Stalberg et ai., 1976; Sanders, 1987) include: 1) normal jitter values;
2) jitter values above the normal range but without impulse blocking; and 3) as dependent
on the severity of the disease, recordings with increased jitter and intermittent impulse
blocking. The latter usually first appears in association with a jitter near 100 !ls (Fig. 2). Of
445 patients with myasthenia, the EDC muscle was abnormal in 99% of those with moderate
or severe generalized disease and in 75% of the patients in clinical remission (StaIberg &
Trontelj,1994).
Jitter that is initially abnormal may increase during continuous activity particularly
at increasing stimulation rates. However, when the stimulation rate subsequently decreases,
the jitter and degree of blocking also decrease. In contrast, jitter that is initially normal
changes little during activity. Sometimes a decrease injitter and blocking is seen at increasing
activation rates, and, occasionally, some muscle fibers belonging to the motor unit under
study may become activated whereas previously, they had been in a state of persistent block.

A B c

1 ms

Figure 2. SFEMG jitter recordings from the EDC muscle of a patient with myasthenia gravis. The oscilloscope
sweep is triggered by the first action potential and the jitter is the interval variability between the two single
fiber action potentials. Upper traces show individual sweeps and the lowest trace is the superimposed sweeps.
A, normal jitter; B, increased jitter but no impulse blocking; C, increased jitter and occasional blocking. In the
lower part, the action potential discharges are superimposed (reprinted from Trontelj & StAlberg, 1991 by
permission of John Wiley & Sons, Inc.).
114 J. V. Trontelj and E. Stalberg

Changes in Neuromuscular Transmission Efficiency Related to Changes

in Activation Rate
These changes can be conveniently studied at the individual end-plates by use of
SFEMG measurements. Valuable additional information, not on the affected postsynaptic
apparatus, but rather on the function of the nerve terminal, may be provided. Indeed,
myasthenia may be regarded as a valuable model to study the normal function of the motor
nerve terminal.
A recent study (Trontelj & StAlberg, 1991; Fig. 3A) in 10 myasthenic patients
observed the response of 58 NMJs to different stimulation rates. The lowest stimulation rates,
0.5 and 1.0 Hz, produced the lowest jitter and incidence of blocking which increased at
stimulation rates of 2, 5, and 10 Hz. However, increasing the stimulation rate to 20 Hz
reduced the amount of jitter and blocking to the level observed at the low stimulation rates.
This improvement is considered to be due to an increase in ACh released per stimulus; i.e.,
intratetanic potentiation (Fig. 3).
Functionally, the most important part of the stimulation frequency spectrum investi-
gated lies between 10 and 20 Hz. Most of the activity of motor units in the limb muscles
occurs in this range during sustained contractions involving submaximal effort (Stalberg et
aI., 1976). The most significant intratetanic potentiation is expected to occur at these
frequencies and it follows that the prominent potentiation observed at the NMJs with large
jitter might be an adaptive presynaptic mechanism to safeguard function of an affected or a
normal low safety NMJ.
Jitter [Link])
4 0 0 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - -- -- - - - _ .

300

200

100

o~_,----_.------._----._----._----,_~

0.5 Hz 1 Hz 2 Hz 5 Hz 10 Hz 20 Hz
Stimulation rate

Jitter Ills)
400 ~~~~-----------------------------.

300

200 Figure 3. The effect of intramuscular

electrical stimulation at different fre-
quencies. Median and range values as
100 in Fig. 1. Top: Jitter changes at 40
NMJs in myasthenia gravis. Bottom:
Jitter changes at 18 NMJs in Lambert-
o ~--~--------~- Eaton myasthenic syndrome (reprinted
2 Hz 5 Hz 10 Hz 15 Hz with permission from StaIberg & Tron-
Stimulation rate telj, 1994; their Figs. 15.30 and 15.33).
Single Fiber Electromyography in Studies of Neuromuscular Function 115

However, in a severely affected muscle, a proportion of end-plates may be unable

to follow a stimulation rate of 2 Hz or more since they develop a persistent block within
the first few stimuli (Trontelj & Stalberg, 1993). Also, a few NMJs in myasthenic muscle
demonstrated moderate jitter at low and intermediate rates with an increased jitter at 20
Hz. These suggest insufficient transmitter synthesis and/or mobilization at the higher
rate.

Lambert-Eaton Myasthenic Syndrome (LEMS)

This rare neuromuscular transmission disorder is often associated with broncho-
genic carcinoma (Eaton & Lambert, 1957). It is caused by an impaired neurotransmitter
release, resulting from an autoimmune attack on calcium channels in the presynaptic
nerve terminal (Lambert & Elmqvist, 1971). Even in clinically strong muscles, the jitter
is often grossly abnormal. The most striking effect is a dramatic reduction of jitter and,
in particular, blocking as the stimulation rate is increased from 1 - 2 Hz to 10 - 20 Hz
(Trontelj & Staib erg, 1991). Many NMJs (often up to one-third or more in a given muscle)
are in a state of complete conduction block. This is only overcome when the stimulation
rate is increased above 10-20 Hz. This seems to be the SFEMG hallmark of LEMS
(Trontelj & Staib erg, 1993).

PROPAGATION VELOCITY OF SINGLE MUSCLE FIBERS

Measurement of the propagation velocity of the single muscle fiber action potential
along the muscle fiber (St!ilberg, 1966) requires use of a multi-electrode recording with two
arrays of recording surfaces. To calculate the propagation velocity, the time interval is
measured between potentials recorded from the muscle fiber at the two recording sites. The
range of normal muscle fiber propagation velocity values lies between 1.5 and 6.5 mls. It
varies from muscle to muscle and even within a given muscle, the major determinant being
the diameter of the muscle fiber (Hakansson, 1956).

Propagation Velocity during Repetitive Activity

Propagation velocity usually decreases during the first minute of continuous single
fiber activation (Stalberg, 1966; Fig. 4 ). This decrease can amount to 50% of the initial value
over a period of 3 min. When the stimulation period is interrupted, the amount the
propagation velocity recovers depends on the length of the interruption.
When using concentric needle EMG recording (Lindstrom et aI., 1977), such a
decrease in propagation velocity explains the change in power spectrum observed when the
higher stimulation frequencies decrease, and the lower frequencies increase during continu-
ous activity. It also explains, in part, changes in the number of turns in analysis of the EMG's
interference pattern (Fuglsang-Frederiksen & Ronager, 1988). Using arrays of surface
electrodes, it is possible to study the mean conduction velocity of fibers in individual motor
units and to determine the direction of action potential propagation (Masuda & Sadoyama,
1986). A specialized method for analyzing the surface EMG (Lindstrom et aI., 1977) provides
an indirect measure of the mean propagation velocity in the entire active muscle. This value
decreases during continuous activity. A similar decrease in velocity is seen during high
frequency stimulation of individual muscle fibers, in which an increase in latency is also
observed.
116 J. V. Trontelj and E. Stalberg

5.0

-;; 4.5
.........
E
'-'
>- 4.0
I-
U
g 3.5
w
>
Z 3.0
o
i=
oct
~ 2.5
Q..
o
g: 2.0

oTo ,
30
I i i
60 90 120
i i i
150 180 210
i i i
240 270 300
I
330
TIME (5)
Figure 4. Propagation velocity of five muscle fibers during repetitive stimulation. Lines connected by
asterisks show decrease in velocity offour fibers discharging at 11, 16, 12 and 9 discharges/s (top to bottom,
respectively). A fifth fiber (the two continuous lines) was tested at stimulation rates of6/s and 16/s. It exhibited
no decrease in propagation velocity (reprinted with permission from St41berg & Trontelj, 1994; their Fig. 10.2).

MUSCLE FIBER DEPOLARIZATION AND CONDUCTION IN

MYOTONIA
In myotonic disorders, muscle force may decline during ongoing contraction, par-
ticularly when started after a period of rest. This is most pronounced in the recessive
(Becker's) form of myotonia congenita, in which weakness may be more disturbing for the
patient than stiffness. Weakness is also seen in the dominant (Thomsen's) form.
Repetitive electrical stimulation (5 Hz) of the ulnar nerve in a patient with Becker's
type of myotonia congenita produced a decline of twitch force to zero within 5 s. Force
gradually recovered after 20 s of continuous stimulation (StAlberg & Trontelj, 1994). The
initial decline in force was associated with a decrement of the amplitude of the compound
muscle action potential to a very low value, but not to zero. This indicates, in part,
undisturbed neuromuscular transmission and preserved muscle fiber depolarization with
impulse conduction, but failing excitation-contraction coupling (Ricker & Meinck, 1972;
Buchthal & Rosenfalck, 1963). At a stimulation frequency of 15 Hz, paralysis occurred
within 3 s and lasted about 80 s. However, this happened only when the muscle was allowed
to rest for about 15 min before the stimulation period. If nerve stimulation was performed
after a voluntary contraction of 5 min, there was barely any decrement in the amplitude of
the muscle action potentials.
A stimulation SFEMG study was performed in 5 patients, 3 with the Thomsen type
and 2 with sporadic myotonia, with characteristics of the Becker type (Trontelj, 1986;
Trontelj et aI., 1987). Forty-five muscle fibers were studied at various stimulation rates. In
any individual, there was a large range in degree and slope of action potential decrement.
For most ofthe fibers in patients with Thomsen's type of myotonia congenita, stimulation
at 20 Hz produced a rather rapid decrement in amplitude to between 40 and 70% within the
first half second. In the two patients with probable Becker's type of myotonia congenita,
profound action potential decrements could be observed at relatively low stimulation
Single Fiber Electromyography in Studies of Neuromuscular Function 117

~---:#=2.5 Hz

1 Hz

5 Hz

5 Hz

-v~ J
f
~8E2.5HZ
-1.,'
(

--v \A.-
f
\IVL-
1 Hz 1 Hz
,,~
~Vl-
r
"\I~
'r ~lA- I

2ms

Figure 5. Profound single fiber action potential decrement in a patient with probable Becker's type of
myotonia congenita. Repeated trains of stimuli separated by short pauses of 2 s show successively smaller
decrement (related to the warm-up phenomenon) (reprinted with permission from Stalberg & Trontelj, 1994;
their Fig. 15.41).

frequencies. In several fibers, action potential amplitudes were reduced to zero after just
10-30 s of stimulation at 2 Hz. This degree of decrement was not commonly observed in the
three patients with the Thomsen type. When a series of stimulations were repeated at intervals
as short as 2-10 s, the decrement usually became progressively less prominent, thereby
indicating a warm-up effect (Fig. 5). However, upon occasion, the opposite behavior was
seen.
In addition to amplitude decrement, the SFEMG recordings in patients with Becker's
type of myotonia congenita showed prominent changes in action potential shape. A progres-
sively longer negative after-potential was followed, in some cases, by a reversal of action
 118 J. V. Trontelj and E. Stalberg

potential polarity just before its amplitude reduced into the noise level. These changes were
largely unrelated to muscle fiber conduction velocity. Decrements to less than 1% of the
original action potential amplitude are compatible with unimpeded conduction; but probably
not without unimpaired excitation-contraction coupling (StAlberg & Trontelj, 1994). SFEMG
also suggests transient focal block of conduction of the action potentials in some fibers,
which may provide an explanation for a part of a myotonic weakness.
Several abnormalities in ion channels in the sarcolemma may underlie the myotonic
stiffness and weakness in different myotonic disorders (Barchi, 1992). Exactly which of the
abnormal ion fluxes and concentrations on either side of the sarcolemma is responsible for
the extracellular action potential decrement and deformities described above cannot be
determined on the basis of these observations. Rather, one has to resort to techniques such
as single channel studies. However, SFEMG can serve as a useful method to monitor some
of these abnormalities in vivo and to assess various treatments (Lagueny et aI., 1994).

ACKNOWLEDGMENTS

This work was supported by the Ministry of Science and Technology of Slovenia
(grant P3-5285-0312), the Medical Research Council of Sweden (grant 135), and the
Commission of the European Communities (grant PL-93-1 033). Attendance of J. v.T. at the
1994 Bigland-Ritchie conference was supported, in part, by the United States Public Health
Service (National Center for Medical Rehabilitation Research, National Institute of Child
Health and Human Development), and the Muscular Dystrophy Association (USA).

REFERENCES
Barchi RL (1992). The nondystrophic myotonic syndromes. In: Rowland LP, DiMauro S (eds.), Myopathies,
Vinken PJ, Bruyn GW, Klawans RL (eds.), Handbook o/Clinical Neurology, Revised series 18,. vol.
62, pp. 261-286. Amsterdam: Elsevier Science Publishers.
Buchthal F & Rosenfalck P (1963). Electrophysiological aspects of myopathy with particular reference to
progressive muscular dystrophy. In: Bourne GH, Golarz MN (eds.), Muscular Dystrophy in Man and
Animals, pp. 193-262. New York: Rather.
Burke RE, Rudomin P & Zajac FEIII (1970). Catch property in single mammalian motor units. Science 168,
122-124.
Dahlbiick L-O, Ekstedt J & StiHberg E (1970). Ischemic effect on impulse transmission to muscle fibres in
man. Electroencephalography and Clinical Neurophysiology 29,579-591.
Desmedt JE (1973). The neuromuscular disorder in myasthenia gravis. I. Electrical and mechanical responses
to nerve stimulation in hand muscles. In: Desmedt JE (ed.), New Developments in Electromyography
and Clinical Neurophysiology, vol. 1, pp. 241-304. Basel: Karger.
Eaton LM & Lambert EH (1957). Electromyography and electric stimulation of nerves in diseases of motor
unit observations on the myasthenic syndrome associated with malignant tumors. Journal 0/Ameri-
can Medical Association 163, 1117-1124.
Ekstedt J (1964). Human single muscle fiber action potentials. Acta Physiologica Scandinavica 61 (suppl.
226), 1-96.
Elmqvist D, Hofmann WW, Kugelberg J & Quastel DM (1964). An electrophysiological investigation of
neuromuscular transmission in myasthenia gravis. Journal of Physiology (London) 174, 417-434.
Engel AG, Tsujihata M, Lambert JM & Lennon VA (1976). Experimental autoimmune myasthenia gravis; a
sequential and quantitative study of the neuromuscular junction ultrastructure and electrophysiologic
correlations. Journal ofNeurophathology and Experimental Neurology 35,569-587.
Fuglsang-Frederiksen A & Ronager J (1988). The motor unit fIring rate and the power spectrum ofEMG in
humans. Electroencephalography and Clinical Neurophysiology 70, 68-72.
Hakansson CH (1956). Conduction velocity and amplitude of the action potential as related to the circumfer-
ence in the isolated fibre of frog muscle. Acta Physiologica Scandinavica 37, 14-34.
Single Fiber Electromyography in Studies of Neuromuscular Function 119

Lagueny A, Marthan R, Sheuermans P, Le Collen P, Ferrer X, Julien J (1994). Single fiber EMG and spectral
analysis of surface EMG in myotonia congenita with and without transient weakness. Muscle & Nerve
17,248-250.
Lambert EH & Elmqvist D. (1971). Quantal components of the end-plate potential in the myasthenic syndrome.
Annals of the New York Academy of Sciences 183, 183-199.
Lambert EH, Lindstrom JM & Lennon VA (1976). End-plate potentials in experimental autoimmune myasthe-
nia gravis in rats. Annals of the New York Academy of Sciences 274, 300-318.
Lindstrom L, Kadefors R & Petersen I (1977). An electromyographic index for localised muscle fatigue.
Journal ofApplied PhYSiology 43, 750-754.
Magleby KI & Zengel JE (1976). Long term changes in augmentation, potentiation and depression of
transmitter release as a function of repeated synaptic activity at the frog neuromuscular junction.
Journal of PhYSiology (London) 257, 471-494.
Masuda T & Sadoyama T (1986). The propagation of single motor unit potentials detected by a surface
electrode array. Electroencephalography and Clinical Neurophysiology 63, 590-598.
McComas AJ & Einhorn RW (1990). Electric responses of human muscles to prolonged repetitive stimulation.
Muscle & Nerve 13, 879.
Ricker K & Meinck HM (1972). Muscular paralysis in myotonia congenita. European Neurology 7, 221-227.
Sanders DB (1987). The electrodiagnosis of myasthenia gravis. Annals ofNew York Academy ofSciences 505,
539-556.
Schiller HH, Stalberg E & Schwartz MS (1975). Regional curare for the reduction of the safety factor in human
mot end-plates studied with single fibre electromyography. Journal ofNeurology, Neurosurgery and
Psychiatry 38, 805-809.
Stalberg E (1966). Propagation velocity in single human muscle fibres in situ. Acta Physiologica Scandinavica
70 (supp!. 287),1-112.
Stalberg E, Ekstedt J & Broman A (1971). The electromyographic jitter in normal human muscles. Elec-
troencephalography and Clinical Neurophysiology 31, 429-438.
Stalberg E, Ekstedt J & Broman A (1974). Neuromuscular transmission in myasthenia gravis studied with
single fibre electromyography. Journal ofNeurology, Neurosurgery and Psychiatry 37,540-547.
StaIberg E & Thiele B (1972). Transmission block in terminal nerve twigs: a single fibre electromyographic
finding in man. Journal of Neurology, Neurosurgery and Psychiatry 35, 52-59.
Stalberg E & Trontelj N (1994). Single Fiber Electromyography. Studies in Healthy and Diseased Muscle.
New York: Raven Press.
Stalberg E, Trontelj N & Schwartz MS (1976). Single-muscle-fiber recording of the jitter phenomenon in
patients with myasthenia gravis and in members of their families. Annals of the New York Academy
of Sciences 274, 189-202.
Trontelj N (1986). Intramuscular stimulation: Technique and findings in the normal and abnormal motor unit.
Muscle & Nerve 9, S83.
Trontelj JV, Mihelin M & Stalberg E (1987). Extracellularly recorded single muscle fibre responses to electrical
stimulation in myotonia congenita. Electroencephalography and Clinical Neurophysiology 66, S 106.
Trontelj N & Stalberg E (1991). Single motor end-plates in myasthenia gravis and LEMS at different firing
rates. Muscle & Nerve 14, 226-232.
Trontelj JV & Stalberg E (1993). Stimulation SFEMG: A stimulation rate of 2 Hz might be inadequate to find
the recording position (a reply). Muscle & Nerve 16, 564-565.
SECTION III
Fatigue of Single Motor Units

A motor unit is no more than a collection of relatively homogeneous muscle fibers

innervated by a single motoneuron. The force exerted by a single motor unit depends not
only on the state of neuromuscular transmission and the intracellular environment but also
on the discharge properties of the motoneuron. This section evaluates the behavior of the
interface between the CNS and the peripheral motor. This perspective is important because
it defines the constraints imposed on the CNS "upstream" of the motoneuron pool.
A functional match is presumed to exist between the electrical properties of mo-
toneurons and the mechanical behavior of the muscle fibers they innervate. The intrinsic
properties of motoneurons that affect their recruitment and repetitive firing characteristics
are reviewed in Chapter 8 (Sawczuk and colleagues). Particular attention is given to factors
affecting the threshold current required to activate motoneurons of different types and those
that influence the duration of the motoneuron's after-hyperpolarization. This duration
appears to parallel the duration of the twitch force produced by the motor unit. Thus, both
durations are long in the least fatigable, slow twitch, type S motor units and their mo-
toneurons but brief in fatigable, fast-twitch, type F motor units and motoneurons. Clearly,
factors that control the excitability of a motoneuron and its after-hyperpolarization are
important for understanding how it will behave under the influence of the powerful net
excitatory drive during strong contractions. The biophysical properties that underlie the early
and late adaptation of motoneuron-discharge rate to sustained input are discussed. Because
the adaptation is greater in type F than type S units, these properties may be tuned to optimize
the tension-firing frequency relation in individual motoneurons. Furthermore, the extent to
which a motoneuron may take up one or more stable states under different tasks is currently
the subject of much investigation.
In Chapter 9, Kernell continues to explore how the motoneuron's discharge to a
constant input current matches the response of its muscle fibers to different discharge rates.
He reviews evidence that activity-dependent mechanisms at the level of both the motoneuron
(transformation of input current to discharge rate) and the muscle fiber (transformation of
discharge rate to force) may be designed to minimize fatigue and to minimize the central
command (or effort) required for force production by a muscle or motor unit. The sigmoidal
relation between isometric force and discharge rate may be distorted by the exact pattern of
activation of the motor unit leading to a mixture of both potentiation (i.e. an increase in force)
and fatigue (i.e. a reduction in force). Another recurrent theme surfaces here, namely the
simultaneous mix of exercise-induced, intra- and extracellular factors, some of which
increase force and others that decrease it.
122 Fatigue of Single Motor Units

In the next two chapters (Chapter 10, Thomas; Chapter 11, Elek and Dengler), the
concepts developed earlier in the section are translated into studies of the properties of human
motor units. Several methods are available to study single motor units in humans and, not
surprisingly, none is perfect. Thomas reviews the data obtained with spike-triggered aver-
aging and the different forms of motor axonal stimulation. The analysis highlights the
differences between the techniques, the mutability of the force-frequency relationship for
single motor units, some apparent differences in fatigability between human motor units and
those in other species, and how little is known about the contractile properties of human
motor units during normal use. Elek and Dengler compare spike-triggered averaging and
intramuscular stimulation for the first dorsal interosseous muscle of human subjects. Data
are reported for patients with hereditary motor and sensory neuropathy (type 1) and
amyotrophic lateral sclerosis. Predictably, values for twitch contraction times and relaxation
times are shorter when assessed with spike-triggered averaging. The suitability of the two
methods for use in patients is discussed, with the yield being higher with spike-triggered
averaging. However, in patients in whom voluntary contractions for long periods are
precluded or who may have increased motor unit synchronization, intramuscular stimulation
may be preferable.
 8

INTRINSIC PROPERTIES OF MOTONEURONS

Implications for Muscle Fatigue

A. Sawczuk,l R. K. Powers,2 and M. D. Binder2

I Department of Oral Biology

2 Department of Physiology and Biophysics
University of Washington
Seattle, Washington 98195

ABSTRACT
The following is a brief review of the intrinsic properties ofmotoneurons that contribute to
their recruitment and rate modulation. Our emphasis is on properties that may either
accelerate or delay the onset of muscular fatigue. In general, intrinsic motoneuron properties
are regulated in a way that minimizes energy expenditure. The correlation of recruitment
threshold with motoneuron type ensures that the most fatigable motor units are reserved for
the most forceful contractions. The variation in minimum firing rates arising from variations
in AHP characteristics ensures that motoneurons begin to fire at rates that are matched to the
force producing characteristics of their muscle units. Further, it is possible that spike-fre-
quency adaptation contributes to optimization of the tension (force)-firing frequency (T-f)
transform of individual motor units.

INTRODUCTION

The notion of a precise, functional match between the intrinsic electrical properties
of motoneurons and the mechanical properties of the muscle fibers they innervate was
formalized by Henneman and colleagues as the size principle of motor unit recruitment
(Henneman et aI., 1965; Henneman & Olson, 1965). One advantage of size-ordered recruit-
ment arises from the inverse relation between motor unit size and fatigue-resistance, so that
larger, fatigable motor units are normally reserved for brief, forceful contractions (Henneman
& Olson, 1965). Subsequent research has confirmed the generality of this size-based
recruitment order and has provided insights into the underlying physiological mechanisms
(Burke, 1981; Henneman & Mendell, 1981; Heckman & Binder, 1990; Binder et aI., 1995).
Once recruited, the firing rate ofmotoneurons increases as a function of increasing depolar-
izing synaptic input. The intrinsic mechanisms controlling motoneuron spike-frequency are
also matched to motor unit mechanical properties, so that rate modulation normally occurs
over a range that is most efficacious in changing motor unit force (Kernell, 1983; Binder et
aI., 1995). Finally, both the spike-frequency of motoneurons and the mechanical output of

123
124 A. Sawczuk et al.

motor units depend upon their activation history (Binder et aI., 1995). These history-depend-
ent properties of motoneurons may also be matched to those of their motor units. The
following review will briefly cover the intrinsic properties of motoneurons that contribute
to their recruitment and rate modulation, with particular emphasis on those properties that
may be implicated in muscle fatigue.

INTRINSIC MOTONEURON PROPERTIES UNDERLYING

RECRUITMENT

Motoneurons are recruited when the somatic membrane potential is displaced from
its resting value (Vr) to the threshold value for initiating an action potential (Vthr ). The amount
of injected or synaptic current (rheobase; I rh ) needed to recruit a motoneuron will depend
upon the difference between the threshold and resting voltages, and upon the effective input
resistance of the cell (RN): Irh = (Vthr - Vr)lRN
The intrinsic properties important for recruitment are those determining V" Vthr and
RN Input resistances measured in cat lumbar motoneurons in vivo exhibit about a lO-fold
range from about 0.4 - 4.0 MQ (e.g., Gustafsson & Pinter, 1984b; Zengel et aI., 1985).
Average RN values are largest in type S, smaller in type FR, and smallest in type FF
motoneurons (Fleshman et aI., 1981; Zengel et aI., 1985). The finding that the range of
variation of Irh exceeds that of RN may result from a tendency for Vthr to be lowest in low
rheobase, high input resistance (presumably type Sand FR) motoneurons (Gustafsson &
Pinter, 1984a; Carp, 1992).
Once the voltage threshold for spike initiation is exceeded, action potentials arise
from the influx of sodium at the initial segment of the motoneuron. Initial segment sodium
currents are first observed at voltages 10mV above the resting potential, whereas somatic
sodium currents are not observed until the membrane is depolarized to 20 m V above rest
and are maximal at 30 mV of depolarization (Schwindt & Crill, 1982). The somatic voltage
at which the initial segment spike is elicited (Vthr ) is not fixed, but can exhibit both rapid
(Calvin, 1974) and slow fluctuations (Schwindt & Crill, 1982) during repetitive discharge.
This behavior is known as accommodation, and is thought to be due, in part, to subthreshold
changes in the amount of sodium channel inactivation (Frankenhaeuser & Vallbo, 1964;
Vallbo, 1964; Schlue et aI., 1974). Experiments on cat motor axons suggest inactivation time
constants in the range of2 - 4 ms (Richter et aI., 1974), but additional slower inactivation
processes may underlie the slow increase in firing level observed during repetitive discharge
(Schwindt & Crill, 1982). Variations in Vthr among different motoneurons may reflect
variations in the relative density of initial segment sodium channels or perhaps a difference
in their accommodative properties (Burke & Nelson, 1971).
The motoneuron action potential in vertebrates is followed by a prolonged afterhy-
perpolarization (AHP), lasting 50-200 ms. This process has been called the medium-duration
AHP (mAHP; Nishimura et ai, 1989; Viana et aI., 1993) to distinguish it from a distinct rapid
hyperpolarization often observed in the same motoneurons (Hounsgaard et aI., 1988;
Nishimura et aI., 1989; Viana et aI., 1993), and a much slower process observed in other
types of neurons (cf. Nishimura et aI., 1989). There is now abundant supportive evidence
that this medium-duration AHP is mediated by an apamin-sensitive, calcium-activated
potassium conductance (G KCa ; Krnjevic et aI., 1979; Zhang & Krnjevic, 1987; Hounsgaard
et aI., 1988; Mosfeldt-Laursen & Rekling, 1989; Nishimura et aI., 1989; Viana et aI., 1993;
Chandler et aI., 1994). The magnitude and duration of the AHP vary according to motoneuron
size (Zwaagstra & Kernell, 1980) and motor unit type: larger and longer AHPs are recorded
in type S than in type F motoneurons (Kernell, 1983; Zengel et aI., 1985). It is not known
Intrinsic Properties of Motoneurons 125

whether these differences in AHP characteristics reflect differences in the properties or

density of G KCa channels or differences in the various factors controlling the time course of
the calcium concentration in the vicinity of the channels.

REPETITIVE FIRING

If the magnitude of injected or synaptic current is increased above that needed to

elicit a single action potential, repetitive discharge ensues. Subsequent firing rate modulation
depends upon the intrinsic relation between injected or synaptic current (I) and the spike
.frequency (f), which is linear over a wide range of currents: f= 10 + ({II) • (1- 10), where 10 is
the minimum current needed to elicit steady repetitive discharge (generally about 1.5 times
the current needed to produce a single action potential, Irh ; Kernell, 1965c );10 is the minimum
firing rate; and PI is the slope of the frequency-current relation. In some motoneurons, the
steady-state relation between spike frequency and injected current (j-I) can be characterized
by a single linear equation over the entire range of applied currents, whereas in others, the
steady-state f-I relation consists of two or three linear segments of different slope (primary,
secondary and tertiary ranges of firing; cf. Schwindt & Crill, 1984).
The conductance underlying the AHP is one of the primary determinants of the
steady-state repetitive discharge properties ofmotoneurons. The minimum steady spike-fre-
quency is approximately equal to the reciprocal of AHP duration (Kernell, 1965c; 10dkowski
et aI., 1988) and the steady-state frequencies attained at the end of the primary and secondary
ranges of firing are also correlated with the reciprocal of AHP duration (Kernell, 1965c; see
also Kernell, Chapter 9, Fig. 1). Finally, comparisons of steady-state f-I relations before and
after pharmacological manipulations of the AHP conductance (Hounsgaard et aI., 1988;
Nishimura et aI., 1989; Viana et aI., 1993; Chandler et aI., 1994), indicate that the steady-state
f-I slope is inversely proportional to the magnitude of the AHP conductance.
A variety of evidence suggests that two other intrinsic properties make important
contributions to motoneuron spike-frequency behavior 1) a low threshold, persistent inward
current; and 2) variations in Vthr resulting from the accommodative properties of the initial
segment. The persistent inward current was first identified in cat lumbar motoneurons based
on the clamp current records during slow depolarizing voltage-clamp ramp commands
(Schwindt & Crill, 1977). The steady-state current-voltage (I-V) curve exhibits an N-shape
starting with a region of negative slope conductance at about 10-20 m V above the resting
potential, which subsequently reverses to a positive slope due to increasing activation of
outward conductances. Voltage-clamp commands encompassing the region of negative slope
conductance reveal a slowly-activating inward current that is enhanced by external Ba++
(Schwindt & Crill, 1980a), and is not blocked by intracellular injection of lidocaine
derivatives that block the fast sodium currents (Schwindt & Crill, 1980a,b). Based on this
evidence, the persistent inward current in cat motoneurons is probably calcium mediated,
although in other motoneurons, a negative slope region in the I-V curve is probably due to
a persistent sodium current (Mosfeldt-Laursen & Rekling, 1989; Nishimura et aI., 1989;
Rekling, 1992; Chandler et aI., 1994).
The presence of a persistent inward current is likely to have important effects on
steady-state discharge behavior provided that the somatic membrane potential can traverse
voltage ranges over which this current is significantly activated. The increase in Vthr at
increasing rates of discharge allows this condition to occur (Schwindt & Crill, 1982; Powers,
1993). Although relatively little inward current activation may take place at the membrane
voltages traversed at the lowest spike-frequencies, the persistent inward current is likely to
be continuously activated throughout the interspike interval at higher firing rates. The
upward inflection in the steady-statef-I curve that occurs in some cells (i.e., secondary range
126 A. Sawczuk et al.

firing; Kernell, 1965b; Schwindt, 1973; Schwindt & Crill, 1982} may result from the
predominance of the inward current. Once activated, the inward current component may
continue to predominate even when the activating stimulus is withdrawn, leading to repeti-
tive discharge that outlasts the depolarizing synaptic or injected current. This switching
between a quiescent state and sustained repetitive firing in response to transient inputs has
been termed bistable firing behavior (cf. Hounsgaard and Kiehn, 1989). It is thought to
depend upon a reduction in the potassium conductance underlying the AHP, leading to a
change in the relative balance of inward and outward currents (Schwindt & Crill, 1980c;
Hounsgaard et aI., 1984; Hounsgaard & Kiehn, 1989).
A number of aspects of the current-to-frequency transduction vary systematically
across the motoneuron pool. The 10-fold variation in rheobase (/rh) is associated with a
similar range in the minimum current needed to elicit steady repetitive discharge (/0 Kernell
& Monster, 1981). Variations in theiI relations among different motoneuron types also arise
from type-related differences in AHP characteristics (vide supra). Due to their longer AHP
durations, type S motoneurons begin to discharge at significantly lower rates than type F
motoneurons (Kernell, 1979), and the steady-state frequencies attained at the end of the
primary and secondary ranges of firing are also lower in motoneurons with long AHP
durations (presumably type S motoneurons; Kernell, 1965c). In contrast, the primary and
secondary rangef-I slopes do not vary systematically across the motoneuron pool (Kernell,
1979). Nonetheless, the systematic differences between minimum and maximum primary
range spike-frequency are matched to differences in motor unit tension-frequency (T-j)
relations, so that the rate modulation within the primary range modulates force along the
steepest portion of the muscle unit's tension-frequency curve (Kernell, 1965c; Kernell,
1979).

SPIKE-FREQUENCY ADAPTATION

Motoneuron firing rate decreases as a function of time after the onset of a current
step in a process called spike-frequency adaptation. In motoneurons, adaptation has a rapid,
initial phase (Kernell, 1965a,b; Sawczuk et aI., 1995), followed by a slow, gradual decline
that continues throughout the duration of firing which may be several minutes (late adapta-
tion; Granit et aI., 1963; Kernell, 1965a; Kernell & Monster, 1982a; Spielmann et aI., 1993;
Sawczuk et aI., 1995). We have recently conducted a detailed, quantitative analysis of the
complete time course of spike-frequency adaptation in rat hypoglossal motoneurons (Saw-
czuk et aI., 1995). As illustrated in Fig. 1, the initial, rapid phase of adaptation is linear and
is complete within the first few interspike intervals. In 50% of the trials, the slow decline in
spike-frequency over a 60 s period could be characterized by an exponential function with
a single time constant. However, in the other 50% of trials, the slow decline in firing had
two components: an early phase of adaptation that was typically complete within the first 2
s of discharge; and a late phase of adaptation that continued for the duration of firing (Fig.
I). The pattern of adaptation displayed in Fig. 1 was quite typical, with more than 90% of
the total decrease in spike-frequency occurring in the first 2 s, about 6% in the 2 s to 26 s
period, and the remaining 4% occurring after 26 s. Although most of the spike-frequency
adaptation occurs within the first two seconds of discharge, there remains a further 40%
reduction in firing rate during the late phase of adaptation (i.e., the firing rate at 26 s is about
60% of that at 2 s). These results are similar to those reported in cat lumbar motoneurons
(Kernell & Monster, 1982a; Lindsay et aI., 1986). The magnitude of initial adaptation was
correlated with the initial spike-frequency (i.e., the reciprocal of the first interspike interval).
The magnitudes of the early and late phases of adaptation were correlated with the firing
frequency reached at the end of initial adaptation. In rat hypoglossal motoneurons, neither
Intrinsic Properties of Motoneurons 127

e.'iii' 0.0 0.5 1.0 1.5 2.0

l;-
c
• :::I Late
Tlm_(a)

I Adaptation

• ff

I
~

i
0 10 20 30 40 50 60
Tlme(s)

J! 41 ~!U ~illill. ~lliill

ii

o 0.1 0.2 1.9 2.0 2.1 29.9 30.0 30.1 59.9 60.0 60.1

Tlm_(a)

Figure 1. Three phases of spike-frequency adaptation during repetitive discharge of rat hypoglossal mo-
toneurons. An initial, rapid decline in firing frequency (Ii -It) was complete within three interspike intervals.
The subsequent decline in spike-frequency (Ii - fr) can be fit with the sum of two exponential functions. (As
illustrated in the upper right inset, a single exponential curve-fit failed to match the data fromlt to 750 ms. An
additional exponential function was required to match the first 2 s of discharge.) The initial adaptation (Ii -It)
was linear with slope of -2.2 x 103 Hz/s. The early adaptation had a time constant (tE) of 0.3 s, and the late
adaptation had a time constant (tL) of27.0 s. Beneath the spike-frequency-time curve are 200 ms samples of
repetitive discharge taken at four time periods during the trial. fi = initial firing frequency based on the first
interspike interval;!, = firing rate at transition between the initial, linear phase of adaptation and the subsequent
exponential decline in firing rate;ft= firing rate at conclusion of the trial (modified from Sawczuk et aI., 1995).

the magnitude nor the time course of the three phases were correlated with other membrane
properties such as input resistance, rheobase or repetitive firing threshold (Sawczuk et aI.,
1995; see, however, Kernell and Monster, 1982a; Spielmann et aI., 1993).

MECHANISMS UNDERLYING SPIKE-FREQUENCY ADAPTATION

Initial Adaptation
It has been previously shown that increases in AHP amplitude are well correlated
with initial adaptation in motoneurons (Kernell & Sjoholm, 1973; Jack et aI., 1975;
Ba1dissera et aI., 1978; Barrett et aI., 1980; Schwindt & Crill, 1982). Summation of AHP
currents could occur if the calcium concentration in the vicinity of the G KCa channels does
not return to resting levels at the end of the interspike intervals. Initial adaptation is also
associated with an increase in Vthr (Schwindt & Crill, 1982). As mentioned earlier, changes
128 A. Sawczuk et al.

in Vthr during repetitive discharge may result from subthreshold changes in the inactivation
of initial segment sodium channels. Although the fast component of sodium channel
inactivation has a time constant of < 5 ms (Hille, 1992), a much slower component has been
described in some excitable membranes (Brismar, 1977; Howe & Ritchie, 1992), suggesting
that sodium channel inactivation could be involved in all phases of spike-frequency adapta-
tion. Finally, depending upon its activation range and kinetics, a hyperpolarization-activated,
mixed cation current (Ih) could contribute to initial adaptation: If it were partly activated at
rest, it would provide an inward current that would increase spike-frequency at the onset of
a current step and then deactivate at the more depolarized voltages encountered during
maintained repetitive discharge (cf. Spain et aI., 1991).

Early and Late Adaptation

The slow decline in spike-frequency seen during the later phases of adaptation could
reflect a progressive increase in outward currents, a progressive decrease in inward currents, or
some combination of these two mechanisms. Slow increases in the calcium concentration near
the G KCa channels would prolong the increase in G KCa beyond the period of initial adaptation.
Such increases in calcium concentration might reflect the saturation of intracellular calcium
sequestering systems (Barrett et aI., 1980) or release of calcium from intracellular stores
triggered by a second messenger (Zhang, 1990) or by calcium itself (Sah & McLachlan, 1991).
Recent computer simulations (Powers, unpublished) suggest that the time course of early
adaptation is consistent with the saturation of an intracellular calcium buffer with slow kinetics
(cf. Sah, 1992). Alternative mechanisms underlying a progressive increase in outward current
include 1) the activation of an electrogenic Na+ -K+ pump (Sokolove & Cooke, 1971; Kernell
& Monster, 1982a; French, 1989); 2) activation of an M-current (Adams et aI., 1986; Marrion,
1993); and 3) activation of a Na+ -activated potassium current (Schwindt et aI., 1989). The first
mechanism is not likely to be responsible for late adaptation, since late adaptation has been
shown to be unaffected by blockade of the Na+-K+ pump with ouabain in rat hypoglossal
motoneurons (Sawczuk & Binder, 1992; Sawczuk, 1993). It is not yet known if significant M-
currents or Na+ - activated potassium currents are present in motoneurons.
A slow decrease in spike-frequency could also reflect a slow inactivation of inward
currents. Slow inactivation processes in initial segment and somatic sodium channels would
be expected to lead to a progressive increase in Vthr along with changes in spike shape (cf.
Schwindt & Crill, 1982; Lindsay et aI., 1986; Sawczuk, 1993). A progressive increase in Vthr
would act to lengthen interspike intervals even in the absence of changes in G KCa or other
potassium conductances, since a given rate of rise of membrane depolarization would take
longer to cross the spike threshold.
Finally, more complex interactions between inward and outward currents might under-
lie late adaptation. For example, reducing or eliminating G KCa by replacing external Ca++ with
Mn++ leads to an increase in spike-frequency for a given current level, and yet the magnitude
of late adaptation is increased (Sawczuk, 1993; Sawczuk & Binder, 1993). It is possible that
under these conditions, an increase in the mean level of membrane depolarization during the
interspike interval accelerates the development of sodium channel inactivation.

FUNCTIONAL SIGNIFICANCE OF SPIKE-FREQUENCY

ADAPTATION

Although spike-frequency adaptation is observed in many types of neurons, its function

remains largely specUlative. The most obvious functional consequence of spike-frequency
Intrinsic Properties of Motoneurons 129

A
•

10~----~----~------.-----.------r-----.
o 10 20 30 40 50 60
Time(s)

B
110
100
90
'N
:. 80
~ 70
c
GI
::J

- ! 60

.¥
GI 50
Q.
(I)
40
30
20
10
0.9 1
Current (nA)

Figure 2. The slope of the spike-frequency-current (/-1) relation of motoneurons changes during adaptation.
A, firing frequency vs. time curves obtained from a single rat hypoglossal motoneuron in response to three
different levels of injected current. The relation between firing rate and the level of injected current was
measured at five different times in the trials (circles:fi= initial firing rate; filled squares:fi= firing rate at the
transition from initial to the later phases of adaptation; open squares: 17.1.0-, = average rate from 0.7 to 1.0 s;
filled triangles: 13 Os = firing rate at 30 s; open triangles: 160, = firing rate at 60 s). B,J-I relations obtained from
trials illustrated in A. Firing rate was linearly related to current intensity in all cases, but the!-I slope varied
with time from 73.5 Hz/nAmeasured atfito 4.4 Hz/nA measured at 60 s (modified from Sawczuk et a\., 1995).

adaptation in motoneurons is its profound influence on their frequency-current relations (Granit et

ai., 1963; Kernell, 1965a; Sawczuk et al., 1995). The slope of the f-1 relation decreases
dramatically during initial adaptation (Granit et aI., 1963). The decrease inf-1 slope continues
throughout early adaptation (Kernell, 1965a), and a true steady-state value is not reached until
after the early phase of adaptation (Sawczuk et ai., 1995; their Fig. 2).
130 A. Sawczuk et al.

The time course of spike-frequency adaptation is remarkably consistent in repeated

trials (Sawczuk et aI., 1995) suggesting that adaptation may be ajingerprint of intrinsic
motoneuron behavior. As such, the pattern of spike-frequency adaptation may be correlated
with a number of other motor unit properties including fatigability. A correlation between
the extent of adaptation and motor unit type has been reported for cat lumbar motoneurons,
with the greatest amount of adaptation observed in the type FF units and the least in the type
S units (Kernell & Monster, 1982a; Spielmann et aI., 1993).
Since only a few, high frequency spikes may be needed to attain near maximal
force in some motor units in the cat, the initial phase of adaptation may prevent excessive
and wasteful discharge during short bursts of discharge (Kernell, 1983). Thus, adaptation
may function to optimize the tension (force)-frequency (T-j) transduction of individual
motor units. However, although this T-f optimization can be clearly demonstrated under
controlled, isometric conditions when single units are stimulated electrically (e.g., Burke
et aI., 1970; Zajac & Young, 1980), initial adaptation has not been a consistent finding
from recordings of cat motor unit activity during locomotion (Hoffer et aI., 1981;
Brownstone et aI., 1992).
The late phase of adaptation may be matched to the progressive increase in twitch
contraction time that occurs during prolonged contractions and allows motor units to
maintain a given force level with a progressively decreasing activation rate (Bigland-Ritchie
et aI., 1983a,b; Bigland-Ritchie & Woods, 1984; Botterman & Cope 1988). The match
between the progressive decrease in motoneuron spike-frequency and the increase in motor
unit contraction time has been suggested as a strategy to optimize force production in the
presence offatiguing conditions (Kernell & Monster, 1982b; Marsden et aI., 1983; Enoka
& Stuart, 1992; Spielmann et aI., 1993; see also Stuart & Callister, 1993). Although there is
evidence to suggest that the decline in motoneuron firing frequency is produced in part by
reflex inhibition (Bigland-Ritchie et aI., 1986; Woods et aI., 1987; Garland et aI., 1988; Stuart
& Callister, 1993) and lor a decline in excitatory synaptic input from muscle spindle afferents
(Macefield et aI., 1991), it is likely that late adaptation also makes a contribution (see,
however, Stuart & Callister, 1993).

SUMMARY

In general, intrinsic motoneuron properties are regulated in a way that minimizes

energy expenditure and muscle fatigue. The variation of 10 according to motoneuron type
ensures that the most fatigable motor units will be reserved for the most forceful contractions.
The variation in minimum spike-frequency ifo) arising from variations in AHP characteristics
ensures that motoneurons begin to fire at rates that are matched to the force-producing
characteristics of their muscle units. Further, variations in the three phases of spike-frequency
adaptation may contribute to optimization of the tension-firing frequency (T-j) transductions
of individual motor units during both phasic and sustained contractions.

ACKNOWLEDGMENTS

The experimental work in our laboratory is supported by United States Public Health
Service grants, DE 00161, NS 31925, NS 26840, and NS 01650. Attendance of A.S. and
MD.B. at the 1994 Bigland-Ritchie conference was supported, in part, by the University of
Miami.
Intrinsic Properties of Motoneurons 131

REFERENCES
Adams PR, Jones SW, Pennefather P, Brown DA, Koch C & Lancaster B (1986). Slow synaptic transmission
in frog sympathetic ganglia. Journal oJExperimental Biology 124, 259-285.
Baldissera F, Gustafsson B & Parmiggiani F (1978). Saturating summation of the afterhyperpolarization
conductance in spinal motoneurones: a mechanism for 'secondary range' repetitive fuing. Brain
Research 146, 69-82.
Barrett EF, Barrett IN & Crill WE (1980). Voltage sensitive outward currents in cat motoneurons. Journal of
Physiology (London) 304, 251-276.
Bigland-Ritchie B, Dawson NJ, Johansson RS & Lippold OC (1986). Reflex origin for the slowing of
motoneurone firing rates in fatigue of human voluntary contractions. Journal ofPhysiology (London)
379,451-459.
Bigland-Ritchie B, Johansson R, Lippold ~C, Smith S & Woods JJ (1983a). Changes in motoneurone firing
rates during sustained maximal voluntary contractions. Journal ofPhysiology (London) 340, 335-346.
Bigland-Ritchie B, Johansson R, Lippold OC & Woods JJ (1983b). Contractile speed and EMG changes during
fatigue of sustained maximal voluntary contractions. Journal oJNeurophysiology 50,313-24.
Bigland-Ritchie B & Woods JJ (1984). Changes in muscle contractile properties and neural control during
human muscular fatigue. Muscle & Nerve 7, 691-699.
Binder MD, Heckman CJ & Powers RK (1995). The physiological control of motoneuron activity. In: Rowell
LF, Shepherd JT (eds.), Smith JL (sec. ed.), Handbook ofPhysiology: Integration ojMotor, Respira-
tory and Metabolic Control During Exercise, sec. A, Neural Control of Movement, pp. 00-00.
Bethesda: American Physiological Society. In press
Botterman BR & Cope TC (1988). Motor-unit stimulation patterns during fatiguing contractions of constant
tension. Journal oJNeurophysiology 60, 1198-1214.
Brismar T (1977). Slow mechanism for sodium permeability inactivation in myelinated nerve fibre of Xenopus
laevis. Journal oj PhYSiology (London) 270, 283-297.
Brownstone RM, Jordan LM, Kriellaars DJ, Noga BR & Shefchyk SJ (1992). On the regulation of repetitive
firing in lumbar motoneurones during fictive locomotion in the cat. Experimental Brain Research 90,
441-55.
Burke RE (1981). Motor units: anatomy, physiology, and functional organization. In: Brookhart JM,
Mountcastle VB (eds.), Brooks VB (vol. ed.), Handbook oj Physiology, sec. 1, vol. IL pt 1, The
Nervous System: Motor Control., pp. 345-422. Bethesda, MD: American Physiological Society.
Burke RE, Dum RP, Fleshman JW, Glenn LL, Lev-Tov A, O'Donovon MJ & Pinter MJ (1982). An HRP study
of the relation between cell size and motor unit type in cat ankle extensor motoneurons. Journal oj
Comparative Neurology 209, 17-28.
Burke RE & Nelson PG (1971). Accommodation to current ramps in motoneurons of fast and slow twitch
motor units. International Journal oJNeuroscience 1, 347-356.
Burke RE, Rudomin P & Zajac FE (1970). Catch property in single mammalian motor units. Science 168,
122-124.
Calvin WH (1974). Three modes of repetitive firing and the role of threshold time course between spikes.
Brain Research 59, 341-346.
Carp JS (1992). Physiological properties of primate lumbar motoneurons. Journal of Neurophysiology 68,
1121-1132.
Chandler SH, Hsaio C, Inoue T & Goldberg LJ (1994). Electrophysiological properties of guinea pig trigeminal
motoneurons recorded in vitro. Journal oj Neurophysiology 71, 129-145.
Enoka RM & Stuart DG (1992). Neurobiology of muscle fatigue. Journal of Applied Physiology 72,
1631-1648.
Fleshman JW, Munson m, Sypert GW & Friedman WA (1981). Rheobase, input resistance, and motor-unit
type in medial gastrocnemius motoneurons in the cat. Journal of Neurophysiology 46, 1326-1338.
Frankenhaeuser B & Vallbo AB (1964). Accomodation in myelinated nerve fibres of Xenopus laevis as
computed on the basis of voltage clamp data. Acta Physiologica Scandinavica 63, 1-20.
French AS (1989). Ouabain selectively affects the slow component of sensory adaptation in an insect
mechanoreceptor. Brain Research 504, 112-114.
Garland SJ, Gamer SH & McComas AJ (1988). Reduced voluntary electromyographic activity after fatiguing
stimulation of human muscle. Journal ofPhysiology (London) 401, 547-556.
Granit R, Kernell D & Shortess GK (1963). Quantitative aspects of repetitive firing of mammalian mo-
toneurones, caused by injected currents. Journal ojPhysiology (London) 168, 911-931.
Gustafsson B & Pinter MJ (1984a). An investigation of threshold properties among cat spinal alpha-mo-
toneurones. Journal ofPhysiology (London) 357, 453-83.
 132 A. Sawczuk et al.

Gustafsson B & Pinter MJ (l984b). Relations among passive electrical properties of lumbar alpha-mo-
toneurones of the cat. Journal of Physiology (London) 356, 401-31.
Heckman CJ & Binder MD (1990). Neural mechanisms underlying the orderly recruitment ofmotoneurons.
In: Binder MD, Mendell LM (eds.), The Segmental Motor System, pp. 182-204. New York: Oxford
University Press.
Henneman E & Mendell LM (1981). Functional organization of motoneuron pool and its inputs. In: Brookhart
1M, Mountcastle VB (sec. eds.), Brooks VB (vol. ed.), Handbook of Physiology, sec. 1, vol. II, pt 1,
The Nervous System: Motor Control., pp. 423-507. Bethesda, MD: American Physiological Society.
Henneman E & Olson CB (1965). Relations between structure and function in the design of skeletal muscle.
Journal ofNeurophysiology 28, 581-598.
Henneman E, Somjen G & Carpenter DO (1965). Excitability and inhibitability of motoneurons of different
sizes. Journal of Neurophysiology 28, 599-620.
Hille B (1992). Ionic Channels of Excitable Membranes (2nd Ed.). Sunderland, MA: Sinauer Assoc. Inc.
Hoffer JA, O'Donovan MJ, Pratt CA & Loeb GE (1981). Discharge patterns of hindlimb motoneurons during
normal cat locomotion. Science 213, 466-7.
Hounsgaard J, Hultborn H, Jespersen B & Kiehn 0 (1984). Intrinsic membrane properties causing a bistable
behaviour of alpha-motoneurones. Experimental Brain Research 55, 391-394.
Hounsgaard J & Kiehn 0 (1989). Serotonin-induced bistability of turtle motoneurones caused by a nifedip-
ine-sensitive calcium plateau potential. Journal of Physiology (London) 414, 265-282.
Hounsgaard J, Kiehn 0 & Mintz I (1988). Response properties ofmotoneurones in a slice preparation of the
turtle spinal cord. Journal ofPhysiology (London) 398, 575-89.
Howe JR & Ritchie 1M (1992). Multiple kinetic components of sodium channel inactivation in rabbit Schwann
cells. Journal of PhYSiology (London) 455, 529-566.
Jack JIB, Noble D & Tsien RW (1975). Electric Current Flow in Excitable Cells. Oxford: Clarendon Press.
Jodkowski JS, Viana F, Dick TE & Berger AJ (1988). Repetitive fIring properties of phrenic motoneurons in
the cat. Journal of Neurophysiology 60,687-702.
Kernell D (1965a). The adaptation and the relation between discharge frequency and current strength of cat
lumbosacral motoneurones stimulated by long-lasting injected currents. Acta Physiologica Scandi-
navica 65, 65-73.
Kernell D (1965b). High frequency repetitive fIring of cat lumbosacral motoneurones stimulated by long-last-
ing injected currents. Acta Physiologica Scandinavica 65, 74-86.
Kernell D (1965c). The limits of fIring frequency in cat lumbosacral motoneurones possessing different time
course of afierhyperpolarization. Acta Physio!ogica Scandinavica 65, 87-100.
Kernen D (1979). Rhythmic properties of motoneurones innervating muscle fIbres of different speed in m.
gastrocnemius medialis of the cat. Brain Research 160, 159-62.
Kernen D (1983). Functional properties of spinal motoneurons and gradation of muscle force. In: Desmedt JE
(ed.), Motor Control Mechanisms in Health and Disease, pp. 213-226. New York: Raven Press.
Kernell D & Monster AW (1981). Threshold current for repetitive impulse fIring in motoneurones innervating
muscle fIbres of different fatigue sensitivity in the cat. Brain Research 229, 193-196.
Kernen D & Monster AW (1982a). Time course and properties oflate adaptation in spinal motoneurones of
the cat. Experimental Brain Research 46,191-196.
Kernen D & Monster AW (1982b). Motoneurone properties and motor fatigue. An intracellular study of
gastrocnemius motoneurones of the cat. Experimental Brain Research 46, 197-204.
Kernell D & Sjoholm H (1973). Repetitive impulse fIring: comparisons between neurone models based on
'voltage clamp equations' and spinal motoneurones. Acta Physiologica Scandinavica 87, 40-56.
Kmjevic K, Lamour Y, MacDonald IF & Nistri A (1979). Effects of some divalent cations on motoneurones
in cats. Canadian Journal of Physiology and Pharmacology 57, 944-56.
Lindsay A, Heckman CJ & Binder MD (1986). Analysis of late adaptation in cat motoneurons. Society for
Neuroscience Abstracts 12, 247.
MacefIeld G, Hagbarth KE, Gorman R, Gandevia SC & Burke D (1991). Decline in Spindle Support to
alpha-Motoneurones During Sustained Voluntary Contractions. Journal ofPhysiology (London) 440,
497-512.
Marrion NV (1993). Selective reduction of one mode M-channel gating by muscarine in sympathetic neurons.
Neuron 11, 77-84.
Marsden CD, Meadows JC & Merton PA (1983). 'Muscular Wisdom' that minimizes fatigue during prolonged
effort in man: Peak rates of motoneuron discharge and slowing of discharge during fatigue. In:
Desmedt IE (eds.), Motor Control Mechanisms in Health and Disease, pp. 169-211. New York: Raven
Press.
Intrinsic Properties of Motoneurons 133

Mosfeldt-Laursen A & Rekling IC (1989). Electrophysiological properties of hypoglossal motoneurons of

guinea-pigs studied in vitro. Neuroscience 30,619-637.
Nishimura Y, Schwindt PC & Crill WE (1989). Electrical properties of facial motoneurons in brainstem slices
from guinea pig. Brain Research 502, 127-142.
Powers RK (1993). A variable-threshold motoneuron model that incorporates time-dependent and voltage-de-
pendent potassium and calcium conductances. Journal ofNeurophysiology 70, 246-262.
Powers RK & Binder MD (1985). Distribution of oligosynaptic group I input to the cat medial gastrocnemius
motoneuron pool. Journal of Neurophysiology 53, 497-517.
Rekling IC (1992). Interaction between thyrotropin-releasing hormone (TRH) and NMDA-receptor-mediated
responses in hypoglossal motoneurones. Brain Research 578, 289-96.
Richter DW, Schlue WR, Mauritz KH & Nacimiento AC (1974). Comparison of membrane properties of the
cell body and the initial part of the axon of phasic motoneurones in the spinal cord of the cat.
Experimental Brain Research 21, 193-206.
Sah P (1992). Role of calcium influx and buffering in the kinetics ofCa(2+)-activated K+ current in rat vagal
motoneurons. Journal of Neurophysiology 68, 2237-2247.
Sah P & McLachlan EM (1991). Ca2+-activated K+ currents underlying the afterhyperpolarization in guinea
pig vagal neurons: a role for Ca2+-activated Ca2+ release. Neuron 7, 257-264.
Sawczuk A (1993) Adaptation in Sustained Motoneuron Discharge. PhD Dissertation, Seattle: University of
Washington.
Sawczuk A & Binder MD (1992). Reduction of Na+-K+ activity does not reduce the late adaptation of
motoneuron discharge. Society for Neuroscience Abstracts 18, 512.
Sawczuk A & Binder MD (1993). Reduction of the AHP with Mn++ decreases the initial adaptation, but
increases the late adaptation in rat hypoglossal motoneuron discharge. The Physiologist 36, A23.
Sawczuk A, Powers RK & Binder MD (1995). Spike-frequency adaptation studied in hypoglossal motoneurons
ofthe rat. Journal of Neurophysiology 73, In press
Schlue WR, Richter DW, Mauritz KH & Nacimiento AC (1974). Mechanisms of accommodation to linearly
rising currents in cat spinal motoneurones. Journal ofNeurophysiology 37, 310-315.
Schwindt P & Crill WE (1977). A persistent negative resistance in cat lumbar motoneurons. Brain Research
120, 173-178.
Schwindt PC (1973). Membrane-potential trajectories underlying motoneuron rhythmic firing at high rates.
Journal ofNeurophysiology 36, 434-439.
Schwindt PC & Crill WE (1980a). Effects of barium on cat spinal motoneurons studied by voltage clamp.
Journal of Neurophysiology 44,827-846.
Schwindt PC & Crill WE (1980b). Properties of a persistent inward current in normal and TEA-injected
motoneurons. Journal ofNeurophysiology 43, 1700-1724.
Schwindt PC & Crill WE (1980c). Role of a persistent inward current in motoneuron bursting during spinal
seizures. Journal of Neurophysiology 43,1296-1318.
Schwindt PC & Crill WE (1982). Factors influencing motoneuron rhythmic firing: results from a voltage-clamp
study. Journal ofNeurophysiology 48, 875-90.
Schwindt PC & Crill WE (1984). Membrane Properties of Cat Spinal Motoneurons. In: Davidoff RA (ed.),
Handbook of the Spinal Cord, pp. 199-242. New York: Marcel Dekker.
Schwindt PC, Spain WI & Crill WE (1989). Long-lasting reduction of excitability by a sodium-dependent
potassium current in cat neocortical neurons. Journal ofNeurophysiology 61,233-44.
Sokolove PG & Cooke 1M (1971). Inhibition of impulse activity in a sensory neuron by an electrogenic pump.
Journal of General Physiology 57, 125-163.
Spain WI, Schwindt PC & Crill WE (1991). Post-inhibitory excitation and inhibition in layer V pyramidal
neurones from cat sensorimotor cortex. Journal ofPhysiology (London) 434, 609-626.
Spielmann 1M, Laouris Y, Nordstrom MA, Robinson GA, Reinking RM & Stuart DG (1993). Adaptation of
cat motoneurons to sustained and intermittent extracellular activation. Journal of Physiology (Lon-
don) 464, 75-120.
Stuart DG & Callister RJ (1993). Afferent and spinal reflex aspects of muscle fatigue: Issues and speculations.
In: Sargeant AI, D. Kernell (eds.), Neuromuscular Fatigue, pp. 169-180. Amsterdam: Royal Nether-
lands Academy of Sciences.
Vallbo AB (1964). Accommodation related to the inactivation of the sodium permeability in single myelinated
nerve fibres from Xenopus laevis. Acta Physiologyica Scandinavica 61, 429-444.
Viana F, Bayliss DA & Berger AI (1993). Calcium conductances and their role in the firing behavior of neonatal
rat hypoglossal motoneurons. Journal ofNeurophysiology 69,2137-2149.
Woods JJ, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal ofNeurophysiology 58, 125-137.
134 A. Sawczuk et al.

Zajac FE & Young JL (1980). Properties of stimulus trains producing maximum tension-time area per pulse
from single motor units in medial gastrocnemius muscle of the cat. Journal of Neurophysiology 43,
1206-1220.
Zengel JE, Reid SA, Sypert GW & Munson JB (1985). Membrane electrical properties and prediction of
motor-unit type of medial gastrocnemius motoneurons in the cat. Journal of Neurophysiology 53,
1323-1344.
Zhang L (1990). Effects of inositol, 1,4,5-trisphosphate injections into cat spinal motoneurons. Canadian
Journal of Physiology and Pharmacology 68, 1062-1068.
Zhang L & Krnjevic K (1987). Effects of intracellular injections ofphorbol ester and protein kinase C on cat
spinal motoneurons in vivo. Neuroscience Letters 77,287-292.
Zwaagstra B & Kernell D (1980). The duration of after-hyperpolarization in hindlimb alpha motoneurones of
different sizes in the cat. Neuroscience Lettters 19, 303-307.
 9

NEUROMUSCULAR FREQUENCY-CODING
AND FATIGUE

D. Kernell

Department of Medical Physiology

University of Groningen
Bloemsingell0, 9712 KZ Groningen, The Netherlands

ABSTRACT

In daily life, muscle fatigue often becomes noticeable as an apparent decline in the efficiency
of force production by central commands, making it necessary to increase drive (or "effort")
to produce a constant motor output. Such aspects of fatigue may be caused by changes in the
way in which synaptic messages arriving at the motoneurons are translated into forces by
the muscle fibers. Therefore, an understanding of these neuromuscular gradation mecha-
nisms is essential for any analysis of motor fatigue. A brief general review is given of I) how
muscle fibers transduce motoneuronal discharge rates into force; 2) how synaptic currents
are transduced into motoneuronal discharge rates; 3) how activity-dependent changes in the
neuromuscular transduction mechanisms contribute to neuromuscular fatigue; and 4) how
the matching between the transduction mechanisms of motoneurons and those oftheir muscle
fibers may help to optimize neuromuscular gradation efficiency and decrease the severity of
fatigue.

INTRODUCTION

In voluntary motor behavior, the wordfatigue has two common-sense connotations:

either that a certain action requires an increasing amount of effort to be continued, or that
any such continuation is rendered impossible because the necessary force cannot any longer
be produced. The sense of effort is likely somehow to reflect (probably via central corollary
discharges; McCloskey, 1981) the amount of central motor drive descending down to the
spinal motoneurons from the brain (for further comments, see Enoka & Stuart, 1992). The
force produced by a given amount of descending motor drive will depend on: I) how the
descending synaptic input is translated into motoneuronal discharge (recruitment procedures
and stimulus current-to-spike-frequency transduction); 2) how, for the recruited cells, the
motoneuronal discharge is translated into relative muscle-fiber force (spike-frequency-to-
tension (force) transduction); and 3) the maximum force-generating capacities of the muscle
fibers, determining what the absolute force of contraction will be for a given relative degree
of activation. Hence, it is obvious that even if the potential maximal muscle force remains

13S
 136 D. Kernell

normal, fatigue may still be present because of, for instance, changes in the transduction
properties of motoneurons and muscle fibers (factors 1-2). The physiological significance
of the frequency-transduction properties of muscle cannot be well understood without also
considering those of their motoneurons. Therefore, although the present brief general review
is primarily focused on muscle properties, some aspects of motoneuron physiology will be
dealt with as well.
Firstly, an account will be given of the unfatigued tension-frequency relation of muscle.
Secondly, the frequency-encoding properties of motoneurons will be briefly described. Thirdly,
the potentiating and/or fatiguing effects of preceding use on these neuromuscular transduction
properties will be analyzed and, lastly, the possible fatigue-alleviating effects of the matching
of transduction properties between motoneurons and muscle fibers will be discussed.

TENSION-FREQUENCY RELATION IN MUSCLE

Effects of Steady-Rate Activation on Force

It has been known since long that, when tested with constant-rate bursts, there is a
markedly sigmoid relation between relative isometric force and the rate of activation (T-f
relation; Fig. IB) in whole muscles (e.g., Cooper & Eccles, 1930; Edwards et aI., 1977; Cooper
et aI., 1988) as well as in single muscle units (e.g., Kernell et aI., 1975, 1983; Botterman et aI.,
1986; Thomas et aI., 1991; Powers & Binder, 1991). The rate-position of this relationship is in
itself an important expression of isometric contractile speed: the faster the muscle (unit) twitch,
the further is the steep portion of the T-fcurve shifted toward higher rates. One practical measure
for this aspect of contractile speed is, for instance, the stimulation rate (or interval) required for
producing 50% of maximum force. Within a given muscle, this midforce interval tends to be

A B
~ ~--------------~ 100 r----------=-~r__~

~
o
u.

00 140 0 0 1 , - - - - - - -----,.1
200
Stimulus current (nA) Activation rate (Hz)

Figure 1. Relationship between motoneuronal and muscle fiber transduction properties (schematic). A,
motoneuronal relation between discharge rate and the intensity of a steady activating current (j-f curve), e.g.,
as applied via an intracellular microelectrode. B, muscle unit relation between isometric force and activation
rate (T-f curve). Numerical dimensions on x- and y-axes of A and B are those characteristic for relatively fast
units of cat hindlimb muscles, for B as studied at the muscle length giving a maximal twitch force. A rightward
shift of the T-f curve will cause less force for a given intermediate spike frequency (B, arrow labeled Fat;
fatigue) and a leftward shift gives the opposite effect (arrow labeled Pot; potentiation). The two linear portions
of the ff curve are referred to as primary and secondary ranges (A, Kernell, 1965a). For stimulus currents
exceeding those of the linear secondary range, the ff relationship flattens off or declines (see dotted lines)
and, finally, firing typically stops (spike inactivation). Lower and upper limits of spike-frequency for primary
and secondary range indicated by lines in A (horizontal) and B (vertical). Firing rates are obtained after initial
phase of adaptation.
Neuromuscular Frequency-Coding and Fatigue 137

highly correlated with twitch time course (e.g., time-to-peak, half-relaxation time; Kernell et
al., 1983; Bottermanetal., 1986; Thomasetal., 1991). Units ofdifferent muscles may, however,
differ in their precise relationship between T-f curve and twitch time-course (Kernell et al.,
1983; Botterman et al., 1986). The steep portion of the T-fcurve has a higher slope for slow than
for faster muscles and units (Cooper & Eccles, 1930; Kernell et al., 1983; Botterman et aI.,
1986). It should be noted that most of the published measurements of T-f relations concern
isometric contractions rather than shortening or lengthening ones (for recent data on non-iso-
metric contractions, see Heckman et aI., 1992).
To understand the genesis of the sigmoid T-frelation, the apparent force summation
has repeatedly been analyzed for two or more single-pulse stimuli at various intervals (e.g.,
Cooper & Eccles, 1930; Burke et al., 1976; Parmiggiani & Stein, 1981). One of the main
results of such analysis is that the summation is markedly non-linear with, at the most
effective intervals, typically more total force-time area being produced than the algebraic
sum of the individual twitches. The apparent summation is likely to depend on the kinetics
of calcium-associated mechanisms (e.g., release, removal, etc.), and also on changes of
muscle fiber stiffness (cf. Parmiggiani & Stein, 1981), that take place after consecutive
muscle action potentials in a stimulus train. The apparent fusion frequency (no oscillations
visible) may be considerably lower than the stimulation rate needed for maximal force; one
should not use the termfused contraction as being equivalent to maximal tetanic contraction.
The presence of apparently fused but non-maximal tetanic contractions (note importance of
measuring at sufficiently high gain, see e.g., Buller & Lewis, 1965) might partly depend on
filtering properties of tissue in series with the contractile elements, rendering small mechani-
cal sarcomere oscillations invisible at the tendon.

Length Effects
Muscle length is known to influence isometric muscle speed such that, over most of
its possible working lengths in the body, the twitches of a muscle become progressively faster
at shorter lengths (e.g., Rack & Westbury, 1969; Stephens et al., 1975). Correspondingly,
also the T-fcurve shifts toward higher rates as a muscle is shortened within its physiological
working range (Rack & Westbury, 1969).

Temperature Effects
Significant effects on twitch-speed and T-f curve take place as changes occur in
muscle temperature. Such effects might well be of importance under fatiguing circumstances
when muscle temperature becomes elevated during work. As compared to cold muscles,
warmer muscles have faster twitches and a T-fcurve shifted toward higher rates (Ranatunga,
1982), thus increasing the amount of motor drive needed for a given relative force. The effect
on absolute force is, however, complex because a warmer muscle may also have a higher
maximal tetanic force (Ranatunga, 1982). Furthermore, in concentric contractions, the
increased shortening speed of a warmer muscle apparently has beneficial results in that it
results in a greater maximal power output (Sargeant, 1987).

Variable-Rate, Tension-Frequency Relations and "Catch-Like" Effects

One or a few brief initial intervals may have a prolonged enhancing effect on the
forces evoked by subsequent firing at lower rates. Such a catch-like facilitation of force
production is more long-lasting in slow than in fast units (Burke et aI., 1976), but effects of
this kind may be relevant in units of any contractile speed (e.g., Bevan et aI., 1992) and such
phenomena are also known to exist in man (e.g., Binder-Macleod & Barker, 1991).
138 D. Kernell

STIMULUS CURRENT-TO-SPIKE-FREQUENCY TRANSDUCTION

IN MOTONEURONS

General Considerations
A first requirement for using the rate gradation offorce is that motoneurons can indeed
vary their discharge rate within a range adequate for the T:frelations of their muscle fibers (see
also Sawczuk et aI., Chapter 8). Motoneurons receive thousands of synapses on the receptive
membrane regions, i.e., on their soma and dendrites. When activated by a presynaptic action
potential, a single synapse produces only a small and rather short-lasting pulse of current in the
motoneuron. Normally occurring repetitive motoneuron discharges are driven by the relatively
steady sum of numerous such unitary pulses, arriving asynchronously at the motoneuron. Due
to their intrinsic membrane properties, the motoneurons (and most other kinds of nerve cell)
are capable oftransducing such steady currents into maintained repetitive discharges (reviews:
Kernell, 1992; Binder et aI., 1993). This current-to-frequency transduction can be investigated
with currents injected through an intracellular microelectrode. In spinal motoneurons of the cat,
the frequency-current curve ([-I curve) generally has a shape such as that shown in Fig. lA,
whereby it should be noted that firing within the lower primary range seems to be more stable
than that within the upper secondary one. The shape and slope of the/-I curve depend on the
combined effects of several membrane properties, including an important contribution of the
calcium-dependent, potassium-conductance changes that lead to the occurrence of a long-last-
ing hyperpolarizing afterpotential following an action potential (afterhyperpolarization, AHP;
for further references, see Kernell, 1992). The duration of the AHP is a major determinant of
the minimum rate of maintained discharge (Kernell, 1965b; for other contributing factors, see
Carp et aI., 1991), and its time-course and size are important for the steepness of the/-I relation,
i.e., for the motoneuronal gain.

Modifications of I-I Gain

In various species, the conductance changes underlying the AHP of motoneurons can
be affected by transmitter substances, such as serotonin (review: Binder et aI., 1993). This
provides the synaptic means for changing the/-I gain. Duringfictive locomotion, maintained
in decerebrate cats with brain stem stimulation, motoneurons have been seen to discharge at
a relatively high rate in a state of apparently very low /-1 gain (i.e., f not varied by extra

Figure 2. A, late phase of spike-frequency adaptation in motoneurons. Measurements of discharge rate from
motoneurons of cat medial gastrocnemius, as obtained during intracellular stimulation with a weak steady
current of 5 nA above the threshold for rhythmic firing. Mean data for 6 cells that were all discharging
steadily during 180 s or more (based on data from Kernell & Monster, 1982a). B, simultaneous presence
of fatigue and potentiation. Normalized measurements of isometric force from 1st deep lumbrical muscle
of cat while the muscle nerve was electrically stimulated with pulse rates of alternating 360 ms bursts set
to 121 Hz (fused contractions, filled
A B
25 100 .... symbols) and 30 Hz (non-fused con-
.............. 121 Hz tractions, open symbols), respec-
-
........ J'
........... tively. Horizontal dotted line drawn
.... through initial force value obtained
with 30 Hz stimulation. Note con-
tinuous fall in upper curve (fatigue)
while there is an initial phase of
growth in the lower curve (potentia-
2 4 tion) (based on data from Kernell et
Time (min) Time (min) a!., 1975).
Neuromuscular Frequency-Coding and Fatigue 139

injected current) and with AHPs of substantially decreased amplitude (Brownstone et aI.,
1992). It is, however, still unknown whether this interesting behavior represents a common
functional state of these cells.

Spike-Frequency Adaptation
As a motoneuron is activated with a step of steady injected current, it shows two
phases of spike-frequency adaptation: 1) the initial phase ofadaptation, during which there
is a rapid decline of discharge rate as well as ofJ-I slope; and 2) the late phase ofadaptation
during which a slowly progressive frequency decline takes place without practically any
change ofJ-I slope (Kernell & Monster, 1982a, their Fig. 2A; see also Sawczuk et aI., Chapter
8). The initial phase mainly occupies the first few intervals of a discharge, and most of the
late phase takes place during the subsequent 30-60 s. AHP properties (AHP summation) are
important for the initial phase, but probably not for the late phase (see Kernell & Monster,
1982a). Both phases of adaptation are markedly more pronounced at high than at low levels
of activity and, as will be further commented upon below, they are both of direct interest in
relation to time-dependent muscle properties such as fatigue. When activated by the same
(small) amount of steady injected current, the late phase of adaptation is much more marked
for fast than for slow motoneurons (probably largely a reflection of their differences in
absolute discharge rate, Kernell & Monster, 1982b; see also Spielmann et aI., 1993).

POTENTIATING AND/OR FATIGUING EFFECTS OF ACTIVITY

ON NEUROMUSCULAR ENCODING PROPERTIES

Steady-Rate Muscle Activation

There are two types of effect on the T-f relation during continued contractions, as
produced by a constant pattern of submaximal activation (e.g., in a fatigue test): 1)
potentiation, a force-increasing effect; and 2)fatigue, a force-decreasing effect. In submaxi-
mal contractions, either one of these effects may manifest itself as a shift of the T-f curve:
leftward along the rate-axis for potentiation, rightward for fatigue (see arrows labeled Pot
and Fat in Fig. 1B). In submaximal or maximal contractions, fatigue may also be caused by
a change in force-generating capabilities of the muscle, e.g., by a drop of maximal tetanic
force. In most instances, potentiation- and fatigue-associated effects are probably more or
less simultaneously present, and the net result will depend on their sum. The simultaneous
presence of both of these aspects of muscle reaction is easily demonstrated by activating the
muscle alternatingly with low and high rates (Kernell et aI., 1975, their Fig. 2B; Cooper et
aI., 1988; Bevan et aI., 1992; for further references, see Enoka & Stuart, 1992).
During moderate to strong on-going motor activity, muscle contractions may become
markedly slower (e.g., Bigland-Ritchie et aI., 1983b), in itself an effect that might impede
normal muscle use in rapidly alternating movements. The slowing is also, however, poten-
tially beneficial in that it acts to shift the T-f curve toward lower rates; i.e., it produces a
potentiating effect (arrow Pot in Fig. IE). Thus, from the gradually increasing to the
gradually decreasing portion of a prolonged, variable-force muscle contraction, the T-fcurve
is typically shifted toward lower rates (Binder-Macleod & Clamann, 1989). For a similarly
arranged series of intermittent contractions caused by steady-rate bursts of different frequen-
cies, net potentiation effects on the T-frelation are typically mainly seen for units that have
a relatively high degree of fatigue-resistance (Kernell et aI., 1975; Thomas et aI., 1991;
Powers & Binder, 1991). In the presence of a fatigue-induced decline in maximal tetanic
140 [Link]

force, the leftward potentiation-shift of the T-fcurve will help to maintain submaximal forces
without having to increase motoneuronal activity, i.e., potentiation may be seen as a
(temporary) fatigue-alleviating mechanism (see Fig. 2B, note delay of about 3 min before
net fatigue is seen for the non-fused contractions).
At the end of (a series of) intense fatiguing contractions there might not only be a
rate-shift of the T-fcurve but also a relatively greater loss of force-generating effects at high
than at low rates (cf. Fig. 2B), i.e., the slope of the T-fcurve will also decline (e.g., Kernell
et aI., 1975). In extreme cases, the T-fcurve may even temporarily flatten out to nearly zero
slope (Nagesser et aI., 1992).
In addition to the acute slowing and potentiating T-feffects associated with ongoing
activity, there are also often long-lasting (delayed) post-activation effects going in the
opposite direction. These tend to cause a marked rightward shift of the T-fcurve such that
low-rate activation now produces relatively less contractile force (low-frequency fatigue;
Edwards et aI., 1977; Jami et aI., 1983; Cooper et aI., 1988; Powers & Binder 1991; see
arrow Fat in Fig. IB). Interestingly, these effects may continue to increase for some time
after the end of the fatiguing exercise, and typically, they markedly outlast the fatigue-asso-
ciated decline of maximum tetanic force.

Variable-Rate Activation of Muscle

Imposed stimulation experiments in animals and humans indicate that the relative
and absolute force decline seen during a fatigue test might differ between constant-rate
activation bursts and those having brief initial interval( s), evoking catch-like muscle effects
(Binder-Macleod & Barker, 1991; Bevan et aI., 1992). Variable-rate patterns of activation
may be of practical value for minimizing muscle fatigue in the prosthetic use offunctional
electrical stimulation (FES).

Motoneuronal "Fatigue-Reactions"
The late phase of spike-frequency adaptation represents, in itself, a fatigue-like
reaction of the motoneuron; firing rate declines continuously in the presence of a constant
drive (Fig. 2A). Thus, this reaction might contribute to central aspects of motor fatigue.
However, it may also playa role in processes alleviating fatigue (see further below). During
late adaptation, the f-I relation is essentially shifted in parallel along the I-axis (rightward
shift).

NEUROMUSCULAR MATCHING OF TRANSDUCTION

PROPERTIES: RELEVANCE FOR ALLEVIATING FATIGUE

Rate-Matching Betweenf-I Relations of Motoneurons and T-fRelations

of Muscle Units
The middle, steep region of the T-f curve represents the range of rates for which a
variation of activation frequency gives a marked change of muscle output. Within this portion
of the T-f curve, the conditions are optimal for the rate-gradation of force. In rate gradation
of force, endurance (i.e., resistance to fatigue) will presumably be promoted by keeping
motoneuronal discharge rates as low as possible. Hence, it would be advantageous for
endurance to recruit motoneurons at firing rates corresponding to the low end of the steep
portion of the T-fcurve (Fig. 1). This is also how thef-I and T-frelationships are typically
Neuromuscular Frequency-Coding and Fatigue 141

matched (cf. A and B of Fig. 1, motoneuron-muscle unit speed-match; Kernell, 1965b, 1992).
At least in anaesthetized animals, the minimal rate of maintained motoneuronal firing takes
place at intervals equal to the duration of the AHP (Kernell, 1965b). The lower end of the
T-f curve is situated at the activation rate at which consecutive twitches start to sum. This
occurs at a stimulus interval about equal to the total twitch duration. Thus, the motoneuron-
muscle unit speed-match is a consequence of the fact that the total duration of AHP in
motoneurons tends to be very similar to the total duration of the twitch in their muscle fibers
(Bakels & Kernell, 1993a,b). In muscles with a wide variation of unit contractile speed, a
systematic co-variation between AHP and twitch speed has been observed (e.g., medial
gastrocnemius; cat, Zengel et aI., 1985; rat, Gardiner & Kernell, 1990 and Bakels & Kernell,
1993a) in muscles with a more restricted speed range, similar average AHP and twitch
durations may occur in the absence of a systematic co-variation (rat tibialis anterior, Bakels
& Kernell, 1993b). Muscles with units showing a systematic co-variation betweenf-I and
T-fproperties apparently exist also in humans; for example, in the short toe extensor, extensor
digitorum brevis, voluntarily activated motor units were reported to show lower minimal
firing rates for units with a slow twitch than for those with a faster one (Grimby et aI., 1979).
The matching between motoneuronal and muscle fiber properties does not only
concern the lower limits; also, the maximal discharge rates of motoneurons seem well
adapted to those needed for producing forces up to the (near-)maximal ones (Kernell, 1965b,
1992). Furthermore, at high activation rates, the steep slope of the f-I curve tends to
compensate for the relatively flat slope at the upper end of the T-f curve (cf. Fig. 1).

Dynamic Matching of Motoneuronal Frequency to Muscle Length and

Temperature
As was described above, the T-f relation has the complicating property that it is
markedly dependent on muscle length. Attempts to find out whether, in voluntarily activated
muscles, motoneuronal discharge behaviors alter correspondingly has met with mixed
results. Positive findings were obtained in two studies on minimal firing rates (recruitment
rates), which were found to be higher during shortening than lengthening contractions (Tax
et aI., 1989), and during short- than long-length isometric contractions (Vander Linden et
aI., 1991). The mechanisms underlying these adaptations are still unknown (possibly
synaptic effects of monoamines on membrane properties). When studying mean EMG
spike-frequencies of mUltiple units at medium to high levels of isometric force, no relation-
ship was found between this parameter and muscle length (Bigland-Ritchie et aI., 1992a).
Similarly, in a study concerning the effects of muscle temperature, the decreased speed of a
cold vs. a warmer muscle was not associated with a systematic change of mean motoneuronal
discharge rates (Bigland-Ritchie et aI., I 992b).

Initial Spike-Frequency Adaptation

The initial phase of adaptation means that motoneurons may easily be made to fire
at very high frequencies for brief initial durations, reaching rates much faster than those
possible in steady firing (in cats, up to several hundred Hz; Kernell, 1965a, 1992). This fits
well to the fact that such very high rates are also needed for a maximal rate of rise of force
in skeletal muscle (Buller & Lewis, 1965); by virtue of the transduction properties of
motoneurons, such high, initial spike rates can be produced without having to use excessive
amounts of central drive (and effort) (Baldissera et aI., 1987). The ease with which a few
brief initial discharge intervals may be produced by direct intracellular activation of mo-
toneurons fits well to the findings that such brief initial intervals are often found in quick
142 D. Kernell

voluntary contractions (Desmedt & Godaux, 1977). However, this facet of motoneuronal
behavior is not only of interest for the rapidity of force increase; one or a few brief initial
spike intervals may also have a prolonged enhancing effect on the forces evoked by
subsequent firing at lower rates (catch-like effects). As was pointed out above, a series of
contractions produced by such variable-rate patterns might show less force decline than those
evoked by steady-rate bursts.

Possible Protective Value of Late Spike-Frequency Adaptation

During a maximal voluntary contraction (MVC), the gradual slowing of contractile
speed leads to a gradual decrease in the motoneuronal firing rates needed for maximum muscle
activation (Bigland-Ritchie et al., 1983b); and, indeed, their spike-frequencies decline corre-
spondingly (Bigland-Ritchie et al., 1983a; Marsden et al., 1983). This has a protective value in
helping to maintain an adequate neuromuscular transmission and excitation-contraction cou-
pling throughout continuous high-force contractions (Jones et al., 1979; muscular wisdom,
Marsden et aI., 1983). In human MVCs, part of the drop in discharge rate is caused by reflex
mechanisms (e.g., Woods et ai., 1987; Macefield et ai., 1993; further references: Enoka &
Stuart, 1992). However, ifpresent also in humans, late adaptation of the motoneurons (Fig. 2A)
would contribute as well, perhaps mainly in initial phases ofMVCs.

Motoneuronal Recruitment Gradation

Although not a main issue of the present chapter, the spike-frequency-gradation of
motoneuronal activity represents only one of the two best-known mechanisms for the voluntary
control of muscle force, the other one being a gradation in the number of recruited motoneurons
and muscle units. With regard to the recruitment-hierarchy used in force-gradation in most
contractions, recruitment tends to start with the most fatigue-resistant units and proceed toward
those with gradually less endurance (e.g., Henneman & Mendell, 1981; Kernell, 1992; see also
Sawczuk et aI., Chapter 8). As has been repeatedly pointed out, this fractionated muscle use
helps to optimize muscle endurance in normal motor behavior, e.g., by letting weak postural
contractions be executed by units with an adequately high degree of fatigue-resistance. In
addition to such hierarchy considerations, recruitment-gain is a variable of potential impor-
tance for other facets of motor control. This gain (i.e., the ease with which synaptic inputs may
alter the number of recruited motoneurons) may be influenced by steady background activity
of excitatory and/or inhibitory synaptic systems (Kernell & Hultborn, 1990). Hypothetically,
fatigue-associated alterations of such background synaptic biasing might lead to changes in the
effort needed to drive a motoneuron pool, i.e., to changes of conceivable relevance for
neuromuscular fatigue. Furthermore, the recruitment gain would be important for determining
the balance between recruitment and rate modulation when both these mechanisms are used in
parallel in contractions of weak to moderate force. For instance, in long-lasting postural
contractions, a high degree of endurance would be expected to be promoted by emphasizing
recruitment gradation (keeping motoneuronal discharge rates low) for the motoneurons con-
cerned, i.e., by having particularly closely spaced recruitment thresholds for the most easily
recruited units (corresponding distribution of intrinsic electrical thresholds seen for gastroc-
nemius motoneurons; Bakels & Kernell, 1994).

CONCLUDING REMARKS

The considerations of this brief review emphasize that: 1) the frequency-transduction

properties of motoneurons and muscles have to be considered when analyzing neuromuscu-
Neuromuscular Frequency-Coding and Fatigue 143

lar fatigue; 2) these transduction properties have to be viewed together for motoneurons and
their muscle fibers in order to understand their physiological significance; and 3) the manner
in which motoneuronal and muscle fiber properties are combined (matched) helps, III a
significant manner, to decrease the severity of neuromuscular fatigue.

ACKNOWLEDGMENTS
This work was supported, in part, by the Netherlands Organization for Scientific
Research (NWO).

REFERENCES
Bakels R & Kernell D (1993a). Matching between motoneurone and muscle unit properties in rat medial
gastrocnemius. Journal of Physiology (London) 463, 307-324.
Bakels R & Kernell D (1993b). "Average" but not "continuous" speed-match between motoneurons and muscle
units of rat tibialis anterior. Journal of Neurophysiology, 70, 1300-1306.
Bakels R & Kernell D (1994). Threshold-spacing in motoneurone pools of rat and cat: possible relevance for
manner of force gradation. Experimental Brain Research 102, 69-74.
Baldissera F, Campadelli P & Piccinelli L (1987). The dynamic response of cat gastrocnemius motor units
investigated by ramp current injection into their motoneurones. Journal ofPhysiology (London) 387,
317-330.
Bevan L, Laouris Y, Reinking RM & Stuart DG (1992). The effect of the stimulation pattern on the fatigue of
single motor units in adult cats. Journal of Physiology (London) 449, 85-108.
Bigland-Ritchie BR, Furbush FH, Gandevia SC & Thomas CK (1992a). Voluntary discharge frequencies of
human motoneurons at different muscle lengths. Muscle & Nerve IS, 130-137.
Bigland-Ritchie B, Johansson R, Lippold OCJ, Smith S & Woods JJ (1983a). Changes in motoneurone firing
rates during sustained maximal voluntary contractions. Journal ofPhysiology (London) 340, 335-346.
Bigland-Ritchie B, Johansson R, Lippold OCJ & Woods JJ (1983b). Contractile speed and EMG changes
during fatigue of sustained maximal voluntary contractions. Journal ofNeurophysiology 50,313-324.
Bigland-Ritchie B, Thomas CK, Rice CL, Howarth JV & Woods JJ (1992b). Muscle temperature, contractile
speed, and motoneuron firing rates during human voluntary contractions. Journal of Applied Physi-
ology 73,2457-2461.
Binder MD, Heckman CJ & Powers RK (1993). How different afferent inputs control motoneuron discharge
and the output of the motoneuron pool. Current Opinion in Neurobiology 3, 1028-1034.
Binder-Macleod SA & Barker CB (1991). Use of a catchike property of human skeletal muscle to reduce
fatigue. Muscle & Nerve 14, 850-857.
Binder-Macleod SA & Clamann HP (1989). Force output of cat motor units stimulated with trains of linearly
varying frequency. Journal of Neurophysiology 61, 208-217.
Botterman BR, Iwamoto GA & Gonyea WJ (1986). Gradation of isometric tension by different activation rates
in motor units of cat flexor carpi radialis muscle. Journal of Neurophysiology 56, 494-506.
Brownstone RM, Jordan LM, Kriellaars DJ, Noga BR & Shefchyk SJ (1992). On the regulation of repetitive
firing in lumbar motoneurones during fictive locomotion in the cat. Experimental Brain Research 90,
441-455.
Buller AJ & Lewis DM (1965). The rate of tension development in isometric tetanic contractions of mammalian
fast and slow skeletal muscle. Journal of Physiology (London) 176, 337-354.
Burke RE, Rudomin P & Zajac FE (1976). The effect of activation history on tension production by individual
muscle units. Brain Research 109, 515-529.
Carp JS, Powers RK & Rymer WZ (1991). Alterations in motoneuron properties induced by acute dorsal spinal
hemisection in the decerebrate cat. Experimental Brain Research 83, 539-548.
Cooper RG, Edwards RHT, Gibson H & Stokes MJ (1988). Human muscle fatigue: Frequency dependence of
excitation and force generation. Journal of Physiology (London) 397, 585-599.
Cooper S & Eccles JC (1930). The isometric responses of mammalian muscles. Journal ofPhysiology (London)
69,377-385.
Desmedt JE & Godaux E (1977). Ballistic contractions in man: characteristic recruitment pattern of single
motor units of the tibialis anterior muscle. Journal of Physiology (London) 264, 673-693.
144 D. Kernell

Edwards RHT, Hill DK, Jones DA & Merton PA (1977). Fatigue of long duration in human skeletal muscle
after exercise. Journal of Physiology (London) 272, 769-778.
Enoka RM & Stuart DG (1992). Neurobiology of muscle fatigue. Journal of Applied Physiology 72,
1631-1648.
Gardiner PF & Kernell D (1990). The "fastness" of rat motoneurones: time-course of afterhyperpolarization
in relation to axonal conduction velocity and muscle unit contractile speed. Pjliigers Archiv 415,
762-766.
Grimby L, Hannerz J & Hedman B (1979). Contraction time and voluntary discharge properties of individual
short toe extensor motor units in man. Journal of Physiology (London) 289, 191-201.
Heckman CJ, Weytjens JLF & Loeb GE (1992). Effect of velocity and mechanical history on the forces of
motor units in the cat medial gastrocnemius muscle. Journal ofNeurophysiology 68, 1503-1515.
Henneman E & Mendell LM (1981). Functional organization of motoneuron pool and its inputs. In: Brookhart
IM, Mountcastle VB (sec. eds.), Brooks VB (vol. ed.), Handbook of Physiology, sec. ], vol. II, pt 1,
The Nervous System: Motor Control., pp. 423-507. Bethesda, MD: American Physiological Society.
Jami L, Murthy KSK, Petit J & Zytnicki D (1983). After-effects of repetitive stimulation at low frequency on
fast-contracting motor units of cat muscle. Journal of Physiology (London) 340, 129-143.
Jones DA, Bigland-Ritchie B & Edwards RHT (1979). Excitation frequency and muscle fatigue: Mechanical
responses during voluntary and stimulated contractions. Experimental Neurology 64, 401-413.
Kernell D (1965a). High-frequency repetitive firing of cat lumbosacral motoneurones stimulated by long-last-
ing injected currents. Acta Physiologica Scandinavica 65, 74-86.
Kernell D (1965b). The limits of firing frequency in cat lumbosacral motoneurones possessing different time
course of afterhyperpolarization. Acta Physiologica Scandinavica 65, 87-100.
Kernell D (1992). Organized variability in the neuromuscular system: A survey of task-related adaptations.
Archives Italiennes de Biologie 130, 19-66.
Kernell D, Ducati A & Sjoholm H (1975). Properties of motor units in the first deep lumbrical muscle of the
cat's foot. Brain Research 98, 37-55.
Kernell D, Eerbeek 0 & Verhey BA (1983). Relation between isometric force and stimulus rate in cat's
hindlimb motor units of different twitch contraction time. Experimental Brain Research 50, 220-227.
Kernell D & Hultborn H (1990). Synaptic effects on recruitment gain: a mechanism of importance for the
input-output relations of motoneurone pools? Brain Research 507, 176-179.
Kernell D & Monster AW (1982a). Time course and properties oflate adaptation in spinal motoneurones in
the cat. Experimental Brain Research 46,191-196.
Kernell D & Monster AW (l982b). Motoneurone properties and motor fatigue. An intracellular study of
gastrocnemius motoneurones of the cat. Experimental Brain Research 46, 197-204.
Macefield VG, Gandevia SC, Bigland-Ritchie B, Gorman RB & Burke D (1993). The firing rates of human
motoneurones voluntarily activated in the absence of muscle afferent feedback. Journal ofPhysiology
(London) 471, 429-443.
Marsden CD, Meadows JC & Merton PA (1983). "Muscular wisdom" that minimizes fatigue during prolonged
effort in man: peak rates of motoneurone discharge and slowing of discharge during fatigue. In:
Desmedt J.E. (ed.), Motor Control Mechanisms in Health and Disease, pp.169-211. New York: Raven
Press.
McCloskey DI (1981). Corollary discharges: motor commands and perception. In: Brookhart JM, Mountcastle
VB (sec. eds.), Brooks VB (vol. ed.), Handbook ofPhysiology, sec. 1, vol. II,pt 2, The Nervous System:
Motor Control, pp. 1415-1447. Bethesda, MD: American Physiological Society.
Nagesser AS, Van der Laarse WJ & Elzinga G (1992). Metabolic changes with fatigue in different types of
single muscle fibres of Xenopus Laevis. Journal of Physiology (London) 448, 511-523.
Parmiggiani F & Stein RB (1981). Nonlinear summation of contractions in cat muscles. II. Later facilitation
and stiffness changes. Journal of General Physiology 78, 295-311.
Powers RK & Binder MD (1991). Effects oflow-frequency stimulation on the tension-frequency relations of
fast-twitch motor units in the cat. Journal of Neurophysiology 66, 905-918.
Rack PMH & Westbury DR (1969). The effects of length and stimulus rate on tension in the isometric cat
soleus muscle. Journal of Physiology (London) 204, 443-460.
Ranatunga KW (1982). Temperature-dependence of shortening velocity and rate of isometric tension devel-
opment in rat skeletal muscle. Journal of Physiology (London) 329, 465-483.
Sargeant AJ (1987). Effect of muscle temperature on leg extension force and short-term power output in
humans. European Journal ofApplied Physiology 56, 693-698.
Spielmann JM, Laouris Y, Nordstrom MA, Robinson GA, Reinking RM & Stuart DG (1993). Adaptation of
cat motoneurons to sustained and intermittent extracellular activation. Journal of Physiology (Lon-
don) 464, 75-120.
Neuromuscular Frequency-Coding and Fatigue 145

Stephens JA, Reinking RM & Stuart DG (1975). The motor units of cat medial gastrocnemius: Electrical and
mechanical properties as a function of muscle length. Journal ofMorphology 146, 495-512.
Tax AAM, Van der Gon JJD, Gielen CCAM & Van den Tempel CMM (1989). Differences in the activation of
m. biceps brachii in the control of slow isotonic movements and isometric contractions. Experimental
Brain Research 76, 55-63.
Thomas CK, Bigland-Ritchie B & Johansson RS (1991). Force-frequency relationships of human thenar motor
units. Journal of Neurophysiology 65, 1509-1516.
Vander Linden OW, Kukulka CG & Soderberg GL (1991). The effect of muscle length on motor unit discharge
characteristics in human tibialis anterior muscle. Experimental Brain Research 84, 210-218.
Woods JJ, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal ofNeurophysiology 58, 125-137.
Zengel JE, Reid SA, Sypert GW & Munson JB (1985). Membrane electrical properties and prediction of
motor-unit type of medial gastrocnemius motoneurons in the cat. Journal of Neurophysiology 53,
1323-1344.
 10

HUMAN MOTOR UNITS STUDIED BY

SPIKE-TRIGGERED AVERAGING AND
INTRANEURAL MOTOR AXON
STIMULATION

C. K. Thomas

The Miami Project to Cure Paralysis

University of Miami School of Medicine
Miami, Florida 33136

ABSTRACT

When low-threshold motor units are activated at low rates during sustained, weak voluntary
contractions, most unit force profiles exhibit fatigue but some show force potentiation. These
data, obtained by spike-triggered averaging, are compared to the fatigue resistance of human
motor units activated at twitch and tetanic rates by intraneural motor axon stimulation. With
the latter technique, representative sampling of the motor units from one muscle group shows
that unit force fatigue or potentiation at submaximal frequencies, and contractile rate
changes, dictate the shifts in unit force-frequency relationships. More diverse fatigue
protocols, and when possible, careful comparisons of data obtained by both these techniques,
are needed to further our understanding of the force and frequency changes of single motor
units during voluntary and stimulated exercise.

INTRODUCTION

Many studies involving mammalian (cat, rat and rabbit) limb muscles have explored
the loss of force generating capacity or fatigability by single motor units. In contrast, human
studies have regularly examined fatigue at the whole muscle level during either voluntary
or stimulated contractions. These differences between studies relate in part to the physiologi-
cal and technical difficulties of selectively activating single human motor units and the
recording of the associated contractile properties. In human studies, four methods have been
used: 1) spike triggered averaging (Buchthal & Schmalbruch, 1970; Stein et aI., 1972;
Milner-Brown et aI., 1973a); 2) intramuscular micro stimulation (Taylor & Stephens, 1976;
see Elek & Dengler, Chapter 11); 3) percutaneous nerve stimulation (Sica & McComas,
1971); and 4) intraneural motor axon stimulation (Westling et aI., 1990). This chapterreviews
the contributions the techniques of spike-triggered averaging and intraneural motor axon

147
148 C.K. Thomas

stimulation have made to understanding of muscle fatigue. Some of the reasons for the
paucity of data are highlighted. They are followed by a brief discussion of how comparisons
of data from voluntary and stimulation protocols conducted at the whole muscle and single
motor unit level are crucial to further our understanding of the mechanisms underlying
muscle fatigue.

SPIKE-TRIGGERED AVERAGING

General Considerations
The concept of triggering off a steadily firing nerve or muscle potential and averaging
the signals linked to that activity to extract even the weakest of associations (Mendell &
Henneman, 1968) was extended to human experiments by Buchthal and Schmalbruch
(1970). These investigators described the contraction times and fiber types of groups of
muscle fibers for numerous human muscles such as biceps and triceps brachii as well as
gastrocnemius, soleus, tibialis anterior, and platysma. With further refinement (Stein et aI.,
1972), evaluations were made of the contractile properties of human muscle at the single
motor unit level and during voluntary contractions. In these studies, the subjects were given
feedback of the activity of a single motor unit and asked to fire that unit at a slow steady rate
«12 Hz) for about two min. Provided other active motor units responded asynchronously
relative to the test unit, the electrical and mechanical responses of the test unit could be
averaged from the whole muscle responses. This spike-triggered averaging technique was
used to show that the weakest units were activated during low force voluntary contractions
and as the voluntary contraction grew in strength, stronger units were recruited. Unit strength
was inversely related to twitch contraction time (time to peak force; Milner-Brown et aI.,
1973a,b).
These experiments were important in many ways. First, orderly recruitment of motor
units by force output (or size) during voluntary contractions supported the predictions made
from studies examining the reflex recruitment of motor units in cat hindlimb muscles
(Henneman et aI., 1965). Subsequent studies have shown similar patterns of recruitment in
other limbs and jaw muscles (Desmedt & Godaux, 1977b; Goldberg & Derfler, 1977; Yemm,
1977), during ballistic contractions (Desmedt & Godaux 1977a; cf. Grimby & Hannerz,
1974), when muscles act in different directions (Thomas et aI., 1986 cf. Person, 1974;
Thomas et aI., 1978; Desmedt & Godaux, 1981; ter Haar Romeny et aI., 1982), during reflex
contractions (Ca1ancie & Bawa, 1985), in relation to axon conduction velocity (Freund et
aI., 1975; Dengler et aI., 1988) as well as during dynamic contractions (Thomas etaI., 1987a).
Recruitment by increasing order of force output also occurs in reinnervated muscles after
nerve section and resuture (Milner-Brown et aI., 1974) and particularly when the reinner-
vated muscles are synergistic (Thomas et aI., 1987b), in muscles of individuals with
amyotrophic lateral sclerosis (Dengler et aI., 1990), after chronic spinal cord injury (Stein
et aI., 1990; Thomas, unpublished data), and following immobilization (Duchateau &
Hainaut, 1990). Alternative recruitment schemes have been described during speech (Smith
et aI., 1981), eccentric contractions (Nardone & Schieppati, 1988; Nardone et aI., 1989;
Howell et aI., 1994), when proprioceptive afferents from the active muscles are blocked
(Hannerz & Grimby, 1979), and when cutaneous stimulation is superimposed on a voluntary
contraction (Stephens et aI., 1978; Datta & Stephens, 1981). Thus, the spike-triggered
averaging technique has spawned a number of studies that examine recruitment phenomena.
Furthermore, there have been detailed examinations of various factors which can affect the
data. These include: the influence of twitch fusion at different unit firing rates (Monster &
Chan 1977; Calancie & Bawa, 1986; Thomas et aI., 1990a), the changes in twitch parameters
Human Motor Units 149

in relation to the short-term unit history (Nordstrom et aI., 1989), whether the actual twitch
can be predicted from the averaged, partially fused force profile (Lim et aI., 1995), as well
as the implications of unit synchrony on unit force (Milner-Brown et aI., 1973a; Yue et aI.,
1995). At first glance, the characterization of the contractile properties of human motor units
during voluntary contractions also provided the framework to explore how motor units
respond during fatiguing protocols. However, surprisingly few studies have examined these
responses.

Contributions to Understanding Muscle Fatigue

In two studies individuals fired single motor units in the first dorsal interosseus or
masseter muscle at low rates (10 Hz) for five or 15 min, respectively (Stephens & Usher-
wood, 1977; Nordstrom & Miles, 1990). The ratio of the final to initial twitch force was used
as a fatigue index. In the first dorsal interosseous, units were recruited up to 60% of maximum
voluntary contraction (MVC) force. Low threshold units (recruited below 15% MVC) were
fatigue resistant « 25% force reduction) whereas units activated above 25% MVC force
were highly fatigable (> 75% force reduction). In the masseter study, units were recruited
up to 26% MVC force. Units showed one of three fatigue responses: 1) a slow progressive
decline in twitch force over 15 min; 2) an initial increase in twitch amplitude which was
maintained for 15 min; or 3) a dramatic force reduction during the first four min, then force
stabilization. Overall, the units were usually more fatigable after 15 compared to 6 min of
activation. However, the force loss was not well correlated with the initial twitch force or
contraction time.
In general, these studies have sampled units with low recruitment thresholds and have
described changes in the magnitude of the partially fused force without regard to alterations
in contraction rate. Changes in the relaxation phase have been particularly difficult to assess
using spike-triggered averaging because offorce fusion. Moreover, the nature of the orderly
recruitment of motor units during voluntary contractions has made it largely impossible to
either control the activation history of each unit. Even so, some of these spike-triggered
averaging results, and those from intramuscular micro stimulation studies (Gydikov et aI.,
1976; Stephens & Usherwood, 1977; Garnett et aI., 1979; Young & Mayer, 1981) are
discussed as if the units have been fatigued by classical protocols (Burke et aI., 1973). Such
comparisons imply, dangerously, that the actual protocol is unimportant in determining the
fatigue response.

Why Have So Few Studies Used Spike-Triggered Averaging to Explore

How Motor Unit Properties Change with Use?
Various factors underlie the paucity of fatigue-related studies which have used the
spike-triggered averaging technique. Four variables will be reviewed here briefly. They
include: 1) the averaging process; 2) the problems associated with deriving unit responses
from an active muscle; 3) the limits imposed by force fusion; and, 4) the consequences of
the contractile speed changes which accompany use.
When signals are averaged over long time periods (e.g., minutes) to extract unit
properties, the temporal resolution of the data becomes diluted. It can also mask unit
potentiation (see below) and dramatically change the extent of fatigue. Many human
masseter units which were activated at 10Hz for 15 min showed an initial period of
potentiation followed by a decline in force (Nordstrom & Miles, 1990). When the ratio of
the final to initial force (fatigue index) was computed over one min rather than three min
intervals, the population of motor units seemed much more fatigue resistant. While this
150 C.K. Thomas

associated force potentiation with use has been reported previously in both whole muscle
and motor unit studies, it has largely been ignored by studies that have used pre-potentiation
of the muscle by tetanic stimulation. But it is clear from these human and other studies (see
below) that its effect can be substantial when motor units are activated submaximally.
A second problem relates to extracting single motor unit activity from an active
muscle. It is unclear whether the forces of different motor units summate linearly or
non-linearly. Even when firing asynchronously, there may be interaction between units.
Similarly, short term synchrony can occur between motor units (Datta & Stephens, 1980),
but the magnitude of synchrony, and alterations in unit synchrony during fatigue are
unknown. Moreover, as the contraction intensifies, the electrical interference pattern makes
it much more difficult to selectively identify the same single motor unit potential. As a
consequence, most spike-triggered averaging studies have favored the sampling of low
threshold motor units. In those muscles where recruitment is prominent at low forces, for
example, small hand muscles, large fractions of the population may actually be sampled.
However, for large muscles like biceps brachii where recruitment can continue up to near
maximal forces (Kukulka & Clamann, 1981), unit responses to fatigue have been left
unexamined. Furthermore, as more units are sampled during an experiment, earlier contrac-
tions may have already involved some motor units, altering their apparent responses to
fatigue.
Another variable is whether activating different motor units at the same rate (e.g., 10
Hz) is the most appropriate comparison in terms of understanding actual muscle function.
Since recruitment thresholds and force-frequency relationships differ dramatically between
units, it is unlikely that a unit which is recruited at 10% MVC force and another which is
activated at 60% MVC will be active simultaneously at 10Hz. The lower threshold unit may
have been activated for some time before the higher threshold unit is even recruited. Thus,
clearer information on the interactions between motor unit recruitment and frequency
modulation during fatiguing voluntary contractions may be attained by examining unit
behavior at fixed levels of force instead of at constant firing rates. When measuring motor
unit contractile properties by spike-triggered averaging, however, the feasibility of imple-
menting this suggestion is limited. The ability to average the unit forces depends on limited
twitch fusion. The increases in force fusion at higher firing rates are well documented.
Similarly, fusion varies with unit contractile rate (Monster & Chan 1977; Calancie & Bawa,
1986; Thomas et aI., 1990a).
Another important feature which confounds the use of spike-triggered averaging in
studies of fatigue is the changes in contractile speed which accompany use. Although slowing
of muscle relaxation is well documented in response to whole muscle fatigue, changes in
twitch contraction time are not. However, at the motor unit level, studies in cat hindlimb and
human thenar muscles not only show that both occur, but that units can have quite divergent
responses (Dubose et aI., 1987; Thomas et aI., 1991b; see below). Because of these changes
in twitch speed, the relative fusion offorce at a given frequency (e.g., 10Hz) will be altered.
If a unit is activated at the same absolute rate both before and after fatigue, its force may
simply rise or fall depending on the alterations in twitch contraction and/or half-relaxation
time.
Fig. I illustrates this point by showing data recorded from a single motor unit before
and after a 60 s MVC of the triceps brachii muscle. The protocol was performed by a 22 year
old man who suffered a spinal cord injury at C5-6 3 years before. Since the triceps muscle
is now partially paralyzed, and few motor units remain under voluntary control, the sparse
interference pattern made it possible to follow the same motor unit during the entire protocol.
Two tests were performed immediately before and after the sustained MVC. The unit was
first fired at a slow, steady rate and the associated force averaged. Second, the unit firing
Human Motor Units 151

A) Pre-fatigue 8) Post-fatigue C)
Figure 1. Averaged, rectified
surface EMG and force from one
triceps brachii motor unit before
100
l'
/1
(A) and after (B) a 60 s MVC. C, 1/
whole muscle force vs. unit fre- 80 01
I I
quency measured during brief ?;
"
0
voluntary contractions per- > 60
~
formed prior to and after a 60 s :.e
MVC (open and closed symbols,
0
p.",
respectively). Note the reduction ~0 40 I
o /
/

in surface EMG and the apparent

absence of change in the force Io.5mv
u..
20 •o Pre-fatigue
profile with fatigue (A,B). How- 12mN • Post-fatigue
ever, any submaximal contrac- 0
tion level required a higher unit n 0 10 20 30 40 50
10ms
firing rate after fatigue. Frequency, Hz

rate was measured during brief (3-5s), submaximal and maximal contractions (Thomas et
aI., 1994).
During the fatiguing contraction, the whole muscle force and surface EMG declined
and was accompanied by a 40% reduction in unit firing rate (43 Hz to 27 Hz). Within two
min, maximal firing frequency had almost recovered. However, the force-frequency rela-
tionship recorded by voluntary effort moved horizontally to the right such that higher firing
rates were required to produce the same relative force. For example, to exert 50% of maximal
whole muscle force, the unit was activated at 20 Hz before and 28 Hz after the fatigue test.
Since the relative positions of these force-frequency relationships are closely linked to unit
force and speed in various cat muscles and human thenar muscles (Kernell et aI., 1983;
Thomas et aI., 1991a), these data suggest twitch force and contraction time have been
reduced. However, as shown in Fig. lA, the force and contraction time of the motor unit
were comparable before and after fatigue when assessed using spike triggered averaging
(30.5 mN, 26.3 ms and 31.3 mN, 27.0 ms respectively). Thus, any reductions in unit force
and contraction time due to fatigue may simply have been offset by the gains in each
parameter resulting from reductions in twitch fusion (Thomas et aI., 1994). At present there
is no easy way of determining initial twitch force and speed from the partially fused profile
obtained by spike triggered averaging (Lim et aI., 1995). Changes in muscle compliance and
the pattern of muscle co-activation may also occur with fatigue. Thus, the complexity of
interpreting data obtained by spike-triggered averaging after fatigue is apparent.

Conclusion
Despite all the potential pitfalls apparent in using the spike-triggered averaging
technique to examine how motor unit mechanical properties change during fatigue, two
important features much be acknowledged. The technique has focused attention on unit
fatigue during low-force voluntary contractions thereby emphasizing alterations in un/used
force profiles. Twitch parameters, often considered to be unreliable because of their suscep-
tibility to potentiation have important implications for understanding force-frequency rela-
tionships. Moreover, spike-triggered averaging remains the only technique currently
available to examine motor unit properties during voluntary contractions. As such it remains
a valuable method which may be better understood if its use could be carefully coupled to
alternative techniques that permit the stimulation of single motor units at both twitch and
tetanic rates.
152 C.K. Thomas

INTRANEURAL MOTOR AXON STIMULATION

General Considerations
The recent development of measuring human motor unit contractile properties in
response to intraneural motor axon stimulation (Johansson et aI., 1988a, 1988b; Westling et
aI., 1990) is analogous to various methods used in animals in that single motor axons are
stimulated some distance from the muscle, and each allows measurements of axon conduc-
tion velocities and the full time course of the twitch and tetanic forces. In humans, stimulation
of motor axons within the main nerve trunk depends on the use of conventional mi-
croneurographic techniques. The appropriate nerve must first be located by stimulating
through a tungsten microelectrode (typically 0.2mm diameter, insulated except for the tip,
electrode impedance: 200-400 kn measured with low current at 1 kHz). With further minute
manipulation it is then possible to stimulate and record force and EMG selectively from a
single motor unit. However, baseline fluctuations due to respiration and pulse pressure waves
feature prominently when the signals are amplified sufficiently to measure unit twitch forces.
Thus, three adaptations are necessary for successful recordings (Fig. 2). First, stimulation
has to be selective. Second, stimulation of the motor axon must be time locked to the pulse
pressure cycle so that it is delivered during a period of minimum baseline fluctuation. Third,
the evoked force signals of individual units should be recorded only after the baseline is reset
electronically. With simultaneous use of all these methods, individuals force responses can
be recorded without signal averaging (cf. McKeon & Burke, 1983).
One of the important findings from these experiments involving intraneural stimula-
tion of motor axons to thenar muscle has been that the distribution of measured axonal
conduction velocities suggests that it is possible to sample the entire range ofiarge diameter
myelinated axons in the nerve. Thus, for one human muscle group, it was possible to define
for the first time, the entire range of twitch and tetanic forces, their contraction and relaxation
rates as well as their axonal conduction velocities. When F-responses were encountered (two
unitary EMG responses to a single stimulus), proximal conduction velocities could also be
evaluated. Moreover, the measurement of the force responses in two directions demonstrated
that individual units within the same muscle group generate force in widely different
directions. Each of these unit parameters could be measured readily from single sweeps
(Westling et aI., 1990), a feature which makes this method ideal for examining fatigue of
single human motor units in response to various stimulation patterns. Similarly, mi-
croneurographic data has been recorded in experiments involving strong voluntary contrac-

A) No response 8) Single unit C) Multiple units

500ms

Figure 2. Sequential superimposed force records with and without baseline reset when no (A), a single (B)
and several motor axons are stimulated (C). Note that even without stimulation of an axon (A), the pulse
pressure waves produce a characteristic rapid rise in the baseline followed by a slower relatively linear decline.
Provided stimuli are delivered during periods of relatively low baseline fluctuation, and the baseline is reset
electronically just before the stimulation, unitary responses or graded responses from multiple units become
obvious (modified from Westling et aI., 1990).
Human Motor Units 153

tions (Vallbo, 1970), so it should also be possible to measure motor unit contractile responses
before and after voluntary contractions. Such experiments still need to be performed.
To date, the contractile properties of human motor units have been examined by
intraneural stimulation of motor axons to thenar (median nerve; Westling et aI., 1990;
Thomas et aI., 1990a,b, 1991a,b) and toe extensor muscles (common peroneal nerve;
Bigland-Ritchie et aI., 1993; Fuglevand et aI., 1993). In each experimental series, the twitch
forces spanned a narrow range compared to that reported using spike-triggered averaging,
and intramuscular micro stimulation. Twitch contraction and half relaxation times were
generally faster for thenar units than for toe extensor units or for units measured using
alternative human methods. This was even after the significant slowing resulting from
post-tetanic potentiation and fatigue (Thomas et aI., 1990b, 1991b).
The force-frequency relationships measured for single human thenar motor units
were similar to the corresponding force-frequency functions reported for various whole
human muscles (Marsh et aI., 1981; Gandevia & McKenzie, 1988) in that twitch fusion began
at low stimulation frequencies (already between 5 and 8 Hz), half maximal tetanic force was
reached at 12 ± 4 Hz (mean ± SD), and maximal force output was usually obtained between
30 and 50 Hz. A few units reached maximal tetanic force between 50 and 100 Hz (Thomas
et aI., 1991a). Similar results have been found for human toe extensor motor units although
half maximal force was attained at 10 ± 1 Hz and full fusion sometimes occurred at 20 Hz
(Bigland-Ritchie et aI., 1993). Thus, force modulation for these human units occurred over
a narrow frequency range and one which is typical of the motor unit firing rates recorded
during submaxima1 and maximal voluntary contractions of various human muscles (Belle-
mare et aI., 1983; Big1and-Ritchie et aI., 1986).
As expected (Cooper & Eccles, 1930), gradation of force with stimulus frequency
was highly dependent on unit contractile rate. Half initial maximal tetanic force was evoked
at lower stimulation frequencies for units with slow twitch contraction and half relaxation
times, as found for cat limb motor units (Kernell et aI., 1975, 1983; Botterman et aI., 1986).
However, for both human thenar and toe extensor motor units, there was limited separation
between the force- frequency curves of even the fastest and slowest units which is consistent
with the narrow range of measured twitch contraction times (37 to 70 ms and 50 to 107 ms
respectively). In cat limb muscles, twitch contraction times may vary by more than 2 fold,
force is graded over much wider frequency ranges and force-frequency curves of fast and
slow units fail to overlap (Kern ell et aI., 1975, 1983; Botterman et aI., 1986). Similarly, axon
conduction velocities for human thenar units are slower and cover a narrower range
compared to values for various hindlimb muscles in animals (Westling et aI., 1990).
Clearly there is a limited range of human twitch properties, axonal conduction
velocities and frequencies over which force is modulated. These differences may reflect
variations between distal vs. limb muscles rather than discrepancies between species.
Similarly, the stronger and slower force profiles recorded by spike-triggered averaging
compared to intraneural motor axon stimulation are not explained readily by sample bias or
twitch fusion. At present, no stimulation studies have actually imitated the data which is
typically recorded by spike-triggered averaging. That is, synchronous stimulation at fixed
rates is dramatically different from the conditions where contractile responses are averaged
during low force voluntary contractions. With spike-triggered averaging many other units
are active simultaneously and asynchronously whereas with intraneural stimulation all but
one motor unit are relaxed. Recording from an actively contracting muscle may increase
twitch force because of the higher resting tension. Thus, some of the differences in results
may depend on series compliance. Studies that measure unit force responses by spike-trig-
gered averaging and use the same pattern of unit firing to stimulate the motor axon would
be most helpful in explaining some of these differences. Such studies may be feasible using
surface stimulation (Doherty & Brown, 1994).
154 C.K. Thomas

Contributions to Understanding Muscle Fatigue

Intraneural stimulation of thenar motor axons has revealed features of the underlying
muscle anatomy. The most fatigable units exerted more force in abduction. Fatigue resistant
units produced more force in flexion. This gradient of fatigability with the angle of force
exertion seems appropriate for the function of human thenar muscles. Abductor pollicis
brevis typically acts during the spacing of the fingers prior to grasping objects (Johansson
& Westling, 1988). This is an intermittent action and one that generally does not involve
external loads. In contrast, the thumb flexor or opponens muscles serve to maintain the grip
force for long periods of time, a capacity which would seem well matched to the high fatigue
resistance of these units.
Thenar motor units (Thomas et ai., 1991 b) were only subjected to one fatigue
protocol: intermittent 40 Hz stimulation for two min (Burke et ai., 1973). Most units (72%)
lost less than 25% force, or potentiated indicating how fatigue resistant these units were. The
remaining units lost between 25-75% force showing intermediate fatigability. There were
no units with fatigue indices less than 0.25, which would classify them as fatigable using
traditional criteria. Overall, there was no significant change in the mean twitch or maximal
tetanic force with fatigue. Nine units were subjected to a second fatigue run after several
other additional tetanic protocols had been applied. This additional activation induced only
small changes in force output. Thus, this method of assessing fatigability demonstrated that
human thenar motor units were remarkably fatigue resistant (Thomas et ai., 1991 b).
In contrast to many motor units in muscles of other mammalian species, most thenar
motor units produced maximal tetanic forces with 40 Hz stimulation. For many of these units
there was an obvious change in the shape of the force profile with fatigue because of the
marked slowing in the contraction and particularly the relaxation phases. This slowing meant
that consecutive force responses began to fuse together (Thomas et ai., 1991b), phenomena
that have been observed but rarely emphasized in some other mammalian motor units (Burke
et ai., 1973). In terms of tetanic force responses, three trends were apparent. Some thenar
units lost force and showed contractile slowing (28%). Others lost minimal force in the
presence of contractile slowing (36%). The remaining units displayed little change in either
force or contractile speed (36%). Those units with the strongest maximal tetanic forces
fatigued and slowed the most. Units that lost twitch force with fatigue (55%) tended to slow
in the twitch contraction phase but speed up during relaxation. The 45% of units that showed
twitch potentiation with fatigue showed the opposite contractile rate trends (Thomas et ai.,
1991b).
Some other studies report twitch potentiation and fatigue (Garnett et ai., 1979; Young
& Mayer, 1981; Nordstrom & Miles, 1990). Contractile slowing however was only reported
for opponens pollicis units as measured by intramuscular microstimulation (Gydikov et ai.,
1976). In these units, slowing was greatest for units with fast initial contraction and
half-relaxation times, as observed for human thenar motor units (Thomas et ai., 1990b,
1991 b). Similar results were found for cat medial gastrocnemius units (Dubose et ai., 1987).
After the fatigue test, the frequencies needed to evoke 50% of the initial maximal
force were not significantly different for human thenar units (Thomas et ai., 1991 b).
However, there were diverse responses for different units in that five units showed a
potentiated force output at all submaximal stimulation frequencies (potentiating units) while
seven units showed a force loss at all stimulation frequencies (fatiguing units). Fig. 3 shows
the force-frequency curves before and after fatigue for two typical units. These units
produced initial maximal tetanic forces of95 and 75 mN and stimulation rates of 14 and 10
Hz evoked half these maximal tetanic forces (A,B respectively). In response to fatigue, one
unit lost force at all frequencies; therefore, needing higher rates of stimulation to produce
any given absolute force. A stimulation frequency of 24 Hz was required to evoke half the
Human Motor Units ISS

.•..-
100 A) B) C) 0)
.0
80 100
.....• 100 •• :~····i

....
.0· •.• G·· ·0
80 : p.
E 80 0: 80
..!'ieo
.G 60 :I o·
0 «i
e· E .~
o·
.:
Z ·9
E 60 p .~ 60 .~ 60
.; :
••
0 40 E ~
~0 .Q ;,!! :rP
40 :9
.:
0
0 40 0 40
u. .: !'
~0
0·
• 20
0
•
..-
20 u. 20 20
!t 6 0
•. e
0 0 " 1"1 ii'i"i 0 0
i I
10
i

100
I
10
i

100
i

10
I i

100
I i I
10 '''.
100
Frequency, Hz
Figure 3. Initial (open symbols) and post-fatigue (closed symbols) force-frequency relationships for two
thenar motor units (A and B, respectively). C and D show the corresponding relationships expressed relative
to maximal tetanic force. Fatigue was induced by intermittent 40 Hz stimulation for two min (modified from
Thomas et a\., 1991a).

initial maximal tetanic force (C). In contrast, the maximal tetanic force for the unit shown
in B was similar in both force-frequency tests. However, there were increases in the force
evoked by submaximal frequencies. Half of this unit's maximal tetanic force could now be
evoked at 7 Hz compared to 10Hz before the test for fatigability (D).
These divergent unit responses could not be explained by variable activation histo-
ries. Rather, the changes in stimulation rate needed to evoke submaximal forces with
activation history depended on changes in force output and the time course of these changes.
The maximal potentiated twitch response was reached after one or two tetanic tests for all
the fatigable units. During the fatigue test, both the twitch and tetanic force of these units
declined. In contrast, the other units produced similar maximal tetanic forces before and after
fatigue and continued to show potentiation of force at all submaximal frequencies (Thomas
et aI., 1991 b). Thus the effects of potentiation seem to carry over to unfused force responses
as observed by others (Olson & Swett, 1971; Kernell et ai., 1975; Burke et ai., 1976).
These data reaffirm the influence that twitch and subtetanic force responses have on
the unit force-frequency curve. At low activation rates, a combination of contraction and
relaxation phase speed changes are probably important as well. However, contractile slowing
occurs with both potentiation and fatigue. Thus, the slowing which accompanies twitch
potentiation, rather than fatigue, may dictate the direction in which these force-frequency
curves move.

What Is the Relationship Between Twitch and Tetanic Responses to

Fatigue
Most human studies have examined how twitch forces change with fatigue but most
animal studies have first pre-potentiated the muscle with tetanic stimulation, then examined
the fatigue response to tetanic stimulation. Direct comparisons between the twitch and tetanic
forces in response to the same fatigue test are necessary to understand these different
approaches. Fig. 4 compares for human thenar motor units, the twitch and tetanic fatigue
indices (ratio of final to initial value) for force, maximal contraction rate (MeR) and
half-relaxation time (HRT; A-C respectively), In each plot, the dashed line represents where
the data would fall ifboth the twitch and tetanic responses changed by comparable amounts.
156 C.K. Thomas

A) Force B) MCR C) HRT

0
2 / 2 / 2 /
~ / /
0 /

fP

/i
"'C / / /
.5

t:
/ 0/
Q)
:>
~ o~o
oS o /0 0
.r=
~ o ''i, , /'0 0
~ ,'0 0 CO
9
/ /
/ /
0 I 0 I 0 I
0 2 0 2 0 2
Tetanic fatigue index

Figure 4. Twitch vs. tetanic fatigue index (final to initial ratio) for force (A), maximal contraction rate (B) and
half relaxation time (C). Data falling on the dashed lines indicate comparable change in each parameter. The
solid lines indicate a regression between twitch and tetanic fatigue indices.

The changes in the twitch and tetanic responses were correlated (Fig. 4A, r=O.61, p<[Link]).
However, the tetanic forces of some of the fatigue resistant units showed mild potentiation
and most of these units showed even greater potentiation of twitch forces. Thus, tetanic force
responses may be unchanged after fatigue while the twitch force is potentiated. Since the
fatigable units also underwent a period of twitch potentiation that was over prior to the fatigue
test, the twitch to tetanic force relationship depends heavily on the unit activation history.
Changes in twitch and tetanic maximal contraction rates with fatigue were also
correlated (Fig. 4B, r=O.67, p<[Link]). Thus, for the fatigable units, both these rates slow, along
with the twitch contraction time. The converse is the case for fatigue resistant units.
Relaxation measures from tetanic responses were not correlated to the corresponding
measures taken from twitches (Fig. 4C), so twitch relaxation parameters have little or no
value for predicting the corresponding tetanic parameters. Thus care must be taken when
force and relaxation properties measured from twitches are used to draw inferences about
tetanic contractile properties and vice versa (Thomas et aI., 1991 b).

Why Have So Few Studies Used Intraneural Motor Axon Stimulation to

Explore How Motor Unit Properties Change with Use
Intraneural stimulation of motor axons and measurement of the associated unit
contractile properties is a new technique. Numerous other reasons could underlie the small
data pool. First, it is technically difficult to locate a peripheral nerve with a fine tungsten
electrode. When the nerve lies deep within a limb, even the smallest of adjustments in the
angle of the needle can dramatically alter the location of the electrode tip. Patience, as well
as care are not only needed for nerve location, but also to ensure an absence of nerve damage.
Second, long experiments (up to six hours) require the relaxed and cooperative participation
of the subject. Any movements by the subject can dislodge the position of the stimulating
electrode and thus disrupt recordings. Descriptions of what the subject feels, particularly
during the initial search for the nerve, are also helpful in guiding the experimenter towards
the nerve. When perception to touch has been lost, for example after certain spinal cord
injuries, visible criteria must be used (e.g., muscle contraction). Third, special equipment is
needed to electronically reset the force baseline to minimize fluctuations from respiration
Human Motor Units 157

and pulse pressure waves (see Fig. 2). Force transducers must be sensitive enough to register
the weak twitch forces of single motor units without averaging. It is also preferable to use a
constant current stimulator. Finally, anatomical constraints can limit the choice of test nerves
and muscles. Experimental conditions are more stable if the stimulation site is some distance
from where the muscle recordings are made (e.g., thenar motor axons in the median nerve
can be accessed above the elbow). The particular motor axons of interest must be sufficiently
spaced within the nerve to permit selective stimulation. For example, thenar motor axons
are dispersed across the entire median nerve trunk above the elbow but tend to focus in only
one of two nerve fascicles near the wrist. Isolation of thenar motor axons by the surrounding
axons and the ratio of sensory to motor axons in the median nerve are two reasons proposed
for the success of these earlier recordings. In summary, all these factors translate into low
data yields. Despite this, the technique has much promise for examining single human motor
unit fatigue. Future studies would benefit by varying activation rates and protocols and
coupling these to voluntary contraction protocols.

CONCLUSION
With spike-triggered averaging, only a limited number of twitch parameters are
measured. Force fusion during weak voluntary contractions, which results in reductions in
twitch amplitudes and contraction times, means that the spike-triggered averaging method
is only appropriate for measuring force responses at low firing rates. During sustained, low
force voluntary contractions of masseter or first dorsal interosseous muscles, there is either
decline or potentiation of unit twitch force. However, recording force profiles with spike-
triggered averaging at the same rate before and after fatigue produces data which is difficult
to interpret. Any changes in contractile speed with fatigue will alter how the twitches sum
together and the apparent force profile. Other factors to consider include the inability to
control unit activation history, alterations in muscle compliance, and muscle co-contraction.
Studies with intraneural stimulation of motor axons have allowed the full time course
of twitch and tetanic forces to be recorded for a representative sample of units from one
muscle group. When activated intermittently and maximally at rates which could be defined,
human thenar units either showed twitch potentiation and tetanic fatigue resistance or a
decline in both twitch and tetanic force. After fatigue, these units needed lower and higher
rates respectively, to produce any given submaximal force. These data not only emphasize
how subtetanic force responses influence force-frequency unit behavior but also illustrate
that twitch and tetanic force relationships depend on activation history. Caution is therefore
required when making inferences about tetanic forces from twitch responses and vice versa.
Similarly, it would seem equally dangerous to impose unit classifications traditionally used
for motor units in mammalian muscles on human units if the protocols used to activate the
units vary, and if the evaluation criteria are based primarily on twitch parameters.
Other factors also make it difficult to compare data obtained during evoked vs.
voluntary contractions. For example, most animal studies and human motor axon stimulation
studies have evoked contractions at fixed stimulation rates (cf. Botterman & Cope, 1988).
Motor unit firing rates are modulated continuously during voluntary contractions. Second,
the same fatigue protocol (Burke et aI., 1973) has been used monotonously. It does not reflect
usual muscle activation because it is impossible, by voluntary effort, to maximally activate
a human muscle for a fraction of a second, each second for two min. Third, stimulation drives
motor units synchronously. Motor unit activity during voluntary contractions is generally
asynchronous. Moreover, one unit is typically stimulated whereas numerous units are often
active during voluntary effort. This may change muscle compliance. All these poorly
understood phenomena argue for an infusion of new experiments that aim to define the
158 C.K. Thomas

diversity of unitary responses and the differences between voluntary and stimulated contrac-
tions. Such information will enhance our appreciation of whole muscle responses, the
balance and importance of recruitment and rate modulation in force production, the interac-
tions between muscle potentiation and fatigue, and whether motor unit behavior is organized
around groups of motor units within or across muscles.

ACKNOWLEDGMENTS
The author wishes to thank Drs. Mary Pocock and James Broton for constructive
comments on the manuscript. The author's laboratory is supported by United States Public
Health Service grant NS 30226, and The Miami Project to Cure Paralysis. Attendance of the
author at the 1994 Bigland-Ritchie conference was supported, in part, by NS 30226.

REFERENCES
Bellemare F, Woods JJ, Johansson R & Bigland-Ritchie B (1983). Motor-unit discharge rates in maximal
voluntary contractions of three human muscles. Journal of Neurophysiology 50,1380-1392.
Bigland-Ritchie B, Cafarelli E & and Vollestad NK (1986). Fatigue of submaximal static contractions. Acta
Physiologica Scandinavica Supplement 556, 137-148.
Bigland-Ritchie B, Fuglevand AJ & Macefield VG (1993). Force modulation by rate-coding in single motor
units of human toe extensor muscles. Society for Neuroscience Abstract 19, 154.
Botterman BR & Cope TC (1988). Motor-unit stimulation patterns during fatiguing contractions of constant
tension. Journal ofNeurophysiology 60,1198-1214.
Botterman BR, Iwamoto GA & Gonyea WJ (1986). Gradation of isometric tension by different activation rates
on motor units of cat flexor carpi radialis muscle. Journal ofNeurophysiology 56, 494-506.
Buchthal F & Schmalbruch H (1970). Contraction times and fibre types in intact human muscle. Acta
Physiologica Scandinavica 79, 435-452.
Burke RE, Levine DN, Tsairis P & Zajac FE (1973). Physiological types and histochemical profiles in motor
units of the cat gastrocnemius. Journal ofPhysiology (London) 234, 723-748.
Burke RE, Rudomin P & Zajac FE (1976). The effect of activation history on tension production by individual
motor units. Brain Research 109, 515-529.
Calancie B & Bawa P (1985). Voluntary and reflexive recruitment of flexor carpi radialis motor units in
humans. Journal of Neurophysiology 53,1194-1200.
Calancie B & Bawa P (1986). Limitations of the spike-triggered averaging technique. Muscle & Nerve 9,78-83.
Cooper S & Eccles JC (1930). The isometric response of mammalian muscles. Journal ofPhysiology (London)
69,377-385.
Datta AK & Stephens JA (1980). Short-term synchronization of motor unit firing in human first dorsal
interosseous. Journal ofPhysiology (London) 308, 19-20P.
Datta AK & Stephens JA(1981). The effects of digital nerve stimulation on the firing of motor units in human
first dorsal interosseous muscle. Journal of Physiology (London) 318, 501-510.
Dengler R, Konstanzer A, Kiither G, Hesse S, WolfW, & Struppler A (1990). Amyotrophic lateral sclerosis:
macro-EMG and twitch forces of single motor units. Muscle & Nerve 13, 545-550.
Dengler R, Stein RB, & Thomas CK (1988). Axonal conduction velocity and force of single human motor
units. Muscle & Nerve 11,136-145.
Desmedt IE & Godaux E (1977a). Fast motor units are not preferentially activated in rapid voluntary
contractions in man. Nature 267,717-719.
Desmedt IE & Godaux E (1977b). Ballistic contractions in man: characteristic recruitment patterns of single
motor units of the tibialis anterior muscle. Journal of Physiology (London) 264, 673-693.
Desmedt IE & Godaux E (1981). Spinal motoneuron recruitment in man: rank deordering with direction but
not with speed of voluntary movement. Science 214,933-936.
Doherty TJ & Brown WF (1994). A method for the longitudinal study of human thenar motor units. Muscle
and Nerve 17, 1029-1036.
Dubose L, Schelhorn TB & Clamann HP (1987). Changes in contractile speed of cat motor units during activity.
Muscle & Nerve 10, 744-752.
Human Motor Units 159

Duchateau J & Hainaut K (1990). Effects of immobilization on contractile properties, recruitment and firing
rates of human motor units. Journal ofPhYSiology (London) 422, 55-65.
Freund H-J, Budingen H-J & Dietz V (1975). Activity of single motor units from human forearm muscles
during voluntary isometric contractions. Journal of Neurophysiology 38, 933-946.
Fuglevand AJ, Macefield VG & Bigland-Ritchie B (1993). Twitch properties of human toe extensor motor
units. Society for Neuroscience Abstract 19, 154.
Gandevia SC & McKenzie DK (1988). Activation of human muscles at short muscle lengths during maximal
static efforts. Journal ofPhysiology (London) 407, 599-613.
Garnett RAF, O'Donovan MJ, Stephens JA & Taylor A (1979). Motor unit organisation of human medial
gastrocnemius. Journal of Physiology (London) 287, 33-43.
Goldberg LJ & Derfler B (1977). Relationship among recruitment order, spike amplitude, and twitch tension
of single motor units in human masseter muscle. Journal ofNeurophysiology 40, 879-890.
Grimby L & Hannerz J (1974). Differences in recruitment order and discharge pattern of motor units in the
early and late flexion reflex components in man. Acta Physiologica Scandinavica 90, 555-564.
Gydikov A, Dimitrov G, Kosarov D & Dimitrova N (1976). Functional differentiation of motor units in human
opponens pollicis muscle. Experimental Neurology 50,36-47.
Hannerz J & Grimby L (1979). The afferent influence on the voluntary firing range of individual motor units
in man. Muscle & Nerve 2, 414-422.
Henneman E, Somjen G & Carpenter DO (1965). Functional significance of cell size in spinal motoneurons.
Journal ofNeurophysiology 28,560-580.
Howell IN, Fuglevand AJ, Walsh ML & Bigland-Ritchie B (1994). Motor unit firing during shortening and
lengthening contractions. Society for Neuroscience Abstract 20, 1759.
Johansson RS, Thomas CK, Westling G & Bigland-Ritchie B (1988a). A new method for examining contractile
properties of human single motor units. European Journal of Physiology 411, S 1 R196.
Johansson RS & Westling G (1988). Coordinated isometric muscle commands adequately and erroneously
programmed for the weight during lifting task with precision grip. Experimental Brain Research 71,
59-71.
Johansson RS, Westling G, Thomas CK & Bigland-Ritchie B (1988b). Recording human single motor unit
properties: a new method. Society for Neuroscience Abstract 14, 1232.
Kernell D, Ducati A & Sjiiholm H (1975). Properties of motor units in the first deep lumbrical muscle of the
cat's foot. Brain Research 98, 37-55.
Kernell D, Eerbeek 0 & Verhey BA (1983). Relation between isometric force and stimulation rate in cat's
hindlimb motor units of different twitch contraction time. Experimental Brain Research 50,220-227.
Kukulka CG & Clarnann HP (1981). Comparison of the recruitment and discharge properties of motor units in human
brachial biceps and adductor pollicis during isometric contractions. Brain Research 219, 45-55.
Lim KY, Thomas CK & Rymer WZ (1995). Computational methods for improving estimates of motor unit
twitch contraction properties. Muscle & Nerve 18, 165-174.
Marsh E, Sale D, McComas AJ & Quinlan J (1981). Influence ofjoint position on ankle dorsiflexion in humans.
Journal ofApplied Physiology 51,160-167.
McKeon B & Burke D (1983). Muscle spindle discharge in response to contraction of single motor units.
Journal ofNeurophysiology 49, 291-302.
Mendell LM & Henneman E (1968). Terminals of single Ia fibers: Distribution within a pool of 300
homonymous motor neurons. Science 160, 96-98.
Milner-Brown HS, Stein RB & Lee RG (1974). Pattern of recruiting human motor units in neuropathies and
motor neurone disease. Journal of Neurology. Neurosurgery & Psychiatry 37, 665-669.
Milner-Brown HS, Stein RB & Yemm R (1973a). The contractile properties of human motor units during
voluntary isometric contractions. Journal of Physiology (London) 228, 285-306.
Milner-Brown HS, Stein RB & Yemm R (1973b). The orderly recruitment of human motor units during
voluntary isometric contractions. Journal of Physiology (London) 230, 359-370.
Monster AW & Chan H (1977). Isometric force production by motor units in the extensor digitorum communis
muscle in man. Journal ofNeurophysiology 40,1432-1443.
Nardone A, Romano C & Schieppati M (1989). Selective recruitment of high-threshold human motor units
during voluntary isotonic lengthening of active muscles. Journal of Physiology (London) 409,
451-471.
Nardone A & Schieppati M (1988). Shift of activity from slow to fast muscle during voluntary lengthening
contractions of the triceps surae muscle in humans. Journal ofPhysiology (London) 395, 363-381.
Nordstrom MA & Miles TS (1990). Fatigue of single motor units in human masseter. Journal of Applied
Physiology 68, 26-34.
160 C.K. Thomas

Nordstrom MA, Miles TS & Veale JL (1989). Effect of motor unit firing pattern on twitches obtained by
spike-triggered averaging. Muscle & Nerve 12, 556-567.
Olson CB & Swett CP (1971). Effect of prior activity on properties of different types of motor units. Journal
ofNeurophysiology 34, 1-16.
Person RS (1974). Rhythmic activity of a group of human motoneurons during voluntary contraction of a
muscle. Electroencephalography 36, 585-595.
Sica REP & McComas AJ (1971). Fast and slow twitch units in a human muscle. Journal of Neurology,
Neurosurgery & Psychiatry 34,113-120.
Smith A, Zimmerman GN & Abbas PJ (1981). Recruitment patterns of motor units in speech production.
Journal of Speech and Hearing Research 24, 567-576.
Stein RB, Brucker BS & Ayyar DR (1990). Motor units in incomplete spinal cord injury: electrical activity,
contractile properties and the effects of biofeedback. Journal ofNeurology. Neurosurgery & Psychia-
try 53, 880-885.
Stein RB, French AS, Mannard D & Yemm R (1972). New methods for analyzing motor function in man and
animals. Brain Research 40, 187-192.
Stephens JA Garnett R & Buller NP (1978). Reversal of recruitment order of single motor units produced by
cutaneous stimulation during voluntary muscle contraction in man. Nature 272, 362-364.
Stephens JA & Usherwood TP (1977). The mechanical properties of human motor units with special reference
to their fatiguability and recruitment threshold. Brain Research 125,91-97.
Taylor A & Stephens JA (1976). Study of human motor unit contractions by controlled intramuscular
microstimulation. Brain Research 117, 331-335.
Ter Haar Romeny BM, Denier van der Gon JJ & Gielen CAM (1982). Changes in recruitment order of motor
units in the human biceps muscle. Experimental Neurology 78, 360-368.
Thomas CK, Bigland-Ritchie B & Johansson RS (1991a). Force-frequency relationships of human thenar
motor units. Journal ofNeurophysiology 65,1509-1516.
Thomas CK, Bigland-Ritchie B, Westling G & Johansson RS (1990a). A comparison of human thenar motor
unit properties studied by intraneural motor axon stimulation and spike-triggered averaging. Journal
ofNeurophysiology 64,1347-1351.
Thomas CK, Johansson RS & Bigland-Ritchie B (1991b). Attempts to physiologically classify human thenar
motor units. Journal ofNeurophysiology 65, 1501-1508.
Thomas CK, Johansson RS, Westling G & Bigland-Ritchie B (1990b). Twitch properties of human thenar
motor units measured in response to intraneural motor axon stimulation. Journal ofNeurophysiology
64,1339-1346.
Thomas CK, Pocock ME & Evans JJ (1994). Fatigue and post-fatigue responses to sustained MVCs after
chronic cervical spinal cord injury. Neural and Neuromuscular Aspects ofMuscle Fatigue Abstracts
H20, 39.
Thomas CK, Ross BH & Calancie B (1987a). Human motor-unit recruitment during isometric contractions
and repeated dynamic movements. Journal ofNeurophysiology 57, 311-324.
Thomas CK, Ross BH & Stein RB (1986). Motor-unit recruitment in human first dorsal interosseous muscle
for static contractions in three different directions. Journal ofNeurophysiology 55, 1017-1029.
Thomas CK, Stein RB, Gordon T, Lee RG & Elleker MG (1987b). Patterns of reinnervation and motor unit
recruitment in human hand muscles after complete ulnar and median nerve section and resuture.
Journal of Neurology, Neurosurgery & Psychiatry 50, 259-268.
Thomas JS, Schmidt EM & Hambrecht FT (1978). Facility of motor unit control during tasks defined directly
in terms of unit behaviors. Experimental Neurology 59, 384-395.
Vallbo, AB (1970). Discharge patterns in human muscle spindle afferents during isometric voluntary contrac-
tions. Acta Physiologica Scandinavica 78, 315-333.
Westling G, Johansson RS, Thomas CK & Bigland-Ritchie B. (1990). Measurement of contractile and
electrical properties of single human motor units in response to intraneural motor axon stimulation.
Journal ofNeurophysiology 64, 1331-1338.
Yemm R (1977). The orderly recruitment of motor units of the masseter and temporal muscles during voluntary
isometric contraction in man. Journal ofPhYSiology (London) 265, 163-174.
Young JL & Mayer RF (1981). Physiological properties and classification of single motor units activated by
intramuscular microstimulation in the first dorsal interosseous muscle in man. In: Desmedt JE (ed.),
Motor Unit Types, Recruitment and Plasticity in Health and Disease. Progress in Clinical Neuro-
physiology 9, pp. 17-25. Basel: Karger.
Yue G, Fuglevand AJ, Nordstrom MA & Enoka RM (1995). Limitations of the surface-EMG technique for
estimating motor unit synchronization. Biological Cybernetics In press.
 11

HUMAN MOTOR UNITS STUDIED BY

INTRAMUSCULAR MICRO STIMULATION

J. M. Elek and R. Dengler

Department of Neurology and Clinical Neurophysiology

Medical School of Hannover
D-30623 Hannover, Germany

ABSTRACT

We recorded twitch parameters of motor units of the human fIrst dorsal interosseus muscle
applying low-rate intramuscular microstimulation of motor axons (IMS) or spike-triggered
averaging (STA). The values of contraction time, half-relaxation time and maximal rate of
rise of force were signifIcantly smaller in the STA studies. The reduction corresponded to
the effect expected from partial twitch fusion. Twitch amplitudes, however, were not smaller
than those in the IMS studies, indicating that other factors, such as motor unit synchroniza-
tion, compensate for the effects of partial fusion.

INTRODUCTION

Single motor units in human muscles can be easily studied electromyographically by

assessing their action potentials. The associated motor unit contractions, however, are more
difficult to measure, although they represent the more relevant parameter from a functional
point of view and could contribute to the pathophysiological understanding of diseases of
the motor system (Young & Mayer, 1982, Dengler et aI., 1990, Yang et aI., 1990). Several
methods have been developed to assess motor unit twitch parameters in human muscles. The
two techniques most widely used are spike-triggered averaging (STA) (Stein et aI., 1972)
and intramuscular microstimulation of individual axons of motor nerves (lMS; Taylor &
Stephens, 1976). Intraneural microstimulation of motor axons (Thomas et aI., 1990a, 1990b)
is another elegant technique to assess motor unit contractions and has been reviewed by
Thomas, Chapter 10.
STA is carried out during voluntary motor unit activity and does not involve
stimulation. It is easy to apply and provides both contraction parameters and the voluntary
recruitment threshold of a motor unit (Stein et aI., 1972; Milner-Brown et aI., 1973a, 1973b;
Stephens & U sherwood, 1977).
However, it has some methodological limitations: I) motor unit twitches may be
distorted by partial fusion of twitch responses (Monster & Chan, 1977; Andreassen &

161
162 J. M. Elek and R. Dengler

Bar-On, 1983; Calancie & Bawa, 1986, Nordstrom et aI., 1989); 2) synchronization of
simultaneously firing motor units may lead to an overestimation of twitch forces (Milner-
Brown et aI., 1973a); and 3) sampling of motor units may be biased towards slow,
fatigue-resistant, low-threshold motor units (Calancie & Bawa, 1986). The technique of
low-rate IMS circumvents the problems associated with STA (Elek et aI., 1992). We,
therefore, compared twitch parameters of a representative number of normal motor units
obtained by STA and IMS in a human hand muscle. In addition, data from patients with
diseases with various forms of chronic partial denervation are presented.

MATERIAL AND METHODS

Thirty-seven normal subjects were investigated, 12 by STA (6 men and 6 women,

mean age 36.5 yrs., range 23 to 63 yrs.) and 25 by IMS (15 men and 10 women, mean age
36 yrs., range 23 to 87 yrs.). Studies were approved by the local ethics committee and all
subjects gave informed consent.

Force Measurements
The first dorsal interosseus muscle of the hand was investigated. The same set-up
was used for STA- and IMS-experiments. An isometric strain gauge was positioned at the
radial side of the index finger with a lever arm of 6.5 cm from the metacarpo-phalangeal
joint, and measuring the contraction force of the first dorsal interosseus muscle in the
direction of abduction only.

STA-Experiments
A total of 236 motor units was investigated using STA. Subjects produced stable
isometric contractions at force levels sufficient to recruit a motor unit and keep it firing regularly
(tonic recruitment threshold, up to 60% of maximal voluntary force) at low rate (8 to 12 Hz).
Motor unit action potentials were picked up by the single fiber lead of a macro-EMG electrode
(Stalberg & Fawcett, 1982) and by a pair of surface electrodes. The single-fiber signal triggered
a computer to average the surface EMG potentials (sampling rate 5 kHz) and the twitch
contractions (sampling rate 3 kHz) of the motor unit (200 to 500 trials). Twitches were recorded
using a high-gain, AC-coupled version ofthe force signal (bandwith 0.3 to 100 Hz). The surface
EMG potentials were used to identify different motor units.

IMS-Experiments
251 motor units were studied. A bipolar needle electrode was inserted into the first
dorsal interosseus muscle at the border between the proximal and middle third to stimulate
single motor axons. Surface electrodes were placed over the muscle belly and the knuckle of
the index finger to record the electromyographic responses. The position of the needle electrode
and the stimulus intensity (up to 3 rnA, duration 0.05 to 0.1 ms) were adjusted to evoke a
reproducible all-or-none EMG potential representing the response of a single motor unit.
Stimuli were applied at a low rate (0.9 Hz) to obtain individual twitch contractions. Force
signals were low-pass filtered (0 to 100 Hz) and amplified at high gain. The evoked EMG
responses were monitored continuously and all trials with irregularities, e.g. missing and
occasionally occurring F-responses, were rejected on-line. Twenty to 50 trials were averaged.
In 10 subjects with particularly stable experimental conditions, stimuli could also be
applied at higher rates (up to 10 Hz in 20 motor units, up to 12 Hz in 15 motor units and up
Human Motor Units Studied by Intramuscular Microstimulation 163

~ _ _I_o'2mv
120mN
Figure 1. Averaged EMG response (upper trace) and
twitch contraction (lower trace) of a single motor unit
obtained by IMS at 0.9 Hz: twitch force (TF); maximum
rate of rise offorce (MRRF); contraction time (CT); half
relaxation time (HRT).

to 14 Hz in 10 motor units) to induce partial fusion of twitches mimicking the situation in

STA. Averaging was started when the contraction had stabilized after 3 to 5 stimuli.
Parameters were measured from the unfused force fluctuations (trough to peak).

Twitch parameters
The following parameters were assessed off-line for twitches in both STA and IMS
studies (Fig. 1): twitch force (TF), contraction time (CT), half relaxation time (HRT) and
maximal rate of rise of force (MRRF).

Statistical Procedures
For statistical comparison of twitch parameters, medians were compared using the
Mann-Whitney test in case of skewed distributions (twitch force and maximal rate of rise of
force, see below); for normal distributions (contraction and half-relaxation times, see below)
means were compared with student's t-test.

RESULTS

Comparison of Motor Unit Twitch Parameters Obtained by STA and

IMS
Twitch force: the distributions of twitch forces obtained by STA and IMS are
illustrated in Fig. 2, left panel. They appeared fairly similar and were both skewed to the
right. The mean and median values of the STA twitchforce were slightly larger than those of
the IMS twitch force (Table 1). The difference, however, was not significant (Mann-Whitney
test). Another slight difference was seen in the low force spectrum. With STA, 13 motor units
(6.5%) revealed twitch forces smaller than 1 mN, while the smallest twitch force value
obtained by IMS was 1 mN.
Maximal rate of rise of force: the distributions of maximal rate of rise of forces
resembled those of the twitch forces (Fig. 2, right panel). The median values obtained by
STA, however, were significantly smaller than the IMS-values (p < 0.05, Mann-Whitney
test, Table 1).
Contraction and half-relaxation times: Both parameters showed an approximately
Gaussian distribution for both STA and IMS (Fig. 3). The mean values obtained by STA,
164 J. M. Elek and R. Dengler

0.3 STA 0.3 SfA

0.2 0.2

0.1 0.1
~
c:
...
:::l
0.0 0.0
~ 0 40 80 120 160 0 2 3 4 5 6
E
"0
i
..c
E
:::l
c: 0.3 IMS 0.3 IMS
Q)

~
Qi
II: 0.2 0.2

0.1 0.1

0.0 0.0
0 40 80 120 160 0 2 3 4 5 6
TF [mN] MRRF [N/s]

Figure 2. Distributions of twitch forces (TF, left) and maximal rate of rise of force values (MRRF, right)
obtained by STA (top) and by IMS (bottom).

however, were significantly smaller for both contraction time and half-relaxation time than
the corresponding IMS values (p < 0.05, t-test, Table 1).

Effect of Stimulus Rate on Twitch Parameters in IMS

Fig. 4 illustrates the effect of stimulus rate (0.9 to 14 Hz) on twitch parameters
obtained with IMS. Values were normalized to the corresponding values at 0.9 Hz stimula-
tion. A reduction of all twitch parameters with increasing stimulus rate was found beyond 4
Hz . Most prominent was the decrease of twitch force reaching a relative value of 0.45 ±
0.20 at 10Hz while maximal rate of rise of force proved to be fairly robust with 0.88 ± 0.18.

Table 1. Motor unit twitch parameters obtained by IMS and STA

Expected Values
IMS (n = 251) STA(n = 236) for 10 Hz
Mean± SD Median Mean± SD Median Mean Median
TF (mN) (14±15) 9 (17.7±19.8) 10.3 (6.3) 4.1
MRRF (N/s) (0.79±0.53) 0.61 (0.76±0.84) 0.42* (0.70) 0.54
CT (ms) 64±14 (63) 47.3±l2.8* (44.8) 42 (42)
HRT (ms) 61±l6 (59) 33.9±lO.3* (33.6) 36 (35)
*significant (p < 0.05) difference between values obtained by IMS and STA.
Human Motor Units Studied by Intramuscular Microstimulation 165

0.3 STA 0.3 STA

0.2 0.2

0.1 0.1
~
c:
:::I
l5 0.0 0.0
15 0 40 80 120 0 40 80 120
E
'0
(jj
..c
E
:::I
c: 0.3 IMS 0.3 IMS
Ql

~
Qj
II: 0.2 0.2

0.1 0.1

0.0 0.0
0 40 80 120 0 40 80 120

CT ems] HRT ems]

Figure 3. Distributions of contraction times (CTs, left) and half-relaxation time (HRTs, right) obtained by STA
(top) and by IMS (bottom).

Contraction time and half-relaxation time decreased moderately to 0.66 ± 0.10 and 0.59 ±
0.15, respectively (at 10 Hz).
The above relative values were multiplied by the mean and median values to give
expected twitch parameters at 10Hz for the whole sample of motor units studied with IMS.
The resulting expected mean and median values (Table 1, right columns) were then compared
with the STA data (Table I, middle columns) which were collected at very similar, voluntary
firing rates of 8 to 12 Hz. The expected values for contraction time and half-relaxation time,
to a lesser extent maximal rate of rise offorce corresponded nicely to those obtained by STA.
The twitch forces obtained by STA, however, were much larger than the expected values and
resembled those recorded by IMS at low stimulus rate.

MOTOR UNIT TWITCH PARAMETERS IN DISEASES WITH

CHRONIC PARTIAL DENERVATION
We investigated 31 patients with chronic partial denervation using low-rate IMS: 8
patients with hereditary motor and sensory neuropathy type I, age 13 to 52 yrs., mean 40
yrs.), 6 patients with spinal muscular atrophy (age 21 to 51 yrs., mean 36 yrs.) and 17 patients
with amyotrophic lateral sclerosis (age 28 to 70 yrs., mean 50 yrs.).
Maximal voluntary force of abduction of the first dorsal interosseus muscle was
reduced in all patient groups as compared to controls (14 - 55 N, mean 26.4 ± 10.2 N);
hereditary motor and sensory neuropathy type I 0.4 - 33 N, mean 10.9 ± 10.6 N; spinal
166 J. M. Elek and R. Dengler

IMS
1.4 1.4
1.2 1.2

~iit1 ~
1.0 ~ 1.0
I
..!... 0.8 ~

u.
0.8
u. 0.6 a:: 0.6
f-
a:: a::
::;;
0.4 a:: 0.4
0.2 0.2
0.0 0.0
a 2 4 6 8 10121416 a 2 4 6 8 10121416

1.2 1.2
1.0 1.0

~
0.8 0.8
I I'
'--'
0.6
~

f- f- 0.6
u a::
0:: 0.4 :x:
a:: 0.4
0.2 0.2
0.0 0.0
a 2 4 6 8 10121416 a 2 4 6 8 10 121416

STIMULUS RATE [ijs]

Figure 4. Mean values and standard deviations (20 motor units) of relative motor unit twitch parameters at
different stimulus rates: RTF, RMRRF, RCT and RHRT - relative twitch parameters normalized to individual
unfused twitch values ofTF, MRRF, CT and HRT at 0.9 Hz, respectively.

muscular atrophy 0.5 - 17 N, mean 8.4 ± 6.8 N; amyotrophic lateral sclerosis 0.4 - 28 N,
mean 9.2 ± 7.9 N.

Motor Unit Twitch Parameters in Hereditary Motor and Sensory

Neuropathy Type I
The mean and median values of twitch parameters are summarized in Table 2 for 64
units. Overall, twitch forces did not differ significantly from controls, maximal-rate-of-rise-

Table 2. Motor unit twitch parameters in hereditary motor and

sensory neuropathy type I (HMSN I)
controls HMSNI MVC>5N MVC::;5N
(n=251) (n = 64) (n = 46) (n = 18)
TF (mN) 9 9 12* 4*
MRRF(N/s) 0.61 0.46* 0.46* 0.30*
CT (ms) 64±14 74±20* 7l±16* 83±27*
HRT(ms) 61±16 70±23* 71±24* 71 ± 24*
Median values for TF and MRRF; mean ± SD for CT and HRT
*significant difference (p < 0.05) compared to controls.
Human Motor Units Studied by Intramuscular Microstimulation 167

Table 3. Motor unit twitch parameters in spinal muscular atrophy (SMA)

Controls SMA SMA (MVC > 5 N) SMA(MVC :0;5 N)
(n = 251) (n = 32) (n = 28) (n = 4)
TF (mN) 9 15* 20* 9
MRRF (N/s) 0.61 0.83* 0.88 0.82
CT (ms) 64±14 59±10 59±9 58±18
HRT (ms) 61±16 57±16 58±!7 42±6
Median values for TF and MRRF, means ± SD for CT and HRT.
*Significant difference (p < 0.05) compared to controls.

of-force medians were significantly reduced, and mean contraction and half-relaxation times
were significantly prolonged.
Patients were then subdivided in those with a severely affected first dorsal interosseus
muscle with a maximal voluntary contraction (MVC) force of 5 N or less and those with a
slightly or moderately affected first dorsal interosseus muscle (MVC > 5 N). In slightly/mod-
erately affected muscles twitch force medians were significantly larger than in controls. The
severely affected muscles, however, showed significantly smaller twitch force medians with
a more pronounced decrease in maximal-rate-of-rise-of-force medians and prolongation of
contraction time means (Table 2).

Motor Unit Twitch Parameters in Spinal Muscular Atrophy

In the 32 motor units investigated, twitch force and maximal-rate-of-rise-of-force
medians were significantly increased compared to controls. Interestingly, twitch force
medians were also increased in the subgroup with MVC > 5 N, but not in the subgroup with
MVC > 5 N. No difference was found for contraction and half-relaxation times (Table 3).

Motor Unit Twitch Parameters in Amyotrophic Lateral Sclerosis

In the amyotrophic lateral sclerosis group (112 motor units), twitch force medians
were significantly increased, contraction times were slightly but significantly prolonged. In
both subgroups (MVC > 5 Nand MVC ::; 5 N) twitch force medians were about equal and
significantly increased, contraction time means were only prolonged in the subgroup with
MVC force::; 5 N (Table 4).

Table 4. Motor unit twitch parameters in amyotrophic lateral

sclerosis (ALS)
controls ALS (MVC >5 N) (MVC:O;5 N)
(n=251) (n = 112) (n = 67) (n = 45)
TF (mN) 9 17* 16* 17*
MRRF (N/s) 0.61 0.69 0.71 0.54
CT (ms) 64±14 69±21* 65±23 75±16*
HRT (ms) 61±16 62±15 61±16 64±14
Median values for TF and MRRF, means ± SD for CT and HRT.
*Significant difference (p < 0.05) compared to controls.
168 J. M. Elek and R. Dengler

DISCUSSION

STA is easier to apply and has a somewhat higher yield (about 8 motor units per hour
of investigation). It has, however, methodological limitations as pointed out in the introduc-
tion. Furthermore, it requires that the subject drive a motor unit at a low and constant rate
which may be difficult for patients with motor disorders. It should also be kept in mind that
STA only allows investigation of motor units that can be activated voluntarily. This may be
of importance when investigating patients with peripheral or central weakness. IMS is more
difficult to apply and has a somewhat lower yield (about 4-5 motor units per hour of
investigation), particularly in patients with partially denervated muscles. It is carried out
with the muscle completely relaxed; and, therefore, needs no assistance from the subject and
precludes motor unit synchronization. Motor units are investigated independently of whether
they can be recruited voluntarily or not. Employing low stimulus rates, IMS also circumvents
twitch distortion associated with partial twitch fusion.
It is generally assumed that the all-or-nothing response evoked by IMS is produced
by a single entire motor unit. On the other hand it appears possible that IMS may not evoke
a response of all muscle fibers of a given motor unit due to stimulation of terminal axon
branches. Although this cannot be ruled out entirely, it is unlikely to play an essential role
(see Elek et al., 1992). An argument against partial activation of motor units with IMS is the
finding that frequently occurring F-responses (in about 30% of motor units) show the same
wave forms as the preceding M-responses (Dengler et ai., 1992). Therefore, we assume that
motor axons proximal to terminal branching are the site of stimulation. From this argument,
a distortion of the twitches, particularly of the rising phase, due to unphysiological an-
tidromic action potential propagation from a stimulated terminal axon branch into others
also appears very improbable.
Direct muscle fiber stimulation occurs occasionally and produces small short-latency
EMG responses that characteristically increase continuously, not stepwise, with increasing
stimulus intensity (EMG amplitude not above [Link]). The associated forces are extremely
small (below 1 mN) and often unmeasurable. It cannot be excluded, however, that stimulation
of individual motor axons also activates muscle fibers in the close vicinity of the stimulation
electrode that belong to other motor units that could result in a slight overestimate of twitch
forces.
We compared motor unit twitch parameters in the human first dorsal interosseus
muscle obtained by STA and IMS. Our results show relevant differences only for the time
dependent parameters: The values of maximal rate of rise of force, contraction time and
half-relaxation time are significantly lower with STA than with IMS, whereas the values of
twitch force reveal no significant difference. This finding is unexpected with regard to the
results of the IMS experiments employing higher stimulus rates where partial fusion
decreased the twitch force. The values of the time dependent twitch parameters in the STA
studies correspond well with the high-rate IMS results (10 Hz) which stresses the role of
partial fusion of twitches in STA. In the case of twitch force, however, other factors must
counteract or compensate the effects of partial fusion in STA.
Milner-Brown and colleagues (1973a) compared twitch parameters ofa small number
of motor units in the human first dorsal interosseus muscle studied by STA and IMS. They
did not find relevant differences for twitch forces and contraction times whereas half-relaxa-
tion times were underestimated by STA. In a larger sample of about 200 motor units in human
thenar muscles, Stein and Yang (1990) obtained about the same twitch forces with STA and
IMS. The predicted effect of partial twitch fusion was also not evident in the data of Thomas
and coworkers (1990a). To their surprise, they found STA twitch forces significantly stronger
Human Motor Units Studied by Intramuscular Microstimulation 169

and STA contraction times significantly longer than the corresponding values obtained by
intraneural microstimulation.
Which factors other than partial fusion influence twitch parameters in STA? An
important candidate is synchrony between firing times of different motor units in a volun-
tarily active muscle (Nordstrom et aI., 1989). In fact, even in very smooth muscle contrac-
tions, simultaneously active motor units do not only show chance synchronization but
frequently also show a more than chance synchronization of the so called short-term type
(Sears & Stagg, 1976, Dengler et aI., 1984). In favor of the role of motor unit synchronization
is our finding that the weakest motor units have smaller twitch force values when STA is
used. These motor units are recruited early (Milner-Brown et aI., 1973b) when only few
motor units are active and when motor unit synchronization would be negligible. Therefore,
partial fusion dominates in this situation and causes lower twitch forces than obtained with
methods using low rate motor axon stimulation.
Differences in the activation history of motor units during STA and IMS may also
influence twitch contractions obtained by both techniques. For STA, twitch contractions are
averaged following up to 500 motor unit discharges, i.e., over a period of at least 1 min given
a mean firing rate of about 10Hz. In addition, this motor unit may have already been active
during the averaging of previously investigated motor units. This may lead to a potentiation
oftwitch contractions (Thomas et aI., 1990b) and could also partly explain the unexpectedly
high twitch forces obtained with STA as compared to low-rate IMS where twitches are
averaged following 25 to 50 stimuli at lis. On the other hand fatigue develops during STA
even in normal subjects after about 5 min of constant motor unit activation at 10Hz (Stephens
& U sherwood, 1977), whereas fatigue is less likely to occur during low-rate IMS. This should
be taken into account when STA is used to investigate patients who might fatigue more
rapidly.
Another factor that may mask the effects of partial fusion on twitch forces in STA is
a more-than-linear summation of contraction forces of simultaneously active motor units as
has been described by Clamann and Schelhom (1988). In cat hindlimb muscles, they found
that the combined forces of two motor units exceeded the algebraic sum of the separate forces
by an average of 5 to 12%, probably due to mechanical interaction of contracting muscle
fibers.
It should further be considered that mechanical conditions (viscous and elastic
properties of muscle, tissue and tendon) are different in STA and IMS. While motor unit
twitches are superimposed on a more or less stable background tension during STA, IMS is
carried out in a relaxed muscle. The lower stiffness of muscle and tendon in the relaxed state
may lead to longer contraction times and reduced twitch forces with IMS. In animal
experiments, however, McPhedran and colleagues (1965) stated that motor unit twitch
parameters were not distorted when recorded from a relaxed muscle, so the effect of these
mechanical factors is probably weak.
The distribution of twitch forces is skewed to the right with both STA and IMS; i.e.,
the motor units with the smallest forces are the most frequent. Whereas such a distribution
can result from a sampling bias in favor of small, i.e., low-threshold motor units in STA; this
can hardly be the case with IMS. One would rather expect a bias towards large, i.e., strong
motor units with IMS because they have thicker axons and should be more readily excited.
Other factors, however, such as the distance of the axon from the stimulating electrode may
be more important. Similar to our study, Stein and Yang (1990) did not find significant
differences in twitch forces and in amplitudes of motor unit potentials obtained by STA and
IMS, so the same group of motor units appears to be sampled by either technique.
To illustrate the use of IMS for clinical studies we obtained in patients with chronic
partial denervation. In brief, we found an increase of twitch forces which corresponds to an
enlargement of motor units due to collateral axonal sprouting. The prolonged contraction
170 J. M. Elek and R. Dengler

times probably reflect slowed conduction velocities of sprouted axons, leading to a less
synchronized contraction of muscle fibers which in the case of hereditary motor and sensory
neuropathy type I also resulted in prolonged half-relaxation times. In patients with severely
denervated muscles, however, twitch forces were reduced. This confirms findings of a
previous study using STA (Dengler et aI., 1990) which also showed a decrease of twitch
forces in severely denervated muscles although macro EMG potentials were not decreased.
An impaired function of the contractile apparatus was hypothesized, although an increased
fatigability during the STA procedure could not be ruled out. The present data obtained with
low-rate IMS clearly exclude the latter possibility.
Investigations of fatigue in human muscles with either STA or IMS have only rarely
been carried out. The use of both techniques for fatigue studies is limited as the recording
(STA) or stimulating electrode (IMS) inserted into the muscle is very easily shifted during
fatiguing contractions. Using STA, Stephens and Usherwood (1977) studied motor unit
twitch contractions in the human first dorsal interosseus muscle before and after a fatiguing
paradigm (voluntary motor unit activation at about 10Hz over 5 min). According to their
results, motor units recruited at higher strengths had higher twitch forces, faster contraction
times and were highly fatigable which is in line with results from animal experiments. In
motor units of the human masseter muscle Nordstrom and Miles (1990) did not find a
correlation of twitch forces, contraction times and fatigability with STA (fatiguing paradigm:
voluntary activation of a motor unit at 10Hz over 15 min) which they attributed to the unusual
histochemical features of the masseter muscle. Fatigue tests with IMS have been carried out
by Young and Mayer (1982) in the first dorsal interosseus muscle of normal subjects and
patients with hemiplegia. They classified motor units as S, FF and FR based on twitch
parameters, sag (8-15 Hz stimulation) and fatigability (30 Hz stimulation for 300 ms repeated
once per second over 2 min). In hemiplegic patients, they also found slow-twitch fatigable
motor units which they called SF. Comparing data obtained in controls and patients they
inferred an increased fatigability of motor units in long-term spastic hemiparesis (see also
McComas et aI., Chapter 36).

ACKNOWLEDGMENTS
The authors' laboratory has been supported by the Federal Ministry for Research and
Technology, Germany (BMFT), and the German Research Society (DFG). Attendance of the
authors at the 1994 Bigland-Ritchie conference was supported, in part, by the Medical School
of Hannover, the United States Public Health Service (National Center for Medical Reha-
bilitation Research, National Institute for Child Health and Human Development; R.D.), and
the Muscular Dystrophy Association (USA).

REFERENCES
Andreassen S & Bar-On E (1983). Estimation of motor unit twitches. IEEE Transactions of Biomedical
Engineering 30, 742-748.
Calancie B & Bawa P (1986). Limitations of the spike-triggered averaging technique. Muscle & Nerve 9,
78-83.
Clamann HP & Schelhom TB (1988). Nonlinear force addition of newly recruited motor units in the cat
hindlimb. Muscle & Nerve 11,1079-1089.
Dengler R, Konstanzer A, Kiither G, Hesse S, WolfW & Struppler A (1990). Amyotrophic lateral sclerosis:
macro- EMG and twitch forces of single motor units. Muscle & Nerve 13, 545-550.
Dengler R, Kossev A, Wohlfarth K, Schubert M, Elek J & WolfW (1992). F-waves and motor unit size. Muscle
& Nerve IS, 1138-1142.
Human Motor Units Studied by Intramuscular Microstimulation 171

Dengler R, WolfW, Birk P & Struppler A (1984). Synchronous discharges in pairs of steadily firing motor
units tend to form clusters. Neuroscience Letters 47, 167-172.
Elek 1M, Kossev A, Dengler R, Schubert M, Wohlfarth K & WolfW (1992). Parameters of human motor unit
twitches obtained by intramuscular microstimulation. Neuromuscular Disorders 2, 261-267.
McPhedran AM, Wuerker RB & Hennemann E (1965). Properties of motor units in a homogeneous red muscle
(m. soleus) of the cat. Journal ofNeurophysiology 28,71-84.
Milner-Brown HS, Stein RB & Yemm R (1973a). The contractile properties of human motor units during
voluntary isometric contractions. Journal of Physiology (London) 228, 285-306.
Milner-Brown HS, Stein RB & Yemm R (1973b). The orderly recruitment of human motor units during
voluntary isometric contractions. Journal of Physiology (London) 230, 359-370.
Monster AW & Chan H (1977). Isometric force production by motor units of extensor digitorum communis
muscle in man. Journal ofNeurophysiology 40, 1432-1443.
Nordstrom MA& Miles TS (1990). Fatigue of single motor units in human masseter muscle. Journal ofApplied
Physiology 68, 26-34.
Nordstrom MA, Miles T & Veale J (1989). Effect of motor unit firing pattern on twitches obtained by
spike-triggered averaging. Muscle & Nerve 12, 556-567.
Sears TA & Stagg D (1976). Short-term synchronization of intercostal motoneurone activity. Journal of
Physiology (London) 263, 357-381.
Stalberg E & Fawcett PRW (1982). Macro-EMG in healthy subjects of different ages. Journal of Neurology,
Neurosurgery & Psychiatry 45, 870-878.
Stein RB, French AS, Mannard A & Yemm R (1972). New methods for analysing motor function in man and
animals. Brain Research 49, 187-192.
Stein RB & Yang JF (1990). Methods for estimating the number of motor units in human muscles. Annals of
Neurology 28, 487-495.
Stephens JA & Usherwood TP (1977). The mechanical properties of human motor units with special reference
to their fatiguability and recruitment threshold. Brain Research 125,91-97.
Taylor A & Stephens JA (1976). Study of human motor unit contractions by controlled intramuscular
micro stimulation. Brain Research 117,331-335.
Thomas CK, Bigland-Ritchie B, Westling G & Johansson RS (1990a). A comparison of human thenar
motor-unit properties studied by intraneural motor-axon stimulation and spike-triggered averaging.
Journal of Neurophysiology 64,1347 -1351.
Thomas CK, Johansson RS, Westling G & Bigland-Ritchie B (1990b). Twitch properties of human thenar
motor units measured in response to intraneural motor-axon stimulation. Journal ofNeurophysiology
64, 1339-1346.
Yang JF, Stein RB, Ihamandas J & Gordon T (1990). Motor unit numbers and contractile properties after spinal
cord injury. Annals of Neurology 28, 496-502.
Young YL & Mayer RF (1982). Physiological alterations of motor units in hemiplegia. Journal ofNeurological
Sciences 54, 401-412.
SECTION IV
Fatigue Studied with NMR Techniques

This section contains four chapters that rely on the benefits afforded by non-invasive
measurements with nuclear magnetic resonance spectroscopy (NMR). It begins with a global
assessment of the metabolic homogeneity of muscle fiber types and ends with a glimpse of
the future measurements which will be possible with NMR.
In Chapter 12, Kushmerick uses data from 31p NMR and muscle histochemistry to
examine fatigue in muscles containing a range of different fiber types. Muscle-specific
differences in the levels of high-energy phosphates are less in humans than in commonly
used experimental animals. Furthermore, the constancy of the ratio of phosphocreatine to
ATP acts to normalize the bioenergetic function of different muscle cells. Hence, large-scale
measurements of cellular bioenergetics provided by NMR are a valid way to examine
comparatively homogeneous human muscles.
The metabolic responses of human muscles to different types of exercise are consid-
ered in Chapter 13 (V011estad). There are limits in the extent to which data from in vitro
studies can be applied to in vivo studies. These limits arise because the type and duration of
the fatiguing exercise and the pattern of neural drive to the muscle influence the extent to
which excitation-contraction coupling or ATP supply cause muscle fatigue. Of particular
interest are observations on repeated submaximal contractions in which force can be
impaired with little change in phosphocreatine, inorganic phosphate or pH. Phosphocreatine
levels are low at exhaustion, but these levels are already reached minutes before exercise
was voluntarily stopped. Evidence is presented that the energy cost of contractions (assessed
by oxygen consumption or heat output) increases progressively during isometric exercise
but not bicycling or running.
An integrated approach to studying human muscle fatigue is presented in Chapter
14 (Miller and colleagues). They review the studies that have applied the Nank technique to
human muscles especially during fatiguing exercise. Force, EMG and metabolites have been
measured during fatigue and recovery in a range of exercise protocols in both control subjects
and patients with neuromuscular disorders. As in the preceding chapter, the extent to which
excitation-contraction coupling can be reliably inferred from in vivo studies of human
muscles is of paramount importance. Initial glycogen stores and arterial pH contribute little
to the variability between subjects in high-intensity exercise performance and this confirms
the view that the fiber-type composition of each subject powerfully determines the exercise
performance. By measurement of the force produced in both voluntary and stimulated
contractions, the added contribution of central fatigue to reductions in force was evaluated.
The goal of these studies is to apportion the fatigue in a particular task to the various
174 Fatigue Studied with NMR Techniques

intracellular components, as well as those that may develop at the neuromuscular junction
or within the eNS.
In Chapter 15, Bertocci outlines the future directions for NMR methods that may
be applied to study fatigue. In addition to its established role in assessment ofthe high-energy
phosphates, it is possible to follow the precise incorporation of substrates and the regulation
of the citric acid cycle with 13e NMR spectroscopy. An advantage is that 13e is a stable,
non-radioactive carbon isotope. By its use, it is now possible to monitor the changing use
of substrates during exercise and their availability. 23Na+ and 39K+ magnetic resonance
spectroscopy can reveal the changing distribution of sodium and potassium ions across the
muscle cell membrane. This may require infused reagents which shift the resonance fre-
quency of extracellular 23Na+ so that intra- and extracellular distributions can be estimated.
Finally, magnetic resonance imaging can be used to assess whether muscle is active or not,
and it can be adapted to depict accurately movement of fatiguing muscles in real time.
 12

BIOENERGETICS AND MUSCLE CELL TYPES

M. J. Kushmerick

Departments of Radiology, Bioengineering, and Physiology

University of Washington
Seattle, Washington 98195

ABSTRACT

Potential complexities in biochemical and bioenergetic interpretation due to fiber type

heterogeneity are not significant for human muscle. Paradigms for understanding muscle
bioenergetics then can be understood from a set of basic premises of biochemical energy
balance I) ATP provides the energy for all forms of muscle work; 2) chemical energy is stored
in cells as phosphocreatine, a biochemical capacitor; 3) the sum of the coupled ATPases sets
the demand side of the balance and defines energetic states; and 4) this demand is supplied
by aerobic metabolism and the products of the coupled ATPases provide control signals for
regulation of energy balance. We speculate that cytoplasmic signals at work in energy
balance may also control muscle plasticity.

INTRODUCTION

A variety of muscle cell types have been identified in vertebrate animals, including
humans. This chapter assesses the bioenergetic distinctions among those stereotyped muscle
fiber types and whether those distinctions make a significant difference for analyzing muscle
bioenergetics; it then outlines the concept of energy balance for muscle bioenergetics, and
lastly shows how energy balance offers a useful paradigm for understanding muscle function.

FIBER TYPES

Properties of individual muscles cells are classified on the basis of a number of

anatomical, physiological and biochemical criteria into various categories of fibers called
fiber types (Saltin & Gollnick, 1983) 1) fatigue-sensitivity; 2) recruitment order of individual
motor units; 3) metabolic repertoire, namely oxidative and glycolytic; 4) mechanical speed,
namely fast- and slow-twitch types; 5) myofibrillar motor, namely myosin types 1 and 2 (and
further subtypes based on light chain components and variants of the heavy chains); 6) other
protein isoforms, which include a large number troponin subunits, calcium pump in the
sarcoplasmic reticulum, creatine kinase, and enzymes of intermediary metabolism, etc.

175
176 M. J. Kushmerick

One notes immediately that anyone or a group of these characteristics defines only
some aspects of muscle structure and function. Thus they do not necessarily generate the
same categories ofjiber types, and so different classification schemes result depending on
the criteria chosen. Moreover the several types classically described are stereotypes; but in
fact there is a continuum of fibers and their properties (Pette & Staron, 1990). Additional
distinctions can be made on the basis of ion channel subunits, membrane receptors, and many
other cellular and molecular characteristics still being discovered. Thus the number of known
isoforms of proteins in the thick and thin filaments alone is sufficient to produce a very large
number of possible cell types (Pette & Staron, 1990). Nonetheless individual fibers and
whole muscles with predominantly one characteristic can be obtained and their energy
metabolism and other functions can be studied (Crow & Kushmerick, 1982; Metzger &
Moss, 1987; Sweeney et ai., 1988; Bottinelli et ai., 1991; Hofmann et ai., 1991; Kushmerick
et ai., 1992; Larsson & Moss, 1993).
In addition to the categories of characteristics listed above, muscle cells also differ
in the concentrations of the metabolites involved in energy metabolism, for which the phrase
high energy phosphates was introduced (Lipmann, 1941). Measurements of metabolite
composition by 31 P NMR demonstrated a higher phosphocreatine (PCr) and lower inorganic
phosphate (Pi) contents in the fast-twitch muscle compared with a slow-twitch muscle
(Meyer et ai., 1985) confirming the results of biochemical assays of muscle extracts of
fast-twitch and slow-twitch muscles (Meyer et ai., 1980; Crow & Kushmerick, 1982).
Chemical analyses of single fiber segments dissected from resting and stimulated rat
plantaris and soleus muscles (Hintz et ai., 1982) showed differences in metabolite contents
at rest which correlated with the fiber type as well as with the extent ofPCr and ATP splitting
during stimulation. A clear difference was found in the composition of high-energy phos-
phates and myosin heavy chain composition in rodent muscles (Kushmerick et ai., 1993).
Slow type 1 myosin heavy chain correlated with lower PCr and total creatine content and
higher Pi content than was observed in fast type 2 myosin-containing muscle. In anatomically
different muscles of the human, leg small but definite differences are measured in high-en-
ergy phosphate metabolite content. 31 P NMR data indicates a higher Pi and lower per content
in the soleus than in the gastrocnemius by the use of localized spectroscopic techniques to
sample regions totally within each muscle (Vandenborne et ai., 1993a). But the magnitude
of the muscle-specific differences in humans is definitely lower than in the animal experi-
ments. This is due in part to optimization of biological materials to demonstrate big
differences in experimental animal preparations. The 31p NMR data confirm analyses of
single fibers from human muscle (sorted histochemically) which showed significantly small
differences in PCr contents between types 1 and 2 fibers (Edstrom et ai., 1982; SOderlund
& Hultman, 1991; Greenhaff et ai., 1993), and no differences in the ATP content. The
conclusion must be that, in a given human muscle, the range of differences in metabolite
composition is smaller than predicted from the extreme examples from laboratory animals,
and that this narrower range is due to a narrower distribution of cell types.
With respect to muscle properties as a chemo-mechanical machine and to quantities
related to bioenergetics, it is known that there is approximately a 4-fold range in the maximal
mechanical power and ATPase rates and a similar range in metabolic capacities (including
oxidative phosphorylation) over the range of skeletal muscle cells. Moreover, these rates are
graded on the intensity of muscle activation, from rest to maximal, for a more than 10-fold
range. Finally due to the compositional differences in the high-energy phosphate content,
all of these chemical changes with muscle performance begin from different starting values.
Thus we see that for whatever property, there is significant possibility for a distribu-
tion of cell types (and therefore heterogeneity) with respect to functional performance. This
suggests the pattern and kinetics of bioenergetics of muscle activity could be highly complex.
If so, it would be difficult to interpret macroscopic measurements in terms of integrated
Bioenergetics and Muscle Cell Types 177

biochemical and molecular mechanisms as has been attempted in a number of studies in

human and animal limb muscles at rest and during exercise (Gollnick et aI., 1973; Meyer &
Terjung, 1979; Taylor et. aI., 1983; Arnold et aI., 1984; Kushmerick & Meyer, 1985; Chance
et aI., 1986; Ren & Hultman, 1989). Non-invasive measurements by 31pNMR spectroscopy
are perhaps the most powerful because they can be repeated with high time resolution in
human limb muscle exercise and yield information on the concentration of metabolically
active, as opposed to bound, bioenergetic compounds of interest. The ratio PilPCr or Pi/(PCr
+ Pi) is commonly used to infer the bioenergetic status of the muscle in human exercise. In
addition to studies of high-energy phosphate, studies of glucose and glycogen metabolism
using l3C NMR spectroscopy (Bloch et aI., 1994; Price et aI., 1994) similarly make cellular
and molecular conclusions from macroscopic measures. Is it possible that these interpreta-
tions and conclusions are fallacious in principle because of macroscopic sampling of
heterogeneous cell types?
No! There are good biological reasons why the overall bioenergetic functions of a
given muscle are interpretable in biochemical and molecular mechanisms. First, the bioen-
ergetic quantities are self-normalizing as indicated by several authors (Connett, 1988; Meyer,
1989; Conley, 1994). Although the total creatine, PCr and ATP concentrations can differ
many fold in different muscle cells, the ratio of PCr/ATP is relatively constant (Connett,
1988; Conley, 1994). The implication of this constancy is that cells have a similar ADP
concentration and chemical potential of ATP, at least much more alike than expected from
the 2-fold absolute concentration differences. Second, the coupling between mechanical
power output and chemical power input means that the mechanical output dominates the rate
of energy utilization. Because cells with higher ATPase rates have higher PCr and ATP
content, different cells each working, for example at 50% of maximal power, will have
similar relative intracellular changes in ADP and chemical potential. Third, the steady rate
of ATP synthesis (and thus of oxygen consumption), depends on the rate of ATP utilization.
The implication of the two preceding points is that there is a scaling of signals controlling
oxidative metabolism, the predominant source of high-energy phosphates in the cells. In
fact, heterogeneity has been difficult to demonstrate in the bioenergetics of single human
muscles. It has been shown that in strong voluntary exercise of calf muscles a heterogeneity
of pH (as measured by a splitting of the Pi peak in NMR spectra) can be detected
(Vandenborne et aI., 1991), but this was due to sampling of more than one muscle (anatomical
heterogeneity) when localization techniques were not done. With proper localization, indi-
vidual muscles were sampled and muscle-specific differences in pH could be shown
(Vandenbome et aI., 1993b) and each muscle was relatively homogeneous. The rate ofPCr
re-synthesis following three identical activations of the forearm musculature in the same
individual was highly repeatable (approximately ± 10%) and mono-exponential, but the
individuals differed among themselves by almost 200% (Blei et aI., 1993). The implication
is that individuals differ because their component fibers differ. Moreover, the PCr breakdown
rate, also highly repeatable, differed in individuals and the magnitude of the breakdown rate
was inversely correlated with the content of Pi (Blei et aI., 1993). Also individuals can be
grouped on the basis of the kinetics of recovery ofPCr following graded voluntary exercise
(Mizuno et aI., 1994) and show recovery kinetics that were explainable on the basis of the
histochemical fiber types of biopsies from the same region sampled by the NMR experiment.
These results (Blei et aI., 1993; Mizuno et aI., 1994) show that the inter-individual NMR
variation in energy metabolism is related closely to fiber type composition of the specific
muscle. The conclusion from this evidence is that any given human muscle can be best
represented as a unimodal continuum of properties of a relatively small range of cell types.
Therefore macroscopic measurements of cellular bioenergetics can be reliably interpreted
in terms of intracellular and molecular mechanisms with the· full realization of differences
178 M. J. Kushmerick

across muscles of the same individual, and in the same muscle of different individuals. We
now consider some rules that apply to all muscles.

A PARADIGM FOR ENERGY BALANCE IN MUSCLE FUNCTION

The classical approach to describe and understand energy balance was defined by
A.V. Hill by the application of equilibrium thermodynamics (energy conservation) to
contracting muscle. The conceptual basis of myothermal energy balance states that the total
energy change between two states of a muscle is exactly equal to the sum of the heat output
and the work done in that change. The many successes of this approach (Wilkie, 1960;
Kushmerick, 1983; Woledge et aI., 1986) have been displaced by elaboration of so many
individual reactions, including highly exothermic and endothermic ones, that the task of
quantifying their extents of reaction and enthalpies led to unmanageable experimental
difficulties and uncertain correlations of the details ofthe chemical processes with myother-
mal events. An alternative, a biochemical energy balance, was elaborated that organizes the
functional integration of the major metabolic pathways in maintaining a chemically defined
energy ba~ance (Kushmerick, 1977).
Lipmann's work suggested how this task might be done. He used the phrase high-en-
ergy phosphate bond to point out that, with all the intricacy of metabolic pathways, there is
only a small set of common biochemicals involved in energy transducing mechanisms, e.g.
ATP (Lipmann, 1941). Energy transductions in muscle (molecular electro-chemical and
chemo-osmotic machines and motors) synthesize or dissipate ATP, the cell's source of
chemical potential energy, in a way that couples the energy to the metabolic, electrical,
osmotic and mechanical work. The extent of energy dissipation uncoupled to ATP is
negligible. The magnitude of ATP hydrolysis uncoupled to the molecular transducers (in
analogy to the proton leak dissipating chemical energy in the mitochondria without coupling
to ATP synthesis (Brown, 1992)) is unknown. It is estimated to be very small on the basis of
the low rate of metabolism in muscles at rest. Although a small amount of ATP is stored in
the cell (on the order of 5 mM), ATP must be produced by metabolism concurrent with work
performance. The reason is that typical rates of ATP utilization in active muscle (on the order
of 1 mM per s) would otherwise exhaust the muscle's source of chemical potential energy
in several seconds. We know that steady states of metabolic activity readily occur over a
wide range of muscle functions from a basal minimum up to some maximum. Beyond this
range of energy balance, called the buffering range (Connett, 1988), the integrated operation
disintegrates into dysfunction for which we use terms like fatigue, pathology, disease, etc.
Energy balance describes those properties of energy using and generating mechanisms and
their control that sets the supply of chemical energy to match the demands of chemical
transducing machines.
The following statements and working hypotheses provide a relatively simple scheme
by which one can integrate the diverse component mechanisms of bioenergetic systems.
These are tentative rules and hypotheses for all types of muscles and suggest novel
experimental approaches. This set of biochemical rules are not a priori predictable from first
principles of physics because the mechanisms involved are the products of evolution.

ATP Provides the Energy for All Forms of Muscle Work

There is no other form of energy coupling known; muscle is not a heat engine or fuel
cell. The essential feature of muscle as a chemo-mechanical system is that free energy of
chemical reactions is coupled to mechanical performance by the actomyosin motor. Chemi-
cal input power matches the mechanical output power. Actomyosin interactions generating
Bioenergetics and Muscle Cell Types 179

force and doing work (moving a load) and ion pumps doing electrical and osmotic work all
rely on energy available from ATP, thereby coupling the exergonic process of ATP splitting
to the endergonic processes of cellular work.

Chemical Energy Is Stored in Cells - Concept of a Biochemical Capacitor

Certain forms of chemical potential energy are biochemically inter-convertible by
means of near equilibrium reactions. Perhaps the best known example is creatine kinase
(Meyer et aI., 1984; Wallimann et aI., 1992) which catalyses the inter-conversion between
ATPandPCr:

ATP + Creatine ~ PCr + ADP + H+.

This reaction defines a chemical energy capacitance because the content of PCr is
effectively a capacitor of chemical energy by the nature of its near equilibration with ATP
and ADP. Note that the rate of charging and discharging this capacitor may not be the simple
exponential form of a typical electrical RC circuit because the flux is defined by the rate
equations for the enzyme catalyzed reaction. Muscle cells have concentrations ofPCr higher
than ATP. Yet none of the chemical energy transducing molecular machines is known to use
PCr directly for coupling chemical energy to work production; PCr is a substrate for only
one enzyme, creatine kinase. Thus, as a capacitor, PCr represents chemical potential energy
previously synthesized by metabolism. Thus endergonic processes can be temporarily
coupled to free energy dissipation without the requirement for simultaneous oxidative
metabolism provided the limits of the chemical energy capacitance, the supply ofPCr, is not
exceeded. The range of ATP chemical potential over which PCr and creatine kinase is an
effective buffer was distinguished from a depleting range (Connett, 1988) in which there is
depletion of ATP, loss of total adenine nucleotides via AMP deaminase, fatigue and cellular
dysfunction.

The Sum of the Coupled ATPases Sets the Demand Side of the Balance
and Defines Energetic States
The overall metabolic rate of a cell changes because its ATPase rates change; the
converse is also true, viz. that if a measure of the overall metabolic rate increases then the
steady-state sum of the ATPases must have increased. In this sense, the rate of the coupled
free energy dissipation mechanisms is the primary and causal mechanism in bioenergetics.
ATP dissipation (coupled to performing work) is in tum activated by signals external to this
molecular motor; that is, the motor is not controlled by the chemical potential of ATP. In this
view neither the availability of extra oxygen or substrate per se nor an increase in the
magnitude of the chemical potential stored in ATP and PCr concentrations will increase the
metabolic rate. Only a change in the rate of the ATPases does that.

The Coupled ATPases Provide Control Signals for Regulation of Energy

Balance
How is the remarkable constancy of the concentration of cellular metabolites
achieved in the face of varied levels of demand? One possible way is that the signals
which activate the ATP utilizing mechanisms also provide information to the coupled
processes synthesizing ATP. A common external signal could result in energy balance by
acting on the ATP utilizing and synthesizing mechanisms in parallel. This kind ofmecha-
nism is essentially the role envisioned for calcium activation of mitochondrial dehydro-
180 M. J. Kushmerick

genases (McCormack & Denton, 1990; McCormack & Denton, 1993) and possibly the
ATPsynthetase itself (Das & Harris, 1990). Although such feedforward mechanisms exist,
it is still mechanistically necessary that a signal (or signals) derived from the coupled
chemo-mechanical machine acts in feedback regulation of ATP synthesis. Because energy
balance is a cellular process, it is a better teleological strategy to have the regulation self
contained. The primary signal molecules must then be one or more of the chemical
products of ATP-coupled molecular machines: Pi, ADP, H+ or creatine. These internal
signals in turn regulate intracellular ATP synthesis by feedback signals (the concept is
independent of the mechanism for whether control of oxidative phosphorylation is kinetic
or thermodynamic - consideration beyond the scope of this essay). The presence of
feedback control does not negate the existence of feedforward ones. The logical necessity
of feedback mechanisms is thus derived from the fact that coupled ATP dissipating
mechanisms are essential to achieve energy balance if the result of the feedforward signals
were in the slightest way imbalanced.

APPLICATION OF THESE PRINCIPLES TO QUESTIONS OF

HUMAN MUSCLE BIOENERGETICS

Homogeneity versus Heterogeneity Revisited

Reasons given above remove major concerns about our ability to study biochemical
and molecular mechanisms in whole muscle by NMR and other sampling methods. We
now consider how the principles in the preceding section apply to bioenergetically
different muscles, and that insights are thereby provided into the function of muscles
with a range of properties. A mechanically faster muscle has a greater rate of coupled
ATPase than a slower muscle. Similarly muscles differ in their mitochondrial content
such that the maximal rate of substrate oxidation and oxidative phosphorylation of slow
muscle is greater than fast muscle. Energy balance is satisfied even when a muscle
transiently depends on its high-energy phosphate capacitor PCr, instead of new ATP
synthesis (Meyer et ai., 1984; Connett, 1988). Mechanical power output is only possible
if there is sufficient chemical potential energy available. When the magnitude of the
ATPases exceeds the maximum of the ATP synthesis and the PCr capacitor is discharged
- its function fails - we commonly refers to this state as fatigue wherein some or all of
the same feedback signals for ATP synthesis are inhibitory for ATP utilizing mechanisms.
That extreme state is out of the continuous range of regulation we have been considering,
and was called the depleting phase (Connett, 1988). Muscles which have a smaller demand
by their ATPases and a larger capacity for oxidative re synthesis may never achieve that
state under physiological conditions (examples include soleus and cardiac muscle). Oxi-
dative phosphorylation will increase as if feedback error signals were absent for three
reasons 1) the size of the error signal will be small and perhaps within the noise of our
experimental measures; 2) the effective gain in the system is large so that small signals
produce big effects; and 3) the effectiveness of a feedforward mechanism may be greater
in some muscles than in others. For all these reasons the control mechanisms may appear
qualitatively different in cardiac vs. skeletal muscle (Heineman & Balaban, 1990) whereas
the same mechanisms and principles can operate with only quantitative differences.
Depending on which view is taken, one's experimental design may differ. One insight of
the view proposed here is that a common regulatory scheme can be considered in general,
with differences between muscle types being quantitative, not qualitative.
Bioenergetics and Muscle Cell Types 181

The Chemical Potential of A TP Provides the Energy for Muscle Power

Further important questions follow from this statement. Does the power of the
actomyosin motor depend on the magnitude of the chemical potential? It appears not. In
isometric conditions, the force produced per unit ATP utilized was independent of the
concentrations of ATP, PCr, Cr, Pi and pH. That is, there was a constant ATP utilization
proportional to the integral of isometric force independent of the changing content of high
energy phosphate compounds over a range in which most of the PCr was depleted in
amphibian (Kushmerick & Paul, 1976) and mouse soleus muscle (Crow & Kushmerick,
1982). Thus the cost per unit mechanical output was constant although the chemical potential
decreased. Experiments answering similar questions about the quantities of work done per
unit ATP utilization (Woledge et aI., 1986; their Fig. 4.37) and the comparison of chemical
power input per unit of mechanical power output for working muscle (Kushmerick & Davies,
1969) indicate a matching of ATP utilization to mechanical work output, also in constant
proportion although the chemical potential decreased substantially over the range of meas-
urements. There is evidence that the soleus muscle may differ from others (Crow &
Kushmerick, 1982; Meyer et aI., 1985), so these issues merit reinvestigation.

Mechanisms of Energy Balance May Control Muscle Plasticity

The phenotype of adult muscle is highly adaptable and is a function of the historical
pattern and intensity of usage (Pette & Vrbova, 1992). An argument from parsimony extends
the mechanisms for control of ATP synthesis to mechanisms for control of gene expression.
The observation that the slow-twitch glycolytic muscle stereotype has not been observed in
Nature suggests there may be an interaction between quantitative energy balance in bioen-
ergetics and muscle phenotype. Feedforward control of ATP synthesis relegates the control
of muscle phenotype to mechanisms outside of the muscle; many of these external influences
are known (e.g., thyroid hormone) and are certainly important. However, normal muscle
activity has strong effects on isoform expression. This effect of muscle use on phenotype
makes control of gene expression by feedback signals originating in energy balance an
attractive possibility. In some systems, hypoxic states are strong stimuli to altered gene
expression (Goldberg et aI., 1988; Firth et aI., 1994). Alterations of the cellular high-energy
phosphate content by creatine analogs is associated with altered expression of myosin heavy
chains (Moerland et aI., 1989), metabolic enzyme profiles (Shoubridge et aI., 1985) and
mechanical function (Meyer et aI., 1986). These several observations suggest a speculative
scheme in which altered bioenergetic states playa role in the processes of maintaining and
altering protein expression to achieve the various phenotypes of adult muscle. This chapter
ends with the speCUlation that if this kind of control occurs where the bioenergetic state
effects gene expression and muscle phenotype, there would be a simple explanation (despite
what likely are complex molecular mechanisms) for the relatively uniform distribution of
cell types in a given limb muscle. This distribution is highly plastic (within genetic
constraints) depending on muscle use, that is the integrated history of its ATPases.

ACKNOWLEDGMENTS

This work is largely due to the experimental and conceptual efforts of Kevin Conley
and Michael Blei who continue to collaborate on this project. The authors' work is supported
by United States Public Health Service grants AR 41928, AG 10853, and AR 01914.
Attendance of MJ.K. at the 1994 Bigland-Ritchie conference was supported, in part, by the
University of Miami.
182 M. J. Kushmerick

REFERENCES
Arnold DL, Matthews PM & Radda GK (1984). Metabolic recovery after exercise and the assessment of
mitochondrial function in vivo in human skeletal muscle by means of31P NMR. Magnetic Resonance
in Medicine 1, 307-315.
Blei ML, Conley KE, Odderson IR, Esselman PC & Kushmerick MJ (1993). Individual variation in contractile
cost and recovery in human skeletal muscle. Proceeding of the National Academy of Science, USA
90,7396-7400.
Bloch G, Chase JR, Meyer DB, Avison MJ, Shulman GI & Shulman RG (1994). In vivo regulation of rat
muscle glycogen resynthesis after intense exercise. American Journal of Physiology (Endocrine
Metabolism) 266, E85-E91.
Bottinelli R, Schiaffino S & Reggiani C (1991). Force-velocity relations and myosin heavy chain isoform
compositions of skinned fibres from rat skeletal muscle. Journal of Physiology (London) 437,
655-672.
Brown GC (1992). Control of Respiration and ATP synthesis in mammalian mitochondria and cells. Biochemi-
cal Journal 284, I-B.
Chance B, Leigh J, Kent J, McCully K, Nioka S, Clark BJ & Maris JM (1986). Multiple controls of oxidative
metabolism in living tissues as studied by phosphorus magnetic resonance. Proceedings of the
National Academy of Science, USA 83, 9458-9462.
Conley KE (1994). Cellular energetics during exercise. Advances in Veterinary Science and Comparative
Medicine 38A, 1-39.
Connett RJ (1988). Analysis of metabolic control: new insights using scaled creatine kinase model. American
Journal of Physiology 254, R949-R959.
Crow MT & Kushmerick MJ (1982). Chemical energetics of slow- and fast-twitch muscles of the mouse.
Journal of General PhYSiology 79, 147-166.
Das AM & Harris DA (1990). Regulation of the mitochondrial ATP synthase in intact rat cardiomyocytes.
Biochemical Journal 266, 355-361.
Edstrom L, Hultman E, Sahlin K & Sjoholm H (1982). The contents of high-energy phosphates in different
fibre types in skeletal muscles from rat, guinea-pig, and man. Journal of Physiology (London) 332,
47-58.
Firth JD, Ebert BL, Pugh CW & Ratcliffe PJ (1994). Oxygen-regulated control elements in the phosphoglycer-
ate kinase 1 and lactate dehydrogenase A genes: Similarities with the erythropoietin 3' enhancer.
Proceedings of the National Academy of Science, USA 91, 6496-6500.
Goldberg MA, Dunning SP & Bunn HF (1988). Regulation of the erythropoietin gene: evidence that the oxygen
sensor is a heme protein. Science 242,1412-1414.
Gollnick PD, Armstrong RB, Sembrowich WL, Shepherd RE & Saltin B (1973). Glycogen depletion pattern
in human skeletal muscle fibers after heavy exercise. Journal ofApplied Physiology 34,615-618.
GreenhaffPL, SOderlund K, Ren JM & Hultman E (1993). Energy metabolism in single human muscle fibres
during intermittent contraction with occluded circulation. Journal of Physiology (London) 460,
443-453.
Heineman FW & Balaban RS (1990). Control of mitochondrial respiration in the heart in vivo. Annual Review
ofPhysiology 52, 523-542.
Hintz, CS, Chi M M-Y, Fell RD, Ivy JL, Kaiser KK, Lowry CV & Lowry OH (1982). Metabolite changes in
individual rat muscle fibers during stimulation. American Journal of Physiology 242, C218-C228.
Hofmann PA, Hartzell HC & Moss RL (1991). Alterations in Ca2+ sensitive tension due to partial extraction
of C-protein from rat skinned cardiac myocytes and rabbit skeletal muscle fibers. Journal of General
Physiology 97,1141-1163.
Kushmerick MJ (1977). Energy balance in muscle contraction: A biochemical approach. In: Sanadi R (ed.),
Current Topics in Bioenergetics, vol. 6, pp. 1-15. New York: Academic Press.
Kushmerick MJ (1983). Energetics of muscle contraction. In: Peachey L, Adrian R, Geiger SR (eds.),
Handbook of Physiology, Skeletal Muscle, pp. 189-236. Bethesda: American Physiological Society.
Kushmerick MJ & Davies RE (1969). The chemistry, efficiency and power of maximally working sartorius
muscles. Proceedings of the Royal Society, London B, 315-353.
Kushmerick MJ & Meyer RA (1985). Chemical changes in rat leg muscle by phosphorus nuclear magnetic
resonance. American Journal of PhYSiology 248, C542-C549.
Kushmerick MJ, Moerland TS & Wiseman RW (1992). Mammalian skeletal muscle fibers distinguished by
contents of phosphocreatine, ATP, and Pi. Proceedings of the National Academy of Science, USA 89,
7521-7525.
Bioenergetics and Muscle Cell Types 183

Kushmerick MJ, Moerland TS & Wiseman RW (1993). Two classes of mammalian skeletal muscle fibers
distinguished by metabolite content. Advances in Experimental Medicine and Biology 332, 749-761.
Kushmerick MJ & Paul RJ (1976). Aerobic recovery metabolism following a single isometric tetanus in frog
sartorius muscle at O°C. Journal ofPhysiology (London) 254, 693-709.
Larsson L & Moss RL (1993). Maximum velocity of shortening in relation to myosin isoform composition in
single fibres from human skeletal muscles. Journal ofPhysiology (London) 472, 595-614.
Lipmann F (1941). Metabolic generation and utilization of phosphate bond energy. In: Nord FF, Werkrnan CH
(eds.), Advances in Enzymology, vol 1, pp. 99-162. New York: Intersciences Publishers, Inc.
McCormack JG & Denton RM (1990). Ca2+ as a second messenger within mitochondria in the heart and other
tissues. Annual Review ofPhysiology 52, 451-466.
McCormack JG & Denton RM (1993). The role of intramitochondrial Ca2+ in the regulation of oxidative
phosphorylation in mammalian tissues. Biochemical Society Transactions 21, 793-799.
Metzger JM & Moss RL (1987). Shortening velocity in skinned single muscle fibers. Biophysical Journal 52,
127-8131.
Meyer RA (1989). Linear dependence of muscle phosphocreatine kinetics on total creatine content. American
Journal of Physiology 257, C 1149-C 1157.
Meyer RA, Brown TR, Krilowicz BL & Kushmerick MJ (1986). Phosphagen and intracellular pH changes
during contraction of creatine-depleted rat muscle. American Journal ofPhysiology 250,C264-C274.
Meyer RA, Brown TR, Kushmerick MJ (1985). Phosphorus nuclear magnetic resonance of fast- and slow-
twitch muscle. American Journal of Physiology 248, C279-C287.
Meyer RA, Dudley GA & Terjung RL (1980). Ammonia and IMP in different skeletal muscle fibers after
exercise in rats. Journal ofApplied Physiology 49, 1037-1041.
Meyer RA, Sweeney HL & Kushmerick MJ (1984). A simple analysis of the "phosphocreatine shuttle."
American Journal ofPhysiology 246, C365-C377.
Meyer RA & Terjung RL (1979). Differences in ammonia and adenylate metabolism in contracting fast and
slow muscle. American Journal of Physiology 237, C lll-C 118.
Mizuno M, Secher NH & QuistorffB (1994). 31P-NMR spectroscopy, rsEMG, and histochemical fiber types
of human wrist flexor muscles. Journal ofApplied Physiology 76, 531-538.
Moerland TS, WolfNG & Kushmerick MJ (1989). Administration ofa creatine analogue induces isomyosin
transitions in muscle. American Journal ofPhysiology 257, C81 0-C816.
Pette D & Staron RS (1990). Cellular and molecular diversity of mammalian skeletal muscle fibers. Reviews
ofPhysiology, Biochemistry & Pharmacology 116,1-76.
Pette D & Vrbova G (1992). Adaptation of mammalian skeletal muscle fibers to chronic electrical stimulation.
Reviews of Physiology, Biochemistry & Pharmacology 120, 115-202.
Price TB, Rothman DL, Taylor R, Avison MJ, Shulman GI & Shulman RG (1994). Human muscle glycogen
resynthesis after exercise: Insulin-dependent and -independent phases. Journal ofApplied Physiology
76, 104-111.
Ren J-M & Hultman E (1989). Regulation of glycogenolysis in human skeletal muscle. Journal of Applied
Physiology 67, 2243-2248.
Saltin B & Gollnick PD (1983). Skeletal muscle adaptability: significance for metabolism and performance.
In: Geiger SR, Adrian RH (eds.), Peachey LD (sec. ed.), Handbook of Physiology, sec. 10, Skeletal
Muscle, pp. 555-631. Bethesda, MD: American Physiological Society.
Shoubridge EA, Challiss RAJ, Hayes DJ & Radda GK (1985). Biochemical adaptation in the skeletal muscle
of rats depleted of creatine with the substrate analogue B-guanidinopropionic acid. Biochemical
Journal 232, 125-131.
SOderlund K & Hultman E (1991). ATP and phosphocreatine changes in single human muscle fibers after
intense electrical stimulation. American Journal of Physiology (Endocrine Metabolism) 261, E737-
E741.
Sweeney HL, Kushmerick MI, Mabuchi K, Sreter FA & Gergely J (1988). Myosin alkali light chain and heavy
chain variations correlate with altered shortening velocity of isolated skeletal muscle. Journal of
Biological Chemistry 263, 9034-9039.
Taylor DJ, Bore PI, Styles P, Gadian DG & Radda GK (1983). Bioenergetics of intact human muscle. A 31P
nuclear magnetic resonance study. Molecular Biology & Medicine 1, 77-94.
Vandenbome K, McCully K, Kakihira H, Prammer M, Bolinger L, Detre JA, De Meirleir K, Walter G, Chance
B & Leigh JS (1991). Metabolic heterogeneity in human calf muscle during maximal exercise.
Proceedings of the National Academy of Science, USA 88, 5714-5718.
Vandenbome K, Walter G, Goelman G, Ploutz L, Dudley G & Leigh JS (1993a). Phosphate content of fast and
slow twitch muscles. Proceedings of the Society ofMagnetic Resonance in Medicine 3, 1140.
184 M. J. Kushmerick

Vandenborne K, Walter G, Leigh JS & Goelman G (1993b). pH heterogeneity during exercise in localized
spectra from single human muscles. American Journal ofPhysiology (Cell Physiology) 265, C 1332-
C1339.
Wallimann T, Wyss M, Brdiczka D, Nicolay K & Eppenberger HM (1992). Intracellular compartmentation,
structure and function of creatine kinase isoenzymes in tissues with high and fluctuating energy
demands: the 'phosphocreatine circuit' for cellular energy homeostasis. Biochemical Journal 281,
21-40.
Wilkie DR (1960). Thermodynamics and the interpretation of biological heat measurements. Progress in
Biophysics and Biophysical Chemistry 10, 260-298.
Woledge RC, Curtin NA & Homsher E (eds.) (1986). Energetic Aspects of Muscle Contraction. New York:
Academic Press.
 13

METABOLIC CORRELATES OF FATIGUE

FROM DIFFERENT TYPES OF EXERCISE IN
MAN

N. K. V011estad

Department of Physiology
National Institute of Occupational Health
Box 8149 Dep, N-0033 Oslo, Norway

ABSTRACT

It is well established that muscle fatigue, defined as a decline in maximal force generating
capacity, is a common response to muscular activity. To what extent metabolic factors
contribute to the reduced muscle function is still debated. Metabolic effects can affect muscle
through different processes, either through a reduced ATP supply or by effects on EC-cou-
piing or crossbridge dynamics. Observations from in vitro experiments are often extrapolated
to interpret fatigue mechanisms from measurements performed in vivo, without recognizing
that the biochemical reactions involved can be quite different depending upon such factors
as activation pattern, mode and duration of exercise. During repeated submaximal contrac-
tions, there is a negligible accumulation of H+ and inorganic phosphate, and hence fatigue
must be ascribed to other factors. Substrate depletion might contribute to exhaustion, but
cannot explain the gradual loss of maximal force. Curiously, the energetic cost of contraction
increases progressively during repeated isometric but not during concentric contractions.
With contractions involving high-force or high power output, fatigue is better related to
H2P04- than to pH, but still other factors seem to playa role.

INTRODUCTION

This chapter describes the metabolic responses during different types of exercise and
relates them to fatigue. Metabolic and biochemical aspects offatigue involve energy release
and utilization, and the consequences of substrate degradation and electrolyte shifts. While
all of these factors may be important in fatigue, only the first two factors will be examined
in this chapter. The effects of pH and electrolyte balance are covered elsewhere (Allen et aI.,
Chapter 3; Sj0gaard & McComas, Chapter 4).
To generate force and power, ATP is hydrolyzed by myosin ATPase. In addition, ATP
is utilized in the re-establishment of electrolyte homeostasis after ion fluxes associated with
propagation of the action potential and Ca2+ release from sarcoplasmic reticulum. Adequate
resynthesis of ATP, therefore, is important in the maintenance of cellular function. During

185
186 N. K. Vellestad

prolonged exercise, oxidative phosphorylation of carbohydrates and free fatty acids are the
main sources for energy. Although the products of this process are not known to affect force,
the availability of substrates or the deterioration of mitochondrial function may contribute
to muscle fatigue. With an inadequate oxygen supply, such as during sustained high-force
contractions, glycogen and phosphocreatine (PCr) are rapidly broken down, and lactic acid
and inorganic phosphate (Pi) accumulate. The possible role of these metabolites and substrate
supply in the development of fatigue is still debated, and I will review some of the available
data with emphasis upon how they relate to fatigue from voluntary activation in humans.

FATIGUE AND EXHAUSTION

The term muscle fatigue is used to denote a decline in the maximal contractile force
of the muscle. Defined this way, fatigue may be assessed from brief maximal voluntary
contractions or tetanically stimulated contractions. Most importantly, fatigue may occur
during exercise at submaximallevels without being reflected in the performance. Fatigue is
different to exhaustion, where the latter is defined as the moment in time when the expected
force level cannot be maintained.
Muscle fatigue can occur with all types of muscular activity (i.e., isometric, shorten-
ing and lengthening contractions), and with both short-lasting intense and prolonged exer-
cise. The various modes and intensities of exercise involve different metabolic processes,
and the importance of biochemical changes to the decline in force may vary accordingly. To
illustrate these possibilities, this chapter focuses on metabolic responses to repeated sub-
maximal contractions and to exercise involving high-intensity force or power. These exam-
ples indicate that it is difficult to extrapolate the biochemical changes seen under one
condition to other exercise protocols.

ANALYTICAL TECHNIQUES

Several different analytical methods have been used to study the effects ofbiochemi-
cal factors on performance. While some early data were obtained by manipulating substrate
levels or metabolic pathways (Lundsgaard, 1930), these approaches were limited and initial
biochemical analysis could only be carried out on stimulated animal muscles. With the
introduction of needle biopsies in the sixties (Bergstrom, 1962), a more detailed picture of
the metabolic changes in human muscle during various types of exercise was obtained.
Substrate degradation and accumulation of metabolic products could readily be determined.
The biopsy technique, however, has several limitations. First, only a limited number of
samples can be obtained from each muscle. Second, the time resolution is poor. And thirdly,
there is evidence that biochemical processes occur in the muscle when the sample is cut
(SOderlund & Hultman, 1986). Hence, interpretation of small metabolic changes measured
with the biopsy technique requires caution.
These disadvantages are not present when metabolites are assessed using 31 P nuclear
e
magnetic resonance spectroscopy 1p-MRS). Relative changes in PCr, Pi and ATP can be
determined from the spectra, and from these ADP, H2P04' and pH can be calculated (see
Bertocci, Chapter 15). This nonivasive technique is based on phosphorous spectra recorded
from muscles by surface coils and usable spectra may be obtained in a few seconds (Degroot
et aI., 1993). The greatest disadvantage of this method is that it has been applied almost
exclusively to the study of phosphorous compounds in relation to fatigue. In the future,
further developments in MRS may expand to analyze other aspects of metabolism. Another
disadvantage of MRS is that it restricts the experimental design to the performance of a task
Metabolic Correlates of Fatigue from Different Types of Exercise in Man 187

within a strong magnetic field. Hence, the exercise-induced changes in metabolites have
mainly been studied during isometric contractions, or dynamic contractions of smaller
muscle groups. A third disadvantage of the 31P_MRS is that it is not possible to calibrate the
quantification of metabolites determined from the recorded spectra. To compensate for this,
high-energy phosphates are most often expressed in relation to total phosphate. A fourth
disadvantage is that the observed changes is concentrations will always be an average of the
muscle fibers within the field of view. Details about cellular changes cannot be obtained
when only a small portion of the muscle fibers are active.
In a recent study, Bangsbo and coworkers (1993) compared changes in metabolites
during fatigue in parallel studies using 31p_MRS and biochemical analyses of muscle
biopsies. Similar changes in PCr were observed with the two methods, but a somewhat larger
decline in ATP was determined biochemically after high-force contractions. The increase in
ADP was much more pronounced when determined by 31P-MRS than biochemically. This
discrepancy was caused by the fact that the biochemical analysis measures total ADP whereas
free ADP is determined from 31P-MRS spectra. Hence, changes in PCr and for most purposes
ATP can be reliably assessed with either method. However, only 31p-MRS is sensitive for
changes in free ADP.
The knowledge gained over the last three decades using different experimental
models and analytical techniques has shown that the biochemical changes associated with
fatigue vary over a wide range. Most investigators concur that several different mechanisms
may playa role depending on the type and intensity of exercise and the fiber-type composi-
tion of the involved muscles.

PROLONGED EXERCISE - REPETITIVE SHORTENING OR

ISOMETRIC CONTRACTIONS

During prolonged repeated submaximal contractions, the maximal force generating

capacity declines continually (Bigland-Ritchie et aI., 1986b; V0llestad et aI., 1988). At the
onset of submaximal exercise only a fraction of the muscle fibers in the working muscles
are activated. This has been shown by biopsy studies from several laboratories (Gollnick et
aI., 1973; Gollnick et aI., 1974; V0llestad et aI., 1984; V011estad & Blom, 1985; Bigland-
Ritchie et aI., 1986a). When performing histochemical staining for the glycogen concentra-
tion in the fibers, it has been shown that only type I fibers are recruited from onset of bicycle
exercise at about 40% of the maximal oxygen uptake (V0 2max) (V011estad & Blom, 1985).
With increasing intensity, a larger fraction of the muscle fibers are recruited at the beginning.
Interestingly, we observed a linear rise in the number of active muscle fibers as the exercise
intensity increased (V0llestad & Blom, 1985). If submaximal exercise is prolonged for an
hour or more, exhaustion of the glycogen stores may occur in some fibers (Gollnick et aI.,
1973; V0llestad et aI., 1984). Due to the lack of glycogen, these fibers will therefore have a
decreased capacity to release ATP. As exercise approaches exhaustion, an increasing number
of fibers show total glycogen depletion.
One appealing hypothesis is thus that exhaustion is caused by an insufficient rate of
ATP release in the muscle fibers. Several lines of evidence, however, argue against such a
causative relationship. Maximal force generating capacity is determined by brief MVCs or
tetanically stimulated contractions. The ATP hydrolysis in these contractions may amount to
1-2 mmollkg ww (Edwards, 1976; Katz et aI., 1986), indicating that only a minor fraction
of the PCr and ATP stores are utilized under the test contractions. In keeping with this, several
reports based on 31p-MRS show that only small, if any, changes occur in ATP during short
lasting high-force contractions (Miller et aI., 1988; Quistorff et aI., 1992). It might be argued
188 N. K. Vollestad

that the fall in ATP concentration is much larger close to the crossbridges than in other parts
of the cell. Hence, the ATP concentration determined as an average for the cell or muscle
does not reflect the true conditions at the force-generating sites. However, as recently
reviewed by Fitts (1994), it is unlikely that even in localized parts of the cell the ATP
concentration should drop to levels below 50 ~M which is shown to be sufficient for maximal
isometric tension in rabbit psoas fibers (Cooke & Bialek, 1979). It may thus be concluded
that ATP availability is not the main cause of fatigue under these conditions.
It is well established that lowered pH or elevated concentrations of inorganic
phosphate (Pi) reduces the force generated by the crossbridges (Cooke & Pate, 1985;
Donaldson & Hermansen, 1978). The effects of these factors during prolonged submaximal
exercise are probably marginal because only a small fraction of the energy is released by
anaerobic pathways. Hence, the accumulation of lactic acid or Pi is less than that required
to affect force generation significantly (V ollestad et aI., 1988).
More recently, different research groups have studied fatigue using repeated sub-
maximal isometric contractions. In several series of experiments, we have examined changes
during quadriceps contractions held at 30% MVC for 6 s with 4 s rest between. The MVC
force fell gradually during the entire exercise period reaching 40-60% of control at exhaus-
tion, whereas serial muscle biopsies revealed only marginal changes in muscle glycogen,
PCr, ATP and lactic acid for the first 30 min (V ollestad et aI., 1988). At exhaustion, however,
almost total depletion of the PCr store was observed, while glycogen was reduced by about
30%. These observations indicate that also with repetitive isometric contractions the gradual
decline in maximal force generating capacity during repeated shortening or isometric
contractions must be explained by other factors than those connected to substrate depletion
or accumulation of Pi and H+.
Further evidence for the lack of a uniform relationship between fatigue and metabo-
lite changes have been provided by experiments using 31p-MRS. Fig. 1 shows results
obtained from two subjects during repeated isometric contractions at 40% MVC with the
quadriceps muscle. In one subject (dotted line), PCr fell by almost 90% within 10 min from
the start of exercise, which was continued for more than 20 min with depleted PCr stores.
Even though he was totally depleted of PCr, only marginal changes were seen in pH. In the
other subject (solid line), PCr fell continually until exhaustion, and no changes were seen in
pH. Fatigue developed gradually and by an equal amount for the two subjects. In addition
to emphasizing that fatigue can develop without large changes in PCr, Pi and pH, these results
show that constant force contractions can be carried out for a long time almost in the absence
ofPCr. This is at variance with our earlier data from repeated isometric contractions, in which
exhaustion seemed to be clearly related to PCr depletion (V ollestad et aI., 1988). An example
of the metabolic changes in one of these subjects is given in Fig. 2, illustrating that PCr was
maintained at almost resting levels for 101 min, but totally depleted at exhaustion which was
reached after 104 min. Also ATP fell rapidly by 2.3 mmol/kg ww in the last 3 min of exercise.
The discrepancy between the data obtained in the two series of experiments is not understood,
but could be related to differences in the angle of the knee (and thus muscle length) or to
circulatory effects due to differences in body position (supine vs. sitting).
Interestingly, there is one distinct difference in the metabolic response to cycling or
running compared with repeated isometric contractions. When this type of dynamic exercise
is kept at a constant intensity, the oxygen consumption (V0 2) rises initially, and reaches a
new steady level within a minute or two. Hence, the energy cost of contraction appears to
remain constant with time. In contrast to this, repeated submaximal isometric contractions
at 30% MVC with the quadriceps muscles results in an initial rise in VOz followed by a
continual further rise until exhaustion (V ollestad et aI., 1990). Similar results have been
reported by Sahlin and coworkers (1992), who also showed that the oxygen cost of isometric
contractions remained elevated for an hour after end of exercise.
Metabolic Correlates of Fatigue from Different Types of Exercise in Man 189

7,4
7.2
7.0
:r::a. 6.8
6.6
6,4
6.2
120

100

Figure 1. Changes in pH and PCr deter-

mined in the quadriceps muscle every 9 s
by 31p-MRS during intermittent isomet-
ric contractions at 40% MVC. The exer-
cise was carried out with the subjects in
a supine position and with each contrac-
tion held for 6 s with 4 s rest between. 20
Data for two subjects shown by dotted
and solid lines. Some of the deflections
are due to maximal test contractions car-
ried out regularly. o 10 20 30 40 50
Time (min)
The causes for the gradual rise in V0 2 during repeated isometric contractions are not
clarified. There are two alternative explanations. Either the ATP demand (ATP turnover:force
ratio) increases, or the mitochondrial ATP production becomes less efficient (decreased p:o
ratio). The latter mechanism would predict that oxygen consumption during any type of
exercise involving the fatigued muscles would be elevated. Sahlin and coworkers (1992)
tested this by comparing the V0 2 during bicycle exercise before and after fatigue from
repeated isometric contractions. They observed no change in V0 2 during bicycling, indicat-
ing that the mitochondrial respiration was not altered. The most probable explanation for the
increased V0 2 thus appears to be that the ATP turnover:force ratio increases. Two lines of

20

1 Cl
15
per

.!!:
(5

.sE 10
d
Figure 2. Changes in muscle concentrations of c:
PCr, ATP and lactate (La) determined from bi- 8 5
opsies taken from the quadriceps muscle of one
subject. Intermittent isometric knee extension
contractions at 30% MVC were carried out until o
exhaustion (104 min). The subject was seated in
a chair and each contraction lasted 6 s with 4 s o 20 40 60 80 100 120
rest between. Time (min)
190 N. K. Vollestad

175 0 30% MVC

~
0
'-
c:
• 50% MVC

0
u
150
~
'-'
Q)
CJ)
'-
Q)
125
'-
:J
~

0
'-
Q)
D- 100
E
Q)
f-
{
-10 0 10 20 30 40 50
Time (min)
Figure 3. Temperature rise in the vastus lateralis muscle during isometric test contractions performed every
4th min of an intermittent isometric knee extension protocol (6 s contraction and 4 s rest) held at 30% and
50% MVC for 6 s with 4 s rest between. The parallel increase in the rate of temperature rise during the test
contractions at the two force levels (with more type II fibers active at 50% MVC) indicate that the recruitment
pattern plays a minor role for the increased energy utilization (reproduced with permission from Sejersted &
V011estad, 1993).

evidence from recent experiments in our laboratory support this conclusion. In a separate
set of experiments, we examined the rate of heat accumulation during constant force test
contractions for 10-15 s carried out at regular intervals during 60 min of repeated 6 s isometric
contractions. As shown in Fig. 3, the rate of temperature rise increased gradually as fatigue
developed. Because the O2 stored in the muscle is insufficient to support 10-15 s contractions
at 30-50% MVC, heat accumulation after the first few seconds probably reflects the rate of
anaerobic energy release and thus should be unrelated to mitochondrial respiration. In
another series of experiments, we examined the changes in relaxation rate, because this
parameter seems to be closely related to the ATP turnover (Edwards et aI., 1975). Despite
profound fatigue (MVC fell by 30-40%), the rate of relaxation rose gradually (V011estad et
aI., 1995). During repeated 30% MVC contractions carried out till exhaustion, the half-re-
laxation time of twitches and tetanic contractions decreased by 27 %. The exercise-induced
increase in rates of heat accumulation and relaxation are in keeping with an increased ATP
utilization. It must be emphasized, however, that there is no direct evidence for this
conclusion. More research is needed to identify the mechanisms behind the increased
metabolic rate during repeated isometric contractions. In addition, it should be sought out
why the metabolic response to repeated concentric and isometric contractions differs.

HIGH-INTENSITY DYNAMIC EXERCISE AND HIGH-FORCE

ISOMETRIC CONTRACTIONS

In contrast to prolonged submaximal exercise, exhaustive short lasting exercise

requires extensive activation of most fibers in the working muscles, and fatigue develops
rapidly. During sustained maximal voluntary contractions, the MVC force will decline by
50% within 1-3 min (Miller et aI., 1988; Wilson et aI., 1988; Degroot et aI., 1993). Using an
isokinetic ergometer or a nonmotorized treadmill, it has been shown that maximal power
Metabolic Correlates of Fatigue from Different Types of Exercise in Man 191

falls by about 50% during 30-45 s of concentric contractions (Thorstensson & Karlsson,
1976; McCartney et aI., 1983; Cheetham et aI., 1986).
Under these conditions, the ATP released from aerobic processes is insufficient to match
the energy demand. Hence, force and power output must rely heavily on ATP from glycolysis
and PCr degradation. Accordingly, a number of studies have shown that lactic acid increases
by 20-30 mmollkg ww and PCr is reduced by more than 80% (Hermansen & Vaage, 1977; Katz
et aI., 1986; Cheetham et aI., 1986; Miller et aI., 1988; Wilson et aI., 1988). Inorganic phosphate
increases as a mirror image ofPCr, whereas only small and insignificant changes are reported
for ATP (Miller et al., 1988; Quistorff et aI., 1992). Because MRS has a relatively good time
resolution, it has been possible to compare the temporal changes in maximal force generating
capacity and metabolite concentration during sustained contractions. Several recent studies
have shown that the fall in PCr and rise in Pi parallels the reduction in maximal force over the
first 1-2 min. Thereafter, the levels ofPCr and Pi almost stabilizes while force continues to
decline (Wilson et aI., 1988; Degroot et aI., 1993). In contrast, pH increases initially, before a
steady decline begins (Miller et aI., 1988; Wilson et aI., 1988; Degroot et aI., 1993). An
illustration of these changes are shown in Fig. 4.
Studies of skinned fibers have shown that both low pH and increased Pi may reduce
maximal force by 50% (Donaldson & Hermansen, 1978; Cooke & Pate, 1985). Accordingly,
both of these factors have been assumed to be responsible for fatigue. Several lines of
evidence, however, indicate that a low pH is not necessarily an important mechanism of
muscle fatigue. By comparing the relationship between muscle force and metabolite levels
during maximal concentric contractions of different speeds, Wilson and coworkers (1988)
found that there were no uniform relation between pH and force, whereas the relationship
between H2P0 4- and force remained unaffected by the duty cycle. Hence, they concludl!d
that fatigue was attributed more to high levels ofH2P04- and less to a low pH. Other studies
with different contraction protocols have arrived at the same conclusion (Miller et aI., 1988;
Le Rumeur et aI., 1990; Degroot et aI., 1993). Furthermore, the recovery of muscle force is

MVC Recovery

~ 100
, ,----------
,,
~
0a..
50 ,,
----!,
0
40
--'I,
,,
~
a , ,,
r ,
~
-- -- --
0
100
. .. . . ' .. . .. .
,,
~

~
~
,
u ,,
~ I
50
Figure 4. Changes in concentrations of PCr and
H+ of calf muscle during and after 4 min of sus- 0 240 1200
tained MVC (adapted from Degroot et aI., 1993). Time (s)
192 N. K. VoUestad

more closely related to H2P0 4- than to pH because MVC force increases immediately after
the end of exercise, while pH remains depressed or continues to decline for a minute or two
before slowly returning to pre-exercise levels (Miller et aI., 1988; Degroot et aI., 1993).
Based on the differences in glycolytic capacity between type I (slow twitch) and II
(fast twitch) fibers, it is often assumed that anaerobic energy release occurs predominantly
in type II fibers. Accordingly, one would expect pH and Pi to influence the force generating
capacity in type I less than in type II. From muscle biopsies obtained before and after short
lasting bicycle exercise of various intensities we calculated the glycogenolytic rate of muscle
fibers of various types (V0llestad et aI., 1992). At three different intensities (120-200%
V0 2max), the glycogen breakdown in type I fibers was about 30% lower than in type II
fibers. During exercise at the highest intensity (200% V0 2max), the glycogenolytic rate in
type I fibers was twice the rate observed in type II fibers during exercise at 120% V0 2max.
This observation suggests that in human muscle marked reductions in pH may occur in both
fiber types, in keeping with single fiber analysis of lactate content (Essen & Hiiggmark,
1975) Because PCr is reported to decline by 80% or more in the quadriceps muscle during
intense exercise, PCr depletion must be substantial in type I fibers. This is shown for
electrically stimulated contractions (SOderlund et aI., 1992). Hence, it appears that fatigue
in both fiber types can potentially be due to low pH or PCr depletion. In contrast to the fast
glycogenolytic rate that may be observed in type I fibers during voluntary activation,
electrically stimulated contractions of the quadriceps muscle fails to stimulate glyco-
genolysis markedly in type I fibers (SOderlund et aI., 1992). The causes for the different
metabolic response to voluntary and electrically stimulated contractions are unresolved.
However, these data emphasize that metabolic regulation may be task dependent.

CONCLUSION

The available data shows that fatigue related changes vary markedly between
different tasks. Furthermore, correlations between fatigue and known metabolic changes all
indicate that the metabolic factors are insufficient to fully explain the changes in maximal
force. There is now substantial evidence that other factors also contribute to fatigue. In
particular, during high-force contractions or exercise at high power output, electrolyte shifts
over the sarcolemma and between sarcoplasmic reticulum and cytosol can be dramatically
altered (Sejersted, 1992). Future studies should examine the interaction between electrolyte
balance and metabolic changes.

ACKNOWLEDGMENTS

The author wishes to thank Eirik Saugen, Dr. Henry Gibson and Prof. Richard H. T.
Edwards for stimulating discussions. Her work is supported by the Research Council of
Norway. Her attendance at the 1994 Bigland-Ritchie conference was supported, in part, by
the American Physical Therapy Association (Research and Analysis Division), and the
Muscular Dystrophy Association (USA).

REFERENCES
Bangsbo J, Johansen L, QuistorffB & Saltin B (1993). NMR and analytic biochemical evaluation ofCrP and
nucleotides in the human calf during muscle contraction. Journal of Applied Physiology 74, 2034-
2039.
Metabolic Correlates of Fatigue from Different Types of Exercise in Man 193

Bergstrom J (1962). Muscle electrolytes in man. Scandinavian Journal of Clinical and Laboratory Investiga-
tion Supplement (Oslo) 14, 9-88.
Bigland-Ritchie B, Cafarelli E & V011estad NK (1986a). Fatigue of submaximal static contractions. Acta
Physiologica Scandinavica 128, 137-148.
Bigland-Ritchie B, Furbush F & Woods JJ (1986b). Fatigue ofintermittent submaximal voluntary contractions:
central and peripheral factors. Journal ofApplied Physiology 61,421-429.
Cheetham ME, Boobis LH, Brooks S & Williams C (1986). Human muscle metabolism during sprint running.
Journal ofApplied Physiology 61, 54-60.
Cooke R & Bialek W (1979). Contraction of glycerinated muscle fibers as a function of the ATP concentration.
Biophysical Journal 28, 241-258.
Cooke R & Pate E (1985). The effects of ADP and phosphate on the contraction of muscle fibers. Biophysical
Journal 48, 789-798.
Degroot M, Massie B, Boska M, Gober J, Miller RG & Weiner MW (1993). Dissociation of [H+] from fatigue
in human muscle detected by high time resolution 31p_NMR. Muscle &Nerve 16, 91-98.
Donaldson SKB & Hermansen L (1978). Differential, direct effects ofH+ on Ca2+-activated force of skinned
fibers from the soleus, cardiac and adductor magnus muscles of rabbits. Pflugers Archiv 376,55-65.
Edwards RHT (1976). Metabolic changes during isometric contractions of the quadriceps muscle. In: Jokl E
(ed.), Medicine and Sport, vol. 9, pp. 114-131. Basel: Karger.
Edwards RHT, Hill DH & Jones DA(1975). Metabolic changes associated with the slowing of relaxation in
fatigued mouse muscle. Journal ofPhysiology (London) 251, 287-301.
Essen B & Hiiggmark T (1975). Lactate concentration in type I and II muscle fibres during muscle contraction
in man. Acta Physiologica Scandinavica 95, 344-346.
Fitts RH (1994). Cellular mechanisms of muscle fatigue. Physiological Reviews 74, 49-94.
Gollnick PD, Armstrong RB, Saubert CW, IV, Sembrowich WL, Sherpherd RE & Saltin B (1973). Glycogen
depletion patterns in human skeletal muscle fibers during prolonged work. Pflugers Archiv 344, 1-12.
Gollnick PD, Piehl K & Saltin B (1974). Selective glycogen depletion pattern in human muscle fibers after
exercise of varying intensity and at varying pedalling rates. Journal of Physiology (London) 241,
45-57.
Hermansen L & Vaage a (1977). Lactate disappearance and glycogen synthesis in human muscle after maximal
exercise. American Journal of Physiology 233, E422-E429.
Katz A, Sahlin K & Henriksson J (1986). Muscle ATP turnover rate during isometric contraction in humans.
Journal ofApplied Physiology 60, 1839-1842.
Le Rumeur E, Le Moyec L, Toulouse P, Le Bars R & de Certaines JD (1990). Muscle fatigue unrelated to
phosphocreatine and pH: an "in vivo" 31-P NMR spectroscopy study. Muscle & Nerve 13, 438-444.
Lundsgaard E (1930). Untersuchungen tiber muskelkontraktionen ohne Mi1chsiiurebildung. Biochemische
Zeitshrift217,162-175.
McCartney N, Heigenhauser GJF, Sargeant AJ & Jones NL (1983). A constant-velocity cycle ergometer for
the study of dynamic muscle function. Journal ofApplied Physiology 55, 212-217.
Miller RG, Boska MD, Moussavi RS, Carson PJ & Weiner MW (1988). 31P nuclear magnetic resonance studies
of high energy phosphates and pH in human muscle fatigue. Journal of Clinincal Investigation 81,
1190-1196.
QuistorffB, Johansen L & Sahlin K (1992). Absence of phosphocreatine resynthesis in human calf muscle
during ischaemic recovery. Biochemical Journal 291, 681-686.
Sahlin K, Cizinsky S, Warholm M & Hoberg J (1992). Repetitive static muscle contractions in humans - a
trigger of metabolic and oxidative stress? European Journal ofApplied Physiology and Occupational
Physiology 64, 228-236.
Sejersted OM (1992). Electrolyte imbalance in body fluids as a mechanism of fatigue. In: Lamb DR, Gisolfi
CV (eds.), Energy Metabolism in Exercise and Sport (Perspectives in Exercise Science and Sports
Medicine), pp 149-207. Carmel, IN: Brown & Benchmark.
Sejersted OM & V011estad NK (1993). Physiology of muscle fatigue and associated pain. In: Vaerl:lY H,
Merskey H (eds.), Progress infibromyalgia and myofascial pain, pp. 41-51. Amsterdam: Elsevier
Science Publications.
SOderlund K, GreenhaffPL & Hultman E (1992). Energy metabolism in type I and type II human muscle fibres
during short term electrical stimulation at different frequencies. Acta Physiologica Scandinavica 144,
15-22.
Soderlund K & Hultman E (1986). Effects of delayed freezing on content ofphosphagens in human skeletal
muscle biopsy samples. Journal ofApplied PhYSiology 61, 832-835.
Thorstensson A & Karlsson J (1976). Fatiguability and fibre composition of human skeletal muscle. Acta
Physiologica Scandinavica 98, 318-322.
194 N. K. VoUestad

V 011estad NK & Blom PCS (1985). Effect of varying exercise intensity on glycogen depletion in human muscle
fibres. Acta Physiologica Scandinavica 125, 395-405.
V0llestad NK, Sejersted I, Saugen E (1995) Increased relaxation rates during and following intermittent
submaximal isometric contractions. Clinical Physiology In press.
V0llestad NK, Sejersted OM, Bahr R, Woods JJ & Bigland-Ritchie B (1988). Motor drive and metabolic
responses during repeated submaximal contractions in man. Journal of Applied Physiology 64,
1421-1427.
V 0llestad NK, Tabata I & Medb0 II (1992). Glycogen breakdown in different human muscle fibre types during
exhaustive exercise of short duration. Acta Physiologica Scandinavica 144,135-141.
V011estad NK, Vaage 0 & Hermansen L (1984). Muscle glycogen depletion patterns in type I and subgroups
of type II fibres during prolonged severe exercise in man. Acta Physiologica Scandinavica 122,
433-441.
V011estad NK, Wesche J & Sejersted OM (1990). Gradual increase in leg oxygen uptake during repeated
submaximal contractions in humans. Journal ofApplied Physiology 68, 1150-1156.
Wilson JR, McCully KK, Mancini OM, Boden B & Chance B (1988). Relationship of muscular fatigue to pH
and diprotonated Pi in humans: a 31p_NMR study. Journal ofApplied Physiology 63, 2333-2339.
 14

MECHANISMS OF HUMAN MUSCLE

FATIGUE
Quantitating the Contribution of Metabolic Factors and
Activation Impairment

R. G. Miller,' J. A. Kent-Braun,2 K. R. Sharma,3 and M. W. Weiner4

, Departments of Neurology, California Pacific Medical Center and

University of California
San Francisco, California 94118
2Magnetic Resonance Unit, VA Medical Center, Department of Radiology,
University of California
San Francisco, California 94121
3Department of Neurology, University of Miami
Miami Florida 33136
4 Magnetic Resonance Unit, VA Medical Center, Departments of Medicine,
Radiology and Psychiatry, University of California
San Francisco, California 94121

ABSTRACT

Both metabolic factors and impairment of activation appear to playa role in human muscle
fatigue. By measuring force, EMG and metabolites during fatiguing exercise and recovery,
we are attempting to estimate the contribution of the different factors which produce fatigue.

INTRODUCTION

Thirteen years have elapsed since the first studies utilizing 31 phosphorous P) e'
nuclear magnetic resonance spectroscopy (NMR) were applied to healthy human skeletal
muscle to study metabolism and fatigue (Chance et aI., 1981). In the intervening years, there
has been increasing access to NMR techniques and as an enlarging literature attests, the
application of NMR has proven extremely useful in broadening our understanding of
mechanisms offatigue and muscle bioenergetics. In this paper, we will attempt to review the
lessons learned from NMR studies in terms of the metabolic basis of muscle fatigue. We will
highlight how these studies, along with force and EMG measures, have helped to clarify the
role of central fatigue and excitation-contraction-coupling impairment as important mecha-

195
 196 R. G. Miller et aI.

nisms of fatigue. An expanded review of the NMR technique may be found in Kent-Braun
and colleagues (1995). In the 1981 elBA symposium on fatigue, there emerged two camps
of investigators: those who subscribed primarily to the theory that metabolic changes are the
principal cause of muscular fatigue, and the other group who postulated that electrophysi-
ological changes (activation impairment) are the major mechanism of muscle fatigue
(Edwards 1981; Wilkie 1981). We have attempted to analyze simultaneously electrophysi-
ological and metabolic measures of fatigue, and these studies have indicated that both
mechanisms play important roles. The complex interaction between altered muscle metabo-
lism and impaired muscular activation will be the major focus of the present review.

ADVANTAGES AND LIMITATIONS OF NMR

A super-conducting magnet is required for NMR studies of human skeletal muscle

during fatiguing exercise. These examinations require both a spectrometer and non-magnetic
exercise equipment, which allows for isolation of the exercise primarily to the specific
muscle group studied by NMR. Data is usually obtained with the use of a surface coil of a
size and shape that determines the volume of tissue under investigation. The major advantage
ofNMR is that continuous biochemical data can be obtained from exercising human muscle
without causing discomfort or altering the exercise conditions in any way. In the past, muscle
biopsies were required to obtain such information but these involved considerable discomfort
and some interruption of the exercise protocol.
In NMR, radio frequency waves are transmitted through a non-magnetic coil in the
presence of a static magnetic field to sample the relative concentration of metabolites that
contain various nuclei (e.g., phosphorous, proton, or carbon). The frequency is plotted on
the X axis, and the peak height and area of the resonance provide information about the
identity and concentration of each compound (Fig. 1). The area of the peak is proportional
to the concentration of the molecule. Sampling rates for individual spectra vary from 15-60
sec. Very fast spectra have been obtained in as short as 2 sec (DeGroot et aI., 1993).

per

\ Figure 1. Phosphorus-31 magnetic res-

onance spectrum from human tibialis
anterior muscle. The spectrum. ob-
tained at rest, clearly shows the reso-
nances (peaks) from inorganic
phosphate (Pi), phosphocreatine (Per),
and the three phosphorus peaks from
adenosine triphosphate (ATP). The
phosphomonoester region is to the left
of Pi and the phospho diester region is
to the right of Pi. This spectrum was
acquired over 4 min with a 3 cm x 5 cm
elliptical surface coil and has ,IS Hz of
line broadening. A broad component to
the baseline was removed with convo-
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I
lution difference (line broadening =
10 5 0 -5 -10 -15 -20 350 Hz) (reproducted with permission
PPM from Kent-Braun et ai., 1994).
Mechanisms of Human Muscle Fatigue 197

There are important limitations to this technique. As the magnetic bore size is
increased, the isocenter of the magnet is enlarged but the muscle under study must be located
in the isocenter of the magnet; this is critical to reduce noise. Although continuous and
non-invasive measures of muscle metabolism can be obtained, the absolute quantitation of
metabolite concentration is difficult. Metabolite concentrations are usually expressed in
relative terms and with some imprecision. Moreover, because it is impossible to restrict the
sample volume, all determinations are approximate and sometimes involve more than a
single muscle and often less than the entire muscle in question. In addition, because all human
muscles represent a mixture of slow and fast fibers, there is only an average expression of
metabolic changes; and it is not possible to distinguish between the changes within different
fiber types in human muscle. In our experience, these studies are extremely difficult to
schedule when there is competition for magnet time to carry out clinical imaging studies. In
the ideal situation, the magnet is completely committed for experimental investigation with
NMR, as is the case in our laboratory. Even under these conditions, there is usually intense
competition for experimental time in the magnet.

MUSCLE ENERGY METABOLISM

A representative spectrum from resting human muscle using 31p NMR is shown in
Fig. 1. The peak for phosphocreatine (PCr) is the most prominent peak in the center of the
spectrum, and the inorganic phosphate peak (Pi) is readily identifiable to the left. There are
three peaks of adenosine triphosphate (ATP), and these can be easily identified as well.
Intracellular pH can be calculated from the chemical shift of the Pi peak to the right towards
PCr (a distance expressed in parts per million). The adenosine diphosphate (ADP) concen-
tration is small but can be calculated from the spectra as well. Muscle contraction requires
the release of energy from ATP; and in the process of energy release, ADP and inorganic
phosphate are formed. With decreasing intracellular pH, the concentration of an important
product of ATP hydrolysis, the monovalent form of inorganic phosphate (H 2P0 4-), is
increased. This compound and other products of ATP hydrolysis have the potential to inhibit
muscular contraction as will be described below. The creatine kinase reaction is important
in both energy production and utilization, because ADP, PCr and hydrogen ions (H+) are
combined to produce ATP and creatine. The creatine kinase equilibrium reaction may be
used to estimate the intracellular ADP. Chance and colleagues (1985) demonstrated that a
linear relationship existed between Pi/PCr and work rate during relatively low exercise
intensities. This linear relationship had previously been observed in preparations of isolated
mitochondria and suggested that ADP regulation of oxidative phosphorylation is important
under these experimental conditions. Chance and colleagues (1986) also described a hyper-
bolic relationship between work level and energy cost (Pi/PCr) wherein unlimited amounts
of oxygen and substrate are available. Because Pi/PCr can be utilized to estimate intracellular
[ADP], the dynamics of ADP control of oxidative phosphorylation can be calculated, and
the capability of the muscle for oxidative phosphorylation and metabolism can be estimated.
Endurance training increases the capacity for oxidative phosphorylation as measured by this
method (Kent-Braun et aI., 1990).
Important validation of the overall accuracy of NMR techniques was provided by
Wilkie and coworkers (1984) who found the total Pi and PCr in muscle was similar using
muscle biopsy and NMR. The level of Pi was slightly higher while that ofPCr was somewhat
lower in biopsy tissue, a finding that was attributed to some continuing hydrolysis of PCr
prior to freezing biopsy tissue. These studies also demonstrated that ischemia produced a
modest reduction in PCr and a slight increase in ADP and pH. Muscle contraction, however,
was required to produce the onset of glycolysis that was not activated by increased ADP
 198 R. G. Miller et aI.

alone. These studies provided substantial support for the value of NMR measurements in
evaluating metabolism in skeletal muscle.
The role oflimiting oxygen supply to muscle has been studied with a blood pressure
cuff to reduce muscle blood flow. With ischemia, both a higher Pi/PCr and lower pH were
observed at each work load in healthy individuals, suggesting the important role of oxygen
during muscular exercise (Wiener et aI., 1986b). Muscle deoxymyoglobin has also been
measured during exercise using proton spectroscopy (Wang et aI., 1990). Deoxymyoglobin
is not present in resting muscle; but with occlusion of the circulation, deoxymyoglobin can
be detected. Exercise under conditions of ischemia produced a prominent desaturation of
myoglobin, which peaked after approximately 6 min of circulatory occlusion during mus-
cular exercise. The measurement of myoglobin desaturation during exercise may provide
additional useful information about oxygen limitation and blood flow during fatiguing
exercise. In another series of experiments, less tissue deoxygenation during exercise was
found when warm-up preceded the fatiguing muscular exercise (Wiener et aI., 1986a); lower
steady state levels ofPi/PCr and higher steady state levels of pH were attained when subjects
warmed up prior to exercise compared with no warm up.
These observations suggested to us that the experimental conditions may be ex-
tremely important in determining not only the experimental results but also the variability
between subjects during fatiguing exercise carried out with NMR. We therefore studied the
variability offatigue in patients with a standardized exercise protocol of the tibialis anterior
muscle and also examined the factors that might contribute to this variability (Miller et aI.,
1995). Between control subjects, there was marked variability in the degree of muscular
fatigue in the tibialis anterior (Fig. 2.) The speed of recovery was also variable, but the degree
of recovery was rather uniform. To evaluate the effects of blood glucose and also glycogen
storage upon these results, we examined the effect of 48 hours of fasting and also 72 hours
of carbohydrate loading to separately determine the effects of a major dietary manipulation
upon the variability of muscle fatigue. There were no significant differences between the
fasting state, the carbohydrate-loaded state, and the control condition in these subjects. These
results suggest that, at least for high-intensity, short-duration exercise, the variability
between subjects is independent of dietary factors. We also examined the effect of serum pH
upon the variability of muscular fatigue by administering 22 grams of ammonium chloride
orally which lowered the serum pH in arterialized venous blood to levels as low as 7.25.
There was no significant difference in muscular fatigability induced by systemic acidifica-
tion compared to controls. These findings suggest that the muscle buffering capability is
sufficient to neutralize the impact of any changes in serum pH. We also examined the
variability of muscular fatigue in the same subject on different days, and found excellent

120

100
.......
"0 80
:;;
:s
~ 60
u
>
::;; 40
Figure 2. Maximal voluntary
20 contraction (MVC) during and
following a 4-min maximal sus-
0 tained contraction of the tibialis
0 2 4 6 8 10 12 14 16 18 anterior (reproduced from Miller
et a!., 1995, by permission of
TIme (min) John Wiley & Sons, Inc.).
Mechanisms of Human Muscle Fatigue 199

reproducibility for each subject. The major factors that were not manipulated in our studies,
and which probably contribute to the variability of muscular fatigue between subjects, are
fiber-type composition of the muscle which is controlled by genetic factors and the level of
physical conditioning. It is likely that these are the major determinants of variability in
muscular fatigability between human subjects, and both of these factors deserve more
systematic examination in the future.

METABOLIC FACTORS IN MUSCLE FATIGUE

In our prior work, we define fatigue as the reduction in maximal force generating
capability of the muscle during exercise. NMR is an excellent technique for continuously
and non-invasively quantitating the changes in muscle metabolism during fatiguing exercise.
In 1986, Taylor and colleagues demonstrated that some depletion of ATP can occur with high
intensity exercise in normal human subjects when per depletion exceeds 80% (Taylor et aI.,
1986). Intracellular pH had fallen to 6.2, indicating the intense degree of muscular exercise.
The calculated free energy of ATP hydrolysis was similar in moderate exercise and during
high intensity exercise. Following intense exercise, the rate of recovery of Pi, per, and pH
was slower than after moderate exercise. The recovery of ATP was slower than that of other
metabolites. Thus, intense fatiguing exercise could lead to a pronounced depletion of
high-energy phosphates, and also produce slowing of metabolic recovery.
Over the past decade, we have carried out experiments to examine the relationship
between muscle fatigue and changes in various metabolites during different types of exercise
(Miller et aI., 1987; Miller et aI., 1988; Boska et aI., 1990; Weiner et aI., 1990; DeGroot et
aI., 1993). We monitored the time course of the changes in force, intracellular pH, per,
compound muscle action potential amplitude, and the rectified integrated electromyogram
(RIEMG) in the exercising adductor pollicis muscle of healthy volunteer subjects (Miller et
aI., 1987). During fatiguing exercise, maximal voluntary force fell to only 10% of initial;
and intracellular pH fell to 6.4, while per was almost completely depleted. Both the
neuromuscular efficiency (forceIRIEMG) and the compound muscle action potential ampli-
tude decreased during fatigue. However, the duration of the compound muscle action
potential amplitude increased, and the overall area of the potential was preserved so that
changes in muscle membrane excitability and neuromuscular transmission were probably
not responsible for fatigue in this experimental paradigm. The pattern of recovery (Fig. 3)
suggested three components of fatigue: first there was a rapidly recovering alteration of
muscle membrane excitation and impulse propagation as reflected by changes in the
compound muscle action potential; second, there was a more slowly recovering alteration
in the metabolic state of the muscle; and third, there was a long duration failure of
excitation-contraction coupling as reflected by impaired neuromuscular efficiency and
reduced twitch tension.
In a subsequent study, we compared the metabolic and physiological changes during
fatigue resulting from intermittent and sustained exercise where the source of energy was
largely from aerobic and anaerobic metabolism, respectively (Miller et aI., 1988). The rapid
force and metabolite changes in the adductor pollicis during sustained exercise are shown
in Figs. 4A and 4C. During intermittent exercise, there were more gradual changes in both
force and metabolites (Figs. 4B and 4D). However, the relationships between force and both
H+ and H2P04- were closer than those between force and either per or Pi. This was a finding
that was consistent in both exercise protocols, and led us to conclude that H+ and H 2P04-
may both play important roles in the development of muscle fatigue.
The close relationships between force and both H+ and H2P04- were also observed
in the less fatigable tibialis anterior muscle, and demonstrated the constant nature of this
200 R. G. Miller et al.

7.0

6.5
a

/
~~o~t~~~~~+-~--'--'~-r--~~--'--'~-r--~~--,-~~~--~~6.0
o 2 3 II 5 6 8 9 10 11 12 13 14 15 20
RECOVERY T11£ (mln.)

Figure 3. Recovery of neuromuscular efficiency, phosphocreatine, pH, and maximal voluntary contraction
after a 4 min fatigue test (reproduced from Miller et a!., 1987, by permission of John Wiley & Sons, Inc.).

7.' A 130 7.' B 140

120 130
7.2 110 7.2 120
too•
.'l
90 ~
80;;
70 ~
60~
50!!

,.
.40 ~

6.2 20
10 10
6.0L..~-~-~~~~~~~-~~....J0 6.0L......~~-~~-~~~--~-~...J0

o to 12 1.4 16 18 o 10 15 20 25 30 35 .40 45 50 55
Tile (1m) hie 'lin)

40 "c.-----------------. IIo 40 r.D;---------------~1I0

~ 100 ~ 100

25
!
~2O
% 15

10 10

10 12 t.4 16 18 W ffi 20 25 30 E 40 6 50
Till (lIn) hlle (1m)

Figure 4. pH (Ll) and MVC (D) plotted as a function of time during 4 min of sustained (A) and intermittent
(B) exercise. H2 P04 - (Ll) and MVC (D) plotted as a function of time during 4 min of sustained (C) and
intermittent (D) exercise (reproduced from Miller et a!., 1988, by copyright permission of The Society for
Clinical Investigation).
Mechanisms of Human Muscle Fatigue 201

relationship (Weiner et al., 1990). A subsequent study of the relationship between the
recovery of force and metabolites following fatiguing exercise also suggested that H2P0 4-
was the metabolic measure most closely related to maximal force generating capacity(Boska
et al., 1990). Thus, we provided evidence for an important role of diprotonated Pi (H2P0 4-)
in the development of muscle fatigue. This is not to say that intracellular pH and total Pi are
not important determinants of muscle fatigue, but the single metabolite that correlated most
closely with muscular fatigue was the build-up ofH 2P04-. Other studies have documented
a strong linear relationship between fatigue and H2P04- in the wrist flexor muscles of healthy
human subjects (Cady et al., 1989a). Thus in 3 muscles and with 3 different exercise
protocols, there is evidence that the relationship with H2P0 4- and fatigue was stronger and
more consistent than the relationship between pH and fatigue.
Cady and co-workers (Cady et al., 1989b; see also Hultman et al., 1981) examined
the relationship between changes in muscle metabolism and muscular fatigue. They found
a linear relationship between H2P04 - and fatigue in the fIrst dorsal interosseous muscle of
healthy subjects. No linear relationship was found, however, in a subject with McArdle's
disease who, because of defective glycogenolysis, could not utilize muscle glycogen (Cady
et al., 1989b). These investigators also noted a dissociation between fatigue and pH that was
particularly prominent during the recovery after fatiguing exercise. They also provided
evidence that there is a pH-independent component of muscular fatigue that may be related
to the slowed rate of force relaxation observed during muscular fatigue. This was demon-
strated by a 50% slowing of relaxation in fatigued muscle at a time when pH remained
unchanged.
The examination of muscle oxidative potential was also the focus of a recent study
of the ankle dorsiflexor muscle group (Kent-Braun et al., 1993a). A sequence of three
different metabolic phases was found during the transition from rest to a fatiguing level of
muscular exercise (Fig. 5).

A 120
•
I
I
100
• • 4 I
r
• •
~

~
'2 80

E •
U 50
>
2
40

20

B 2.5
•

Figure 5. Changes in MVC (A) and PilPCr and [W] (D) vs. +
time in 1 subject. Bilinear fit is included to illustrate the 3 .::. 0.5

phases of metabolism delineated by inflection points for PilPCr

(e) and [W] (XlO-7,O) vs. time. Dashed vertical lines indicate 0.0 &...--......._ _ _ _ _.................

location of inflection points (reproduced with permission from o 2 4 6 B 10 12 14 16

Kent-Braun et ai., 1993a). Time (min)
202 R. G. Miller et al.

The length of the first phase, during which energy for the muscle contraction was
produced primarily by oxidative phosphorylation, was closely related to the oxidative
potential of the muscle (i.e., the initial slope of force vs. Pi/PCr). This first phase was
followed by an intermediate phase, during which Pi/PCr was increasing at a more rapid rate
but H+ remained relatively unchanged. In a third phase, during which glycolytic sources of
ATP began to contribute more significantly to energy production, there was an inflection in
H+ with continued steady increases in Pi/PCr. It was the third, non-steady state phase that
was associated with the production of muscular fatigue. These studies demonstrated the
utility Of31p NMR for the study of relationships between oxidative and glycolytic metabo-
lism during progressive muscular exercise. By comparing the time course of metabolic
changes involved primarily in either oxidative or glycolytic metabolism, insight can be
gained into the interplay between these pathways and the development of fatigue.
Thus, numerous studies of muscle fatigue in healthy human subjects have demon-
strated the importance of altered metabolites (e.g., pH, Pi, H 2P04 -) in the development of
muscular fatigue during various types of exercise. However, fatigue may also develop due
to activation impairment, especially during intermittent long duration exercise. This is
discussed in more detail below.

CENTRAL FATIGUE

Impaired firing rate modulation or reduced motor unit recruitment during fatiguing
exercise is often referred to as central fatigue. Any breakdown in muscular activation that is
proximal to the point of stimulation of the motor nerve is generally considered central.
Impaired central activation of muscle during exercise is measured in two different ways.
During a maximal voluntary contraction, superimposed electrical stimulation of the motor
nerve, which may produce added force from incompletely activated muscle fibers, provides
a quantitative measure of central activation failure. When the muscle is fully activated, the
superimposed electrical stimulus produces no added force (Bigland-Ritchie et aI., 1986a;
Lloyd et aI., 1991). Previous workers have variously utilized the superimposed twitch, paired
stimuli, and brief trains of tetanic stimuli (e.g., Gandevia & McKenzie, 1985; Bigland-
Ritchie et aI., 1986ab; Lloyd et aI., 1991; Kent-Braun et aI., 1993b; Sharma et aI., 1994b).
In a comparative study, we found that a train of 50 Hz tetanic stimulation is a more sensitive
method of detecting added force (unpublished data). Therefore, we have utilized this measure
as a primary means of evaluating central fatigue.
An additional technique utilizes a comparison of the decline in maximal voluntary
contraction strength and the decline in tetanic tension during fatiguing exercise. When
fatigue is entirely peripheral to the point of electrical stimulation of the motor nerve, the
decline in MVC and tetanic tension should be identical. When fatigue is central, at least in
part, there should be a greater decline in MVC compared with the decline in tetanic tension.
We have used both of these techniques to evaluate central fatigue in a number of studies that
will now be summarized.
In the pre-fatigued tibialis anterior muscle of sedentary healthy subjects, we found
during a maximal voluntary contraction that there was no added force (Kent-Braun et aI.,
1993b). During a sustained 4-min maximal isometric voluntary contraction, there was a small
degree of added force with increasing fatigability. At the end of the 4th min, when fatigue
was significant, the added force averaged 15% of the voluntary force generated by the
muscle. Larger degrees of added force were found in a study of healthy children (age 6-10)
using the same isometric sustained contraction protocol (Sharma et aI., 1995c). The children
were quite healthy but had some difficulty with concentration and sustaining motivation.
They demonstrated substantial added force at the end of 4 min, indicating significant central
Mechanisms of Human Muscle Fatigue 203

impairment of activation of the muscle. In both children and adults, there was a tendency
for the added force to steadily increase during the 4 min of sustained contraction, suggesting
that some central fatigue is commonly found in healthy subjects during arduous, sustained
muscular fatiguing exercise. Surprisingly, and by contrast, central fatigue in an age-matched
group of boys with Duchenne muscular dystrophy was much less at the end of 4 min of
fatiguing exercise compared with the healthy children (Sharma et aI., 1995c) suggesting that
boys with Duchenne muscular dystrophy have improved activation of muscle either as a
compensation for muscular weakness or because they are more used to the testing proce-
dures.
In contrast, in a recent study of patients with chronic fatigue syndrome, we found
much greater central fatigue in patients compared with sedentary adults who served as
age-matched control subjects (Kent-Braun et aI., 1993b; cf. Lloyd et aI., 1991). No other
neuromuscular abnormality was found in evaluating muscular fatigue in chronic fatigue
syndrome patients. Specifically, during fatiguing exercise, neither the decline in voluntary
muscle force generation nor the change in tetanic tension were significantly different from
controls. The compound muscle action potential amplitude was preserved during fatiguing
exercise in both patients and controls. Metabolic changes were, if anything, less in patients
than in controls, consistent with reduced activation of muscle by patients during exercise.
These studies document the presence of a central component offatigue in some patients with
chronic fatigue syndrome, although the precise cause of the central fatigue remains unclear.
Possibilities include reduced attention, muscular pain, deconditioning, and reduced motiva-
tion.
We also found that a unique type of central fatigue was produced by rapid repetitive
muscular contractions (Miller et aI., 1993). A series of repetitive isometric contractions was
carried out with the ankle dorsiflexor muscles in normal human subjects beginning with a
rate of 24/min and gradually increasing every 3 min to a maximum frequency of 72
contractions per min. The subject was instructed to execute each contraction as fast as
possible to reach the target force of 40% MVC. Throughout the 15-min exercise period, there
was a steady decline in maximal voluntary contraction strength down to approximately 50%
of initial force, and a decrease in intracellular pH to 6.5. The changes in spectra, both during
exercise and recovery, are shown in Fig. 6.
During the first 1 or 2 min, the changes in metabolism and maximal voluntary
contraction strength were slight. As exercise increased, the metabolic changes steadily
developed. By contrast, there was a decrease in the speed of voluntary force generation
within the first min of fatiguing exercise (Figs. 7 and 8).
At this time, there was only a minimal alteration in metabolites and MVC. Moreover,
the speed of tension generated in the tetanus was unchanged at this stage, and the speed of
tension generated in the twitch was actually increased due to twitch potentiation. But, the
duration of the EMG burst associated with the rapid voluntary movements was prolonged
by approximately 20%-25% (Fig. 8). Thus, during rapid repetitive movements, there are two
components offatigue: a steadily progressive change in metabolism and a decline in maximal
force generation, but an earlier slowing of force generation during rapid voluntary muscle
contractions that is not present during electrically stimulated muscle contraction in either
the twitch or the tetanus. Because the slowed speed of voluntary force generation is correlated
with a prolonged EMG burst duration, there must be a slowing of recruitment and/or firing
rate modulation of motor unit potentials during a rapid voluntary contraction. This type of
central fatigue, which develops so early in this exercise protocol, may be relevant to many
activities of daily living including piano playing, keypunching, typing, etc. Further studies
of rapid repetitive movements are indicated both in healthy individuals and in various disease
states where the ability to execute rapid movements is often selectively impaired.
 204 R. G. Miller et al.

iii

-20 PPM
Figure 6. A stacked plot of the I-min 31p NMR spectra acquired during a typical rapid exercise protocol
(reproduced with permission from Miller et aI., 1993).

--
Figure 7. Superimposed force recordings from the adductor pollicis of a representative subject during rapid
voluntary contractions to a 40% MVC target (A), tetanic stimulation (B), and a single twitch (C). Each
recording was made at the beginning and again after 6 min of rapid repetitive contractions (reproduced with
permission from Miller et aI., 1993).
Mechanisms of Human Muscle Fatigue 205

Figure 8. Force and EMG recordings at the beginning (top) and after I min of rapid repetitive contractions
(bottom) with the adductor pollicis muscle from one subject. EMG signals were recorded with surface
electrodes (bandpass filter settings 1.6 Hz to 16 kHz). Records are aligned at the onset of the increase in force
(reproduced with permission from Miller et aI., 1993).

EXCITATION-CONTRACTION-COUPLING IMPAIRMENT

The other likely source of activation impairment, besides central fatigue, is excita-
tion-contraction coupling. However, this is the most difficult aspect of neuromuscular
function to quantitate during fatiguing exercise. Although it is impossible in healthy human
subjects to measure the impulse propagation in the transverse tubule or the amount and rate
of calcium released by the sarcoplasmic reticulum, indirect measures of excitation contrac-
tion coupling are available. Probably the most reliable measure of impaired excitation
contraction coupling is the existence of a long duration fatigue following the completion of
muscular exercise. Long duration fatigue has also been called low-frequency fatigue, and
the phenomenon was first attributed to excitation-contraction-coupling impairment by
Edwards and colleagues (1977). These investigators found that after exhaustive exercise
there was a long-lasting impairment of the force produced by low-frequency stimulation,
with an earlier recovery of the force produced by high-frequency stimulation. In a subsequent
study of intermittent low intensity exercise, Bigland-Ritchie and colleagues observed a form
offatigue that was associated with very little alteration in metabolites as analyzed by muscle
biopsy (Bigland-Ritchie, et aI., 1986b). A particularly pronounced depression of the twitch
and a more modest depression of maximal voluntary force generation was noted and
attributed to excitation-contraction-coupling impairment. In a subsequent study, we ob-
served that low intensity intermittent exercise could produce not only fatigue associated with
very little metabolic alteration as examined by NMR spectroscopy, but also with some long
lasting impairment of muscular force generation (Moussavi et aI., 1989). Again, this pattern
was attributed to excitation-contraction-coupling impairment.
206 R. G. Miller et al.

A
......100 ~~
o

f S:::·~_~--i--
~ 80
~
(.) 50 .,"
~
40 Ijl
~0--~--5~--~-1~0~--~15~

100 B
......
;g 80
:~
~50
~ Figure 9. Maximal voluntary contraction (MVC) force recov-
::I! 40 ery after long-duration exercise ('Y; n = 5) and short-duration
exercise (0; n = 6). A, nuclear magnetic resonance experi-
ments. B, neurophysiological experiments (reproduced with
Time (min) permission from Baker et aI., 1993).

More recently, our group has compared the recovery from short duration exercise
during two min of maximal sustained contraction under ischemic conditions with a longer
exercise protocol (20-25 min) under aerobic conditions (Baker et ai., 1993). The speed of
recovery of both tetanic tension and maximal voluntary contraction was markedly slowed
in the long duration exercise compared to short duration exercise (Fig. 9). The recovery of
twitch tension was particularly slow in the subjects who carried out long duration exercise.
By contrast, the recovery of metabolites (PCr, Pi, and H 2P04 -) was prompt and complete in
both groups. The PCr recovery was, if anything, slower in the subjects who performed
short-duration exercise. In both groups, metabolic recovery was complete after 10 min at a
time when the recovery of both MVC and tetanic tension was incomplete.
The evoked compound muscle action potential amplitude was not significantly
reduced following either type of exercise. The amplitude of the surface EMG signal was
somewhat depressed after both types of exercise, but was more markedly reduced after short
duration exercise. Within the first min, there was a rapid recovery of the excessive EMG
impairment associated with the short duration exercise.
These observations document the presence of a long-lasting fatigue that is particu-
larly striking for low-frequency force (twitch tension). The absence of any alteration of the
compound muscle action potential suggests that neither neuromuscular transmission impair-
ment nor altered muscle membrane excitability can explain the long-lasting fatigue. More-
over, the prompt recovery of metabolites following both types of exercise suggests that this
is not a defect in energy metabolism induced by long-duration exercise. Similarly, the
comparable decline in tetanic tension and maximal voluntary strength strongly suggests that
fatigue was peripheral rather than central in origin.
The contribution of metabolic changes to long-duration fatigue was estimated from
the time course of recovery of inorganic phosphate and maximal voluntary contraction
strength (Fig. 10).
Following short-duration exercise, the time course of recovery of inorganic phos-
phate and muscular force are almost identical suggesting that the recovery offorce depended
primarily upon the recovery of metabolites after short-duration exercise. By contrast, the
recovery offorce was delayed following long-duration exercise compared with the recovery
of inorganic phosphate and other metabolites. As already indicated, the absence of change
in the compound muscle action potential suggests that the neuromuscular junction and
Mechanisms of Human Muscle Fatigue 207

:r la,

10 I:P.

IS

I
30

20
....
0
:P.

I
10

Figure 10. Mean level offatigue (expressed as %MVC

0
reduction) and mean Pi elevation after short-duration
exercise (A) and long-duration exercise (B) (reproduced o S 10 IS
with permission from Baker et aI., 1993). Time(mio)

muscle membrane did not contribute significantly to the muscle fatigue. The central com-
ponent of fatigue was judged to be insignificant because the parallel and similar changes
were observed in tetanic tension and MVC during fatigue, and also because the recovery of
force was independent of any change in EMG. Thus, the contribution of excitation-contrac-
tion-coupling impairment was estimated as the remaining mechanism of fatigue unaccounted
for by the other mechanisms elaborated above (Baker et aI., 1993).
We attempted to estimate the quantitative contribution of metabolic changes, excita-
tion-contraction-coupling impairment, altered neuromuscular transmission, and central fa-
tigue to the overall muscular fatigue induced by both short-duration exercise and
long-duration exercise (Fig. 11).
At the conclusion of fatiguing short-duration exercise, most of the fatigue could be
explained by metabolic changes. Some alteration of excitation-contraction coupling could
also be present, but was not identifiable using our techniques. Because ofthe reduced surface
EMG signal, a slight degree of central fatigue was also suggested in both types of exercise.
Similarly, because of the very slight reduction in compound muscle action potential ampli-
tude, a mild contribution of neuromuscular-transmission impairment was possible. However,

60
o metobolite
• ee
• nmj/membrone
~ ens

Figure 11. Histogram of recovery from fatigue after short-dura-

tion (S) and long-duration exercise (L) at 0,5, 10, and 15 min of
recovery. Relative contributions of 4 different fatigue mecha-
nisms are indicated by shading in each column. EC, excitation-
contraction coupling; NMJ, neuromuscular junction; CNS,
central nervous system. The mean level of fatigue (expressed as o
% MVC reduction) is indicated on the ordinate (reproduced with Time
permission from Baker et aI., 1993).
 208 R. G. Miller et al.

in the long-duration exercise, a significant impairment of excitation-contraction coupling

appeared to be present immediately after exercise; and this appeared to become more
prominent during the recovery phase as indicated in Fig. 11. We have probably underesti-
mated, if anything, the contribution of excitation-contraction-coupling impairment in the
present study, particularly at the conclusion of exercise. Precise quantitation of these
parameters is clearly not possible using the methods of this study, but in a general way it
helps us to begin to focus on the role of individual mechanisms of fatigue at different stages
of fatigue and recovery and in different types of exercise, both in health and in disease. The
role of excitation-contraction-coupling impairment appears to be even more prominent in
some neuromuscular diseases. In patients with post-polio syndrome and amyotrophic lateral
sclerosis, we have found even more marked impairment of excitation-contraction coupling
(Sharma et aI., 1994b; Sharma et aI., 1995a).
In summary, NMR spectroscopy in conjunction with force and EMG measurements
has been useful in allowing us to broaden our understanding of the mechanisms of fatigue
in health and disease. As emphasized by Bertocci, Chapter 15, when NMR allows measure-
ment of calcium, sodium and potassium levels as well as glycogen, high-energy phosphates,
and intracellular pH, a much more comprehensive understanding of fatigue in human muscle
will be possible.

ACKNOWLEDGMENTS

The authors' research has been supported, in part, by grants from the Multiple
Sclerosis Society (USA), and the Muscular Dystrophy Association (USA).

REFERENCES
Baker AJ, Kostov KG, Miller RG & Weiner MW (1993). Slow force recovery after long duration exercise:
Metabolic and activation factors in muscle fatigue. Journal ofApplied Physiology 74, 2294-2300.
Bigland-Ritchie B, Cafarelli E & Vollestad NK (1986a). Fatigue of submaximal static contractions. Acta
Physiologica Scandinavica 128 (suppl. 556), 137-148.
Bigland-Ritchie B, Furbush F & Woods JJ (1986b). Fatigue of intermittent submaximal voluntary contractions:
central and peripheral factors. Journal of Applied Physiology 61, 421-429.
Boska MD, Moussavi RS, Carson PJ, Weiner MW & Miller RG (1990). The metabolic basis of recovery after
fatiguing exercise in human muscle. Neurology 40, 240-244.
Cady EB, Elshove H, Jones DA & Moll A (1989a). The metabolic causes of slow relaxation in fatigued human
skeletal muscle. Journal of Physiology (London) 418, 327-337.
Cady EB, Jones DA, Lynn J & Newham DJ (1989b). Changes in force and intracellular metabolites during
fatigue of human skeletal muscle. Journal ofPhysiology (London) 418, 311-325.
Chance B, Eleff S, Leigh JS, Sokolow D & Sapega A (1981). Mitochondrial regulation of phosphocreatine-
phosphate ratios in exercising human muscle: A gated 31p NMR study. Proceedings o/the National
Academy o/Sciences o/the United States 0/ America 78, 6714-6718.
Chance B, Leigh JS, Clark BJ, Maris J, Kent J, Nioka S & Smith D (1985). Control of oxidative metabolism
and oxygen delivery in human skeletal muscle: A steady-state analysis of the work/energy cost transfer
function. Proceedings of the National Academy of Sciences of the United States of America 82,
8384-8388.
Chance B, Leigh JS, Kent J, McCully K, Nioka S, Clark BJ, Maris 1M & Graham T (1986). Multiple controls
of oxidative metabolism in living tissues as studied by phosphorus magnetic resonance. Proceedings
o/the National Academy o/Sciences o/the United States 0/America 83,9458-9462.
DeGroot M, Massie BM, Boska M, Gober J, Miller RG & Weiner MW (1993). Dissociation of [H+] from
fatigue in human muscle detected by high time resolution 31p_NMR. Muscle & Nerve 16, 91-98.
Mechanisms of Human Muscle Fatigue 209

Edwards RHT (1981). Human muscle function and fatigue. In: Porter R, Whelan J (eds.), Human Muscle
Fatigue: Physiological Mechanisms. Ciba Foundation Symposium No. 82, pp. 1-18. London:Pitman
Medical.
Edwards RHT, Hill DK, Jones A & Merton PA (1977). Fatigue ofiong duration in human skeletal muscle after
exercise. Journal of Physiology (London) 272, 769-778.
Gandevia SC, McKenzie DK (1985). Activation of the human diaphragm during maximal static efforts. Journal
ofPhysiology (London) 367, 45-56.
Hultman E, Sjoholm H, Sahlin K & Edstrom L (1981). Glycolytic and oxidative energy metabolism and
contraction characteristics of intact human muscle. In: Porter R, Whelan J (eds.), Human Muscle
Fatigue: Physiological Mechanisms. Ciba Foundation Symposia No. 82, pp. 19-40. London: Pitman
Medical.
Kent-Braun JA, McCully KK & Chance B (1990). Metabolic effects of training in humans: a 3Ip_MRS study.
Journal ofApplied Physiology 69, 1165-1170.
Kent-Braun JA, Miller RG, Weiner MW (1994). Magnetic resonance spectroscopy studies of human muscle.
Radiologic Clinics ofNorth America 32, 313-335
Kent-Braun JA, Miller RG & Weiner MW (1993a). Phases of metabolism during progressive exercise to fatigue
in human skeletal muscle. Journal ofApplied Physiology 75, 573-580.
Kent-Braun JA, Miller RG & Weiner MW (1995). Magnetic resonance spectroscopy: studies of human skeletal
muscle metabolism in health and disease. In: Holloszy JO (ed.), Exercise and Sport Science Reviews
23, pp. 305-347. Baltimore: Williams and Wilkins.
Kent-Braun JA, Sharma KR, Weiner MW, Massie B & Miller RG (1993b). Central basis of muscle fatigue in
chronic fatigue syndrome. Neurology 43,125-131.
Lloyd AR, Gandevia SC & Hales JP (1991). Muscle performance, voluntary activation, twitch properties and
perceived effort in normal subjects and patients with chronic fatigue syndrome. Brain 114, 85-98.
Miller RG, Boska MD, Moussavi R, Carson PJ & Weiner MW (1988). 31 P Nuclear magnetic resonance studies
of high energy phosphates and pH in human muscle fatigue. Comparison of aerobic and anaerobic
exercise. Journal of Clinical Investigation 81, 1190-1196.
Miller RG, Carson PJ, Moussavi RS, Green A, Baker A, Boska MD & Weiner MW (1995). Factors which
influence alterations of phosphates and pH in exercising human skeletal muscle: measurement error,
reproducibility, and effects of fasting, carbohydrate loading, and metabolic acidosis. Muscle & Nerve
18,60-67.
Miller RG, Giannini D, Milner-Brown HS, Layzer RB, Koretsky AP, Hooper D & Weiner MW (1987). Effects
of fatiguing exercise on high-energy phosphates, force and EMG: Evidence for three phases of
recovery. Muscle & Nerve 10, 810-821.
Miller RG, Moussavi RS, Green AT, Carson PJ & Weiner MW (1993). The fatigue of rapid repetitive
movements. Neurology 43, 755-761.
Moussavi RS, Carson PJ, Boska MD, Weiner MW & Miller RG (1989). Nonmetabolic fatigue in exercising
human muscle. Neurology 39, 1222-1226.
Sharma KR, Weiner MW, Miller RG, Kent-Braun J, Majumdar S, Huang Y & Mynhier M (1995a). Physiology
of fatigue in amyotrophic lateral sclerosis. Neurology 45,733-740.
Sharma KR, Kent-Braun J, Mynhier MA, Weiner MW & Miller RG (1994b). Excessive muscular fatigue in
the postpoliomyelitis syndrome. Neurology 44, 642-646.
Sharma KR, Mynhier M & Miller RG (1995c). Muscular fatigue in Duchenne muscular dystrophy. Neurology
45, 306-310.
Taylor DJ, Styles P, Matthews PM, Arnold DA, Gadian DG, Bore P & Radda GK (1986). Energetics of human
muscles: Exercise- induced ATP depletion. Magnetic Resonance in Medicine 3, 44-54.
Wang Z, Noyszewski EA & Leigh JS (1990). In vivo MRS measurement of deoxymyoglobin in human
forearms. Magnetic Resonance in Medicine 14, 562-567.
Wiener DH, Fink LI, Maris J, Jones RA, Chance B & Wilson JR (1986a). Abnormal skeletal muscle
bioenergetics during exercise in patients with heart failure: Role of reduced muscle blood flow.
Circulation 73, 1127-1136.
Wiener DH, Maris J, Chance B & Wilson JR (1986b). Detection of skeletal muscle hypoperfusion during
exercise using phosphorus-31 nuclear magnetic resonance spectroscopy. Journal of the American
College of Cardiology 7, 793-799.
Weiner MW, Moussavi RS, Baker AJ, Boska MD & Miller RG (1990). Constant relationships between force,
phosphate concentration, and pH in muscles with differential fatigability. Neurology 40,1888-1893.
Wilkie DR (1981). Shortage of chemical fuel as a cause offatigue: studies by nuclear magnetic resonance and
bicycle ergometry. In: Porter R, Whelan J (eds.), Human Muscle Fatigue: Physiological Mechanisms.
Ciba Foundation Symposium No. 82, pp. 102-119. London: Pitman Medical.
210 R. G. Miller et al.

Wilkie DR, Dawson MJ, Edwards RHT, Gordon RE & Shaw D (1984). 31p NMR studies of resting muscle in
normal human subjects. In: Pollack, GH & Sugi, H (eds.) Contractile Mechanisms in Muscle,
pp.333-347. New York: Plenum Publishing Corporation.
 15

EMERGING OPPORTUNITIES WITH NMR

L. A. Bertocci

Institute for Exercise and Environmental Medicine, Presbyterian Hospital

of Dallas
Dallas Texas 75231 and
Rogers Magnetic Resonance Center, Department of Radiology
UT Southwestern Medical Center
Dallas Texas 75235

ABSTRACT

Several nuclear magnetic resonance (NMR) methods with potential as tools for the study of
neuromuscular fatigue are described briefly. 13C MR spectroscopy (MRS) is presented as a
means of studying the regulation of substrate (fuel) flux into the citric acid cycle. 23 Na and
39K MRS can be used to study the distribution of sodium and potassium ions across the cell
membrane. 1H MR imaging (MRl) takes advantage of the relationship between the environ-
ment of water 1H and the mechanisms of nuclear spin relaxation to make static and cine
images which present spatial (anatomical) information that can be tied to essential physi-
ological, biochemical, or biophysical phenomena of fatiguing muscle.

INTRODUCTION

Since the discovery of the physical property of nuclear magnetic resonance (NMR;
Bloch et aI., 1946, Purcell et aI., 1946) new applications continue to be found for NMR.
Although most of the early NMR work was done using lH magnetic resonance spectroscopy
(MRS) in NMR systems with narrow bore resistive magnets up to 1.5T in strength, it was
the development of larger, superconducting magnets that has made it possible for most of
these new applications and, in particular, has made it possible to apply NMR techniques to
the study of intact organs in vivo. The application ofNMR to living systems, ranging from
chemical analyses of organic and bio-organic compounds to the more recent studies of whole
micro-organisms, intact tissues, through to entire animals and human clinical patients,
represents an extraordinary advance in modern biological and clinical research.
Several NMR techniques have begun to be applied to the study of neuromuscular
fatigue. Foremost is the application Of 31 p MRS to measure the changes in relative concen-
tration of high energy phosphates (ATP, ADP, and PCr) or pH that occur during steady-state
and fatiguing exercise. However, there are other NMR techniques in use in other experimen-
tal settings which, if properly adapted to the fatiguing neuromuscular unit, may provide even

211
212 L. A. Bertocci

more fundamental biological information about neuromuscular fatigue. Several such emerg-
ing opportunities for the study of neuromuscular fatigue will be described in this chapter.
Although by no means an exhaustive list, the more promising methods will be described
theoretically, with examples of their current utility and suggestions for their future applica-
tions.

SPECTROSCOPY

Carbon atoms are the basic building blocks of nearly all organic compounds, from
the CO2 produced during basic respiration to the amino acid sequences of large molecular
weight proteins. To make the energy required for muscle to perform useful work, carbon-
containing compounds (fuels, or more properly called substrates) must flow through the
pathways of intermediary metabolism where in a series of biochemical reactions, they are
ultimately converted to CO2, H20, and energy. Insofar as muscle fatigue can be considered
to be an excess of energy demand vs. an inadequate amount of energy supply, the study of
these reactions is critically important to the understanding of the causes of muscle fatigue.
These reactions have been studied primarily using labeled compounds, either sub-
strates or substrate analogs (for example [1_14C)glucose, [lor 2- 14C)acetate or [1, 2, or
3_ 14C)pyruvate) (Katz, 1985) and or 2-deoxy-d-[2,6-3H]glucose) (Fushiki et aI., 1991)
labeled radioactively or with non-radioactive stable isotopes (for example [U-l3C)glucose
(Katz & Lee, 1991; Lee et aI., 1991). The progress of the chemical reactions of interest can
be monitored by several methods, most often by some combination of one-time or serial
measurements of the 14C02, 3H20 or other 3H_ or 14C_ containing compounds that are
produced. Alternatively, l3C labeled compounds can be introduced and incorporation into
the products can be determined by mass spectrometry and mass isotopomer analysis (Katz
& Lee 1991; Lee et aI., 1991).
Unfortunately, there are disadvantages to all of these methods, such as 1) the need to
handle radioactively labeled compounds; 2) the difficulty in discriminating between different
metabolic fates or pathways; and 3) the small number of discrete results can be determined
from a single experiment. As an alternative to such techniques, l3C MRS has been applied
more recently to the study of intermediary metabolism in vivo. The first demonstrations of
this technique were in bacteria (Walker et aI., 1982) and the isolated, perfused rat liver
(Cohen, 1983), although more rigorous mathematical treatments (Chance et aI., 1983) of the
resultant spectra have been developed which allow a great deal of information to be extracted
from a single spectrum (Malloy et aI., 1990; Jeffrey et aI., 1991).
Isotopomer analysis of glutamate (02C_2CHNH2-3CH24CHz-C02), either the non-
steady state analysis using the C4 alone or the steady-state analysis using C2, C3, and C4 in
combination, provides a straightforward measurement ofthe relative contribution of various
substrates to the citric acid cycle. This information can be used to compare the relative flux
of acetyl-CoA to the TCA cycle from 1) glycogenolysis vs. fatty acid fl-oxidation; 2) labeled
vs. unlabeled substrates; and 3) energy production vs. anaplerosis. The principal advantages
of l3C MRS compared with more traditional approaches to studying substrate regulation
(e.g., measuring plasma concentrations of substrates, calculation of respiratory quotient, 14C
radioisotope studies) include 1) safety from use of radioactive isotopes (l3C is a naturally
occurring, stable isotope of carbon); 2) convenience (a large amount of information is
available from a single spectrum; 3) simultaneous measurement of the contribution of
multiple substrates to the citric acid cycle; 4) measurement of energy producing and
anaplerotic activities; and 5) both steady-state and non-steady state analyses.
Emerging Opportunities with NMR 213

Carbohydrate Fatty Acids

g~~~ ;~d"~ \

O\::JG .(--.GIT
~Citrate~
r
Acetyl-CoA
Citric \
Ie

13c MRS analysis

Figure 1. A simplified schematic of the biochemical rationale for the design of the proposed l3C MRS
experiments.

The relevant properties of the magnetic resonance signal from a detectable atom
include 1) the effect of the chemical environment of the detectable atom on the frequency at
which it resonates; and 2) the proportionality between the intensity of the spectra peaks and
the number of detectable atoms in an individual chemical environment. In the case of an
organic molecule such as glutamate, each carbon experiences a unique chemical environment
and thus a separate spectral peak. However, only the I3C atoms, but not the naturally abundant
12C atoms, are detectable by MRS.
As exogenously supplied I3C atoms are incorporated into the TCA cycle (Fig. I),
individual carbon atoms of each intermediate molecule may become labeled. The position
of the label depends on the labeling pattern of the compound entering the TCA cycle and the
isotopic mixing that occurs with multiple turns of the cycle. Although labeling occurs in all
TCA cycle intermediates, the incorporation of label into a-ketoglutarate is most relevant
because a-ketoglutarate is in chemical equilibrium with glutamate and the equilibrium
conditions are such that the normal concentration of glutamate is much greater than
a-ketoglutarate, or any other TCA cycle intermediate, making it easily visible in a 13C
spectrum. Thus, the labeling pattern of glutamate (which is the same as a-ketoglutarate)
provides all the necessary information about entry of l3C labeled substrate into the TCA
cycle. When the C4 of glutamate has been labeled with I3C, the presence or absence of an
adjacent I3C atom (in C3 and/or C5) alters its chemical environment and gives rise to a unique
signal (Fig. 2). The different combinations oflabel in C3, C4, and C5 combine to produce
a spectrum of9 different peaks, which can be analyzed to deduce the fractional contributions
of the different label combinations (Fig. 3).

....- Citrate « - - Acetyl-CoA

a-Ketoglutarate
oII
.......... SCoA - C - CHa
"" 1 2
Figure 2. The isotopomer combina- Glutamate 0 •
tions resulting from the incorporation HaN HH HH H ••
of [3- l3 C]lactate and [l,2- l3C]acetate \ / \/ \/
-OOC - C - C - C - coo- Glutamate C4
into glutamate. Solid circles indicate Peaks
o.
12345
label which is detectable in the spectral /0 0 0 A
region of the glutamate C4. Label in-
corporation results in the nine line C4
lac GIu1amate 0 0 • • 0 A A
glutamate spectrum from the four dif- lsotopomers 0 0 0 • • A A
ferent isotopomers. '\.0 0 • • • A " " "
 214 L. A. Bertocci

PPM
Figure 3. A representative spectrum from the C4 of glutamate. The abscissa, in units of parts per million (ppm)
hertz, describes the chemical shift differences in resonance frequencies between the individual peaks. The
peaks arising from incorporation of [3- !3C)lactate are shaded; those from [1,2- !3C)acetate are hatched.

By way of example, the spectrum in Fig. 3 is the 9 line resonance arising from the
C4 of glutamate after incorporation of [I ,2- 13 C]acetate and [3 -l3C]lactate. As a consequence
of the label mixing, the 3 peaks arising from the incorporation of [3-l3C]I~ctate can be
distinguished from the 6 peaks arising from incorporation of [1,2- 13 C]acetate. Analysis of
one such spectrum provides information about 1) the relative contribution to the TCA cycle
of [3- l3 C]lactate compared with [1 ,2- l3 C]acetate and unlabeled sources; 2) the relative
contribution of [3- l3 C]lactate to anaplerosis; and 3) the relative rate of anaplerosis to TCA
cycle flux. Comparison of the total area under the peaks arising from the C2, C3, and C4 of
glutamate provides information about how close the system was to metabolic steady state
(Malloy et aI., 1987; Malloy et aI., 1990).
Although this technique has been used primarily for the study of metabolic regulation
in the heart and liver, the technique appears to have potential for the study of skeletal muscle
fatigue. For example, the technique could be used to examine the pattern of substrate
utilization as the muscle makes a transition between rest and exercise, particularly as the
exercise leads to fatigue. It has been known for many years that skeletal muscle alters its
relative utilization of carbohydrates vs. fats as it goes from rest to exercise, that this alteration
is extreme during fatiguing exercise, and that the capacity to perform maximal exercise is
attenuated when certain fuels are unavailable (Hedman, 1957; Gollnick et aI., 1981; Brooks
& Mercier, 1994). Therefore, one could use a l3C MRS protocol to study the relationship
between skeletal muscle fatigue and the flux through one or more contributing metabolic
pathways.
It is well known that skeletal muscle uses fat, mostly in the form offatty acids in the
blood, as its primary fuel during rest. The energy it makes from fat comes from the process
of p-oxidation in which a long chain fatty acid is broken down into 2-carbon units, and then
completely oxidized to CO 2, H20, and chemical energy. During exercise, the fraction of
energy derived from carbohydrate oxidation increases. Although there are several different
carbohydrates that produce pyruvate, most pyruvate is produced from glycogenolysis.
Oxidation of the pyruvate so derived occurs following decarboxylation and activation to
acetyl-CoA where is then has the same fate as acetyl-CoA derived from fat.
With the preceding as the biochemical basis for an experiment, how could one use
l3C MRS to measure the differences in the pattern of substrate utilization between rest and
exercise or between non-fatiguing and fatiguing exercise? This could be done, for example
by studying the depletion of endogenous glycogen and the subsequent repletion with
[I ,6-l3C]glucose. Glycogenolysis of the labeled glucosyl units will produce [3_ 13 C]pyruvate
Emerging Opportunities with NMR 215

Glycogen
[1 ,6- 13Cjglucosyl units

Fat
glycolysis [U- l 3Cjpalmitate

p-oxidation
[3- l 3Cjpyruvate

[13C MRS analYSiS)

~
L ~jacetyl-coA
[2_13C]OAA

( Cit~ic
I~
[1,2- 13Cjacetyl-CoA

t Acid [13Cjcitrate
I Cycle )

[13Cjglutamate ~ [13Cj(X_~

Figure 4. A specific substrate competition experiment where J3C label is introduced as glycogen preloaded
with [1,6- 13 C]glucosyl units and as exogenous [U- 13C]palmitate. This results in label entering the citric acid
cycle non-oxidatively as [2- 13 C]oxaloacetate, oxidatively via pyruvate as [2-13C]acetyl-CoA, or oxidatively
via p-oxidation as [1,2- 13C]acetyl-CoA.

and, following oxidative decarboxylation by the pyruvate dehydrogenase complex, [2-

13C]acetyl-CoA. Incorporation of [2- 13 C]acetyl-CoA into the citric acid cycle via oxidative
pathways would label the C3 and C4 of glutamate. Similarly, two carbon degradation of the
labeled palmitate and subsequent activation will produce [1,2- 13 C]acetyl-CoA. Incorpora-
tion of [1 ,2-13C]acetyl-CoA into the citric acid cycle via oxidative pathways would label the
C3, C4, and C5 of glutamate. By examination of the labeling pattern in the glutamate C4
alone, one can compare the relative rates of incorporation of labeled glycogen, palmitate,
and unlabeled sources. By examining the patterns and the relative amount of labeling in the
C2, C3, and C4 of glutamate, one can also determine the relative contributions of oxidative
vs. non-oxidative (anaplerotic) entry into the citric acid cycle if the system is in isotopic
steady state. Thus, one can determine if there is a change in the pattern of substrate utilization
during fatiguing exercise; or conversely, by altering, and monitoring, the relative availability
of substrates, one can determine the effects on exercise capacity or the tendency for fatigue.

23Na and 39K MRS

The ability of any cell to function properly depends on the maintenance, of its
intracellular ionic composition. For nerve and muscle cells, the maintenance of basal
trans-membrane sodium (Na+) and potassium (K+) gradients is the basis for the initiation
and propagation of action potentials. Each muscle cell contraction is initiated by an action
potential, which results in a net influx ofNa+, and efflux ofK+ across the cell membrane.
Although this ion flux is countered by the Na+/K+ ATPase, an electrogenic ATP-requiring
transmembrane protein which actively transports Na+ out of, and K+ into, the cell against a
bulk concentration gradient, extended periods of exercise result in net increases in intracel-
lular Na+ (Sreter, 1963; Blum et aI., 1988; Lindinger & Heigenhauser, 1988) and extracellular
K+ (Sreter 1963; Ahlborg et aI., 1967; Saltin et aI., 1981; Clausen, 1990; Lindinger &
216 L. A. Bertocci

Heigenhauser, 1991). To the extent that a continuous maintenance of the normal distribution
of[Na+]out and [K+]in depends on the Na+/K+ ATPase, and thus on the energetic state of the
muscle, it would be valuable to be able to measure the relative amounts of these ions in the
intracellular vs extracellular space in vivo during fatiguing exercise. Of particular interest
would be the ability to relate the changes in the steady state distribution ofNa+ or K+ in the
presence of the measurable changes in cellular energetics as seen using 3 1p MRS.
Let us first consider the measurement of sodium by MRS. The sodium in muscle
tissue has several properties that are attractive to the MR scientist attempting to measure
tissue sodium. First, 23Na has a very high natural abundance, such that nearly 100% of all
naturally occurring sodium exists as 23Na isotope (Friebolin, 1993). Second, the sensitivity
of the 23Na nucleus is approximately 9.25 x 10-2 relative to lH for equal numbers of nuclei
and at constant field (Friebolin, 1993). This makes it the second most NMR-sensitive nucleus
found in biological tissues. Third, sodium is present in muscle tissue in high concentration.
In typical muscle tissue, [Na+lout is ;:: 100mM, and [Na+]in is about 10 mM. Thus, the
combination of high natural abundance, high relative sensitivity, and high concentration
make the observation in vivo of biological sodium by magnetic resonance an attractive
prospect.
Unfortunately, it is only in relatively rare cases that it is of value to measure the total
amount of sodium in a tissue; often the values of interest are the individual values of
intracellular and extracellular sodium. The normal concentration of sodium in the extracel-
lular space is almost an order of magnitude greater than that in the intracellular space. In
contrast, the intracellular volume is usually a much larger fraction of the total space than is
the extracellular space. For example, skeletal muscle is about 95% intracellular space. Thus,
if one performs a simple 23Na MRS experiment on skeletal muscle, one will get a single
spectral peak arising from a combination of the >100mM concentration of [Na+]out in the 5%
extracellular volume co-resonating with the::; IOmM concentration of [Na+]in in the 95%
intracellular volume. To separate the signals, more sophisticated techniques must be applied
that take advantage of some of the properties of the 23Na+ nucleus in vivo. There are two
fundamental strategies used to cope with this problem of co-resonance: use of shift reagents
to move one resonance away from the other or use of a quantum filter to attenuate one of
the signals.
The general mechanism of action of a shift reagent is two fold 1) to cause a shift in
the resonance frequency, due in largest part to the paramagnetic component of the reagent;
and 2) to shorten the T I and/or T2 ofthe nucleus by providing another pathway for relaxation
of the perturbed nuclear spins. Depending on the nature of the shift reagent, either the
paramagnetic or relaxation effect will predominate. Although there are cases where a
shortening of relaxation times is of paramount importance, such as gadolinium-based shift
reagents in MRI, the paramagnetic effect is most important in detecting the transmembrane
distribution of the 23Na nucleus in vivo.
Shift reagents bind reversibly to the extracellular Na+ and cause their magnetic
resonance frequency to shift. Shift reagents are usually large inorganic molecules that are
combined with one or more strongly paramagnetic atoms that can interact with the ion of
interest, in this case 23Na+. The interaction of the 23Na+ and the paramagnetic atom in the
complex causes an alteration in the normal interaction between the 23Na+ and the external
magnetic field. If the paramagnetic atom complex is large and polar, it will not cross the
muscle cell membrane. Thus, only the extracellular 23Na+ moieties will interact with the shift
reagent, only resonances arising from the extracellular 23Na+ will be shifted, and one will
see two different spectral peaks in a 23Na MR spectrum.
Fig. 5 is an example of the use of the shift reagent Thulium 1,4,7,IO-tetraazacyclo-
dodecane-l,4,7,10-tetrakis(methylene phosphonate) (TmDOTP), a complex organic ligand
containing the lanthanide thulium as the paramagnetic agent. In this experiment, TmDOTP
Emerging Opportunities with NMR 217

Shifted

Unshifted

Figure 5. Examples of the effect of the shift reagent ~ Kidney

TmDOTP-5 on 23Na spectra collected from rat muscle,
heart, liver, brain, and kidney. In each case, the effect of ~ ~X5
the interaction between the TmDOTP-5 and the 23Na+ ions I Iii I I I I I i I I I I II I i I II I I iii iii I
is to move the resonance downfield by about 5 ppm 40 30 20 10 0 -10
(reproduced with permission from Bansal, 1994). ppm

was infused into a rat via a jugular catheter. After allowing time for the reagent to attain a
steady-state concentration in the body, 23N a MR spectra were collected from several different
tissues. Because the TmDOTP is too large and hydrophilic to cross a cell membrane, the
shifted peak arises from the ion in the vascular space and the unshifted peak from the ion in
the intracellular space.
The other method to differentiate the intracellular and extracellular Na+ is to take
advantage of the way in which sodium undergoes transitions between its nuclear spin states.
Sodium is a quadripolar nucleus, and can undergo nuclear transitions between its four
possible spin states of 312, 112, -112, and -3/2. These transitions can be made singly or in
double or triple quantum steps. By altering the RF pulse sequence imposed on the tissue, it
is possible to select only the double or triple step changes in spin state. In principle, such an
experiment, designed to detect only double or triple quantum transitions, selects only for an
individual pool of sodium atoms. Such a selection is useful only if an individual pool of
sodium represents a single biological, physiological, or anatomical pool. For example, it is
assumed that the intracellular pool of sodium is composed of N a+ which is bound to large
proteins or membranes or is contained in a viscous environment where multiple quantum
transitions would predominate. Conversely, it is assumed that the vascular pool of sodium
is composed ofNa+ which is unbound, where Na+ is allowed to tumble relatively freely, and
where single quantum transitions would predominate. Although the results of such experi-
ments are rarely unambiguous, and require further validation, comparisons between the
spectral distributions of sodium detected with a standard one-pulse experiment vs. those
detected using a multiple quantum filter might provide information about the distribution of
the ion across the membrane.
218 L. A. Bertocci

With the preceding as the biochemical basis for an experiment, how could one use
23Na MRS to measure the transmembrane distribution of Na+ in skeletal muscle, and in
particular during exercise? By means of illustration, let us consider a specific example in
which TmDOTP is infused to a steady state in the rat and the hindquarter muscles are
electrically stimulated using several different patterns of varying or increasing intensity. The
difference in the relative areas of the spectral peaks at these different stimulation rates would
then represent the steady-state distribution of sodium and would reflect the ability of the
Na+/K+ pump to keep up with the influx of Na+ and efflux of K+ mediated by the action
potentials associated with the stimulation protocol. Similarly, the same variables could be
measured using a mUltiple quantum filter experiment. Thus, one can determine if exercise
changes the pattern of ion movement as a function of any number of variables associated
with fatigue, including exercise intensity, exercise training, muscle fiber type, local blood
flow, or any other such perturbation, and all these measurements can be made in the intact
tissue.

IMAGING

One of the most revolutionary developments in the technology of modem medical

science is magnetic resonance imaging (MRl). It provides more detail and tissue-type
information than any other imaging modality available. In the text that follows, two aspects
ofMRl will be described. First, the use ofMRI to discriminate active from inactive muscle
and to use this information to make biophysical and physiological inferences about ion and
water movements. Second, the possibility of using cine MRI, currently used primarily in the
study of the beating myocardium, to study actively contracting skeletal muscle.

tHMRI

The hydrogen nucleus (also referred to as the hydrogen ion, H+, or proton) has the
largest relative sensitivity of any MR detectable nucleus. Its combination of large concen-
tration in living tissues, large gyro magnetic ratio, and large natural isotopic abundance, make
it the nucleus of choice for making MR images. Because the general principles of magnetic
resonance imaging can be learned from any number of excellent textbooks, it is not the intent
to provide a treatise on the details of MRI. However, it is necessary to be reminded that the
physical basis of the imaging process depends on the interaction between magnetic field
strength, the intrinsic resonance properties of the proton, and the interaction(s) between the
proton and its biophysical environment. Based on the potential for all these types of
interactions, one can produce MRIs as remarkably informative as they are by making the
proper adjustments to the collection techniques that take advantage of one or more of these
physical properties so that the anatomical feature of interest is appropriately emphasized.
One of the biophysical properties involved in MRl that can be exploited is relaxation,
and specifically the speed and mechanism of the relaxation of the nuclear spins that have
been perturbed during the process of acquiring an MRI. By altering the details of the MRl
acquisition process, one can create an image that selectively emphasizes protons undergoing
relaxation at specific rates or via specific mechanisms and to use these images to differentiate
different tissue types (such as fat vs. muscle or tissue vs. blood). Nuclear spins decay
according to two different relaxation mechanisms 1) spin lattice (also called longitudinal)
relaxation, described by the relaxation parameter T I, and 2) spin-spin (also called transverse)
relaxation, described by the relaxation parameter T2. The difference in T2 relaxation of water
Emerging Opportunities with NMR 219

protons in different tissue compartments is the basis for using MRI to examine muscle during
rest and exercise.
The use of such T2 weighted images is based on the likelihood that the relaxation
properties of the nuclear spins of water protons should depend on the interaction of the
protons and their environment. Although this idea generates little argument, the exact nature
of the mechanism at work is controversial: it has been argued that the increase in image
signal intensity results from water movement into the interstitial space because the transverse
relaxation (T2) time of extracellular, and in particular interstitial, protons should be longer
than that of intracellular protons (Archer et aI., 1992). The converse argument has also been
proposed: that the increase in image signal intensity results from water movement into the
intracellular space because the transverse relaxation (T2) time of intracellular protons should
be longer than that of extracellular protons (Fisher et aI., 1990). In either case, if there is a
decrease in the concentration of water in one space (with its own characteristic T 2), and an
accompanying increase in concentration of water in another space (with its own, and
different, T2), it will be reflected as a change in the signal intensity of the image pertaining
to one space or another. And more specifically, if changes in image signal intensity can be
associated with a known movement of water during exercise, the T2 characteristics of the
protons in the two different spaces can be used to explain the phenomenon in biophysical
terms.
One promising application of this general technique is the use ofMRI to selectively
detect exercised vs. non-exercised skeletal muscle (Fleckenstein et aI., 1988). Initially, this
was used to identify the spatial distribution of skeletal muscles recruited during forearm
exercise (Fleckenstein et aI., 1989). This identification has been particularly valuable in that
it has provided a means of performing routine 31 P MRS examinations of skeletal muscle with
confidence that the 31p MR spectra are being collected from actively exercising muscle
(Fleckenstein et aI., 1991; Bertocci et aI., 1993). But more importantly, it has provided a
means for determining the extent to which skeletal muscles are recruited during exercise as
well as some of the biophysical and physiological consequences of this contractile activity.
Another application is the use of MRI to examine the movement of water between
compartments in skeletal muscle during exercise. Cell membranes are very permeable to
water molecules and water moves across muscle cell membranes along an ionic concentra-
tion gradient so that osmotic balance is maintained. If there is a change in the distribution
of osmotically active moieties in one of the tissue compartments, water will move in parallel
so that osmotic balance is maintained. For example, if exercise causes glycogenolysis, and
glycolytic flux increases, the intracellular concentration of hexose and triose phosphates will
increase. Water will move into the cell to maintain osmotic balance, the fraction of intracel-
lular tissue water will increase, and the MRI signal intensity associated with intracellular
water will also increase. Conversely, when repeated action potentials cause a net loss of
intracellular osmolality due to extrusion of K+ greater than can be handled by the Na+JK+
ATPase, or when heavy exercise causes a net extrusion of lactate ions, the extracellular
concentration of ions will increase. Water will move out of the cell to maintain osmotic
balance, the fraction of extracellular tissue water will increase, and the MRI signal intensity
associated with extracellular water will also increase. Thus, the nature of the change in the
MRI signal intensity as a result of skeletal muscle exercise can be used to make biophysical
and physiological conclusions about the movement of ions into and out of skeletal muscle
cells.
This type of approach has been used extensively in recent years by several different
research groups, and a proportionality between exercise intensity and an increase in MRI
intensity has been demonstrated. However, some questions remain concerning the meaning
of this proportionality. In general, the question of the exact definition of exercise intensity,
and the real biophysical and physiological meaning of the increase in image intensity remain
220 L. A. Bertocci

to be resolved. The exercise protocols used in these studies varies from group to group, but
they all follow the same general pattern of graded rhythmic exercise. The results have been
similarly general: as the intensity (determined by various means) of the applied force is
increased, so does the image intensity. The increase in image intensity has been assumed to
be due to an increase in the concentration of interstitial (Archer et aI., 1992) or intracellular
(Fisher et aI., 1990) water.
Following the initial demonstration of this technique to discriminate regions of active
vs. inactive muscle (Fleckenstein, 1988), it was demonstrated that the increase in signal
intensity was proportional to exercise intensity (Fisher et aI., 1990). In this study, 1.5 min
bouts of dorsiflexion exercise were used in combination with volume changes caused by
venous occlusion. Most of the increase in image signal intensity was associated with
increasing force generation at a given rate of limb flexion. There was a similar result when
the force output was held constant but the limb flexion rate was increased from 1 to 2 to 3
Hz (Jenner et aI., 1994). Thus, the increase in signal intensity was proportional to power
output in the range observed. When the exercise was designed so that the power output was
relatively constant, but the duration was varied, it was observed that the increase in signal
intensity reached a plateau (Fleckenstein et aI., 1993).
Several hypotheses have been proposed to explain these results. One explanation is
the contributions from bulk water movement. This has been studied using 1) partial vascular
occlusion (external blood pressure cuff at diastolic pressure) to increase total limb water
(Fisher et aI., 1990); 2) total vascular occlusion (external blood pressure cuff at suprasystolic
pressure) to allow water movement between compartments but to prevent any change in total
limb water (Archer et aI., 1992); and 3) hyperemic reflow (Archer et aI., 1992). These results
are generally consistent with a small, but detectable, contribution of bulk water to the
increase in image signal intensity.
Another explanation is the movement of water out of the vascular space and into the
interstitial or intracellular space as a result of changes in osmotic gradients. This has been
studied using 1) exercise routines which result in differing amounts of glycolytic flux, resulting
in differing amounts ofaccumulation of hexose and triose phosphates (Fleckenstein et al., 1993,
Jenner, 1994); and 2) exercise in patients with metabolic disorders which preclude the
accumulation oflactate (Fleckenstein et aI., 1991). These results are consistent with osmotic
drive being the primary contribution to exercise-induced increases in signal intensity.
Where does this lead in the future? Apart from the obvious use of 1H MRI to examine
the biophysical and physiological basis of water movement in muscle tissue, one can imagine
several other possible uses for this technique. One is the use of MRI to determine skeletal
muscle fiber type non-invasively. Although skeletal muscle fiber type has traditionally been
determined histologically from a tissue biopsy, the ability to use MR to estimate this
non-invasively is very attractive. Such efforts have already been employed using MRS
(Kushmerick et aI., 1992, Blei et aI., 1993, Mizuno et aI., 1994). It would be good to be able
to take advantage of the superior spatial and temporal resolution of the proton by using lH
MRI to make such measurements. Additionally, it may be possible to use MRI to study motor
recruitment patterns and the regulation of motor control. MRI has already been used to
demonstrate the difference between voluntary exercise and cutaneous evoked contractions
(Adams et aI., 1993). Such studies can supplement other measures of muscle fiber recruit-
ment, such as EMG. Given the superior MR sensitivity ofthe hydrogen nucleus, it is likely
that other uses for this technology will be found.

Cine
Cine MRI is a relatively recent technical development which owes its existence in
part to the continuing efforts of MR physicists, and in part to the development of computers
Emerging Opportunities with NMR 221

with sufficient computational ability to maintain continuous updates of the image informa-
tion in near-real time. So far, this has been applied to dynamic internal events such as the
beating heart. Relevant to the study of neuromuscular fatigue, the ability to use MRI to
monitor anatomical events in real time may someday provide the ability to study fatiguing
muscle in situ. Although no such work has been done in skeletal muscle, cine MRI has
provided an important new method for viewing the anatomy and mechanics of the beating
myocardium.
As examples of its power, cine MRI has been used to 1) make a precise geometrical
analysis of the ventricular wall (Sacks et ai., 1993); 2) examine left ventricular dysfunction
(Dell 'Italia et ai., 1993); and 3) perform wall motion analysis (Matheihssen et ai., 1993).
These examples indicate that cine MRI can act as a live real-time imaging modality that can
be used to accurately assess the mechanics and the anatomy of moving tissue. To the extent
that skeletal muscle is a moving tissue, the same principles could be applied to the study of
muscle. It is relatively easy to imagine skeletal muscle movements being observed in real
time and correlated to their related neurological or biochemical regulation, particularly as
muscle undergoes fatigue. All that remains is the actual design and conduct of the experi-
ments.

SUMMARY

In this brief text, two types of MRS and two types of MRI have been discussed in
the context of their possible future applications to the study of neuromuscular fatigue.
Although the metabolic aspects of l3C and 23Na MRS make them more likely to produce
important information in the near future, it is the high relative sensitivity of the proton which
makes the possible applications of IH MRI more likely to produce important information in
the more distant future. In particular, when the biophysics of water movement and relaxation
patterns is resolved, T I and T2 weighted imaging experiments may result in fundamental
advances in our understanding of the dynamics of water and ion movements. Also, use of
cine MR to obtain real-time views of actively contracting muscle may provide similarly
fundamental advances in our understanding of the processes which regulate the contractile
process. The future is as bright as our imaginations.

ACKNOWLEDGMENTS

The author's work and laboratory is supported by the National Aeronautics and Space
Administration (NASA; grant NAGW 3582), United States Public Health Service grants RR
02584, HL 06296 and HL 07360, the Department of Radiology, University of Texas (UT)
Southwestern Medical Center, and the Institute for Exercise and Environmental Medicine,
Presbyterian Hospital of Dallas. Attendance of the author at the 1994 Bigland-Ritchie
conference was supported, in part, by the Department of Radiology, UT Southwestern
Medical Center, the Institute for Exercise and Environmental Medicine, Presbyterian Hos-
pital of Dallas, and the University of Miami.

REFERENCES

Adams GR, Harris RT, Woodward D & Dudley GA (1993). Mapping of electrical muscle stimulation using
MRI. Journal ofApplied Physiology 74, 532-537.
222 L. A. Bertocci

Ahlborg B, Bergstrom J, Ekelund LG & Hultman E (1967). Muscle glycogen and muscle electrolytes during
prolonged physical exercise. Acta Physiologica Scandinavica 70, 129-142.
Archer BT, Fleckenstein JL, Bertocci LA, Haller RG, Barker B, Parkey RW & Peshock RM (1992). Effect of
perfusion on exercised muscle: MR imaging evaluation. Journal ofMagnetic Resonance Imaging 2,
407-413.
Bansal N (1994). In vivo 23Na MRS with a shift reagent. In: Weatherall PT, Deslauriers R (eds.), MR Pulse,
Newsletter ofthe Society ofMagnetic Resonance vol. 1, pp. 18-19. Berkeley, CA: Society of Magnetic
Resonance.
Bertocci LA, Lewis SF & Haller RG (1993). Lactate infusion in human muscle PFK deficiency. Journal of
Applied Physiology 74, 1342-1347.
Blei ML, Conley KE & Kushmerick MJ (1993). Separate measures of ATP utilization and recovery in human
skeletal muscle. Journal of Physiology (London) 465, 203-222.
Bloch F, Hansen W & Packard M (1946). Nuclear induction. Physics Reviews 69, 127.
Blum H, Schnall MD, Chance B & Buzby GP (1988). Intracellular sodium flux and high-energy phosphorus
metabolites in ischemic skeletal muscle. American Journal of Physiology 255, C377-C384.
Brooks GA & Mercier J (1994). Balance of carbohydrate and lipid utilization during exercise: the "crossover"
concept. Journal of Applied Physiology 76, 2253-2261.
Chance EM, Seeholzer SH, Kobayashi K & Williamson JR (1983). Mathematical analysis of isotope labeling
in the citric acid cycle with applications to I3C NMR studies in perfused rat hearts. Journal of
Biological Chemistry 258, 13785-13794.
Clausen T (1990). Significance ofNa+ -K+ pump regulation in skeletal muscle. News in Physiological Sciences
5,148-151.
Cohen SM (1983). Simultaneous I3C and 31p NMR studies of perfused rat liver. Journal of Biological
Chemistry 258, 14294-14308.
Dell 'Italia LJ, Blackwell GG, Urthaler F & Pearce DJ (1993). A stable model ofleft ventricular dysfunction
in an intact animal assessed with high fidelity pressure and cinemagnetic resonance imaging.
Cardiovascular Research 27, 974-979.
Fisher MJ, Meyer RA, Adams GR, Foley JM & Potchen EJ (1990). Direct relationship between proton T2 and
exercise intensity in skeletal muscle MR images. Investigative Radiology 25, 480-485.
Fleckenstein JL, Bertocci LA, Nunnally RL & Peshock RM (1989). Exercise-enhanced MR imaging of
variations in forearm muscle anatomy and use: importance in MR spectroscopy. American Journal
of Roentgenology 153, 693-698.
Fleckenstein JL, Canby RC, Parkey RW & Peshock RM (1988). Acute effects of exercise on MR imaging of
skeletal muscle in normal volunteers. American Journal of Radiology 151, 231-237.
Fleckenstein JL, Haller RG, Lewis SF, Archer BT, Barker BR, Payne J, Parkey RW & Peshock RM (1991).
Absence of exercise-induced MRI enhancement of skeletal muscle in McArdle's disease. Journal of
Applied Physiology 71, 961-969.
Fleckenstein JL, Watamull D, McIntire DD, Bertocci LA, Chason DP & Peshock RM (1993). Muscle proton
T 2 relaxation times and work during repetitive maximal voluntary exercise. Journal of Applied
Physiology 74, 2855-2859.
Fleckenstein JL, Weatherall PT, Bertocci LA, Ezaki M, Haller RG, Greenlee R, Bryan WW & Peshock RM
(1991). Locomotor system assessment by muscle magnetic resonance imaging. Magnetic Resonance
Quarterly 7,79-103.
Friebolin H (1993). Basic One- and Two-Dimensional NMR Spectroscopy. New York, NY: VCH Publishers.
Fushiki T, Kano T, Inoue K & Sugimoto E (1991). Decrease in muscle glucose transporter number in chronic
physical inactivity in rats. American Journal of Physiology 260, E403-E410.
Gollnick PD, Pemow B, Essen B, Jansson E & Saltin B (1981). Availability of glycogen and plasma FFA for
substrate utilization in leg muscle of man during exercise. Clinical Physiology 1, 27-42.
Hedman R (1957). The available glycogen in man and the connection between rate of oxygen intake and
carbohydrate usage. Acta Physiologica Scandinavica 40, 305-321.
Jeffrey FMH, Rajagopal A, Malloy CR & Sherry AD (1991). 13C-NMR: a simple yet comprehensive method
for analysis of intermediary metabolism. Trends in Biological Science 16, 5-10.
Jenner G, Foley JM, Cooper TG, Potchen EJ & Meyer RA (1994). Changes in magnetic resonance images of
muscle depend on exercise intensity and duration, not work. Journal of Applied Physiology 76,
2119-2124.
Katz J (1985). Determination of gluconeogenesis in vivo with 14C labeled substrates. American Journal of
PhYSiology 248, R391-R399.
Katz J & Lee PWN (1991). Application of mass isotopomer analysis for determination of pathways of glycogen
synthesis. American Journal of Physiology 261, E332-E336.
Emerging Opportunities with NMR 223

Kushmerick MJ, Moerland TS & Wiseman RW (1992). Mammalian skeletal muscle fibers distinguished by
contents of phosphocreatine, ATP, and Pi. Proceedings of the National Academy of Sciences 89,
7521-7525.
Lee PWN, Somo S & Bergner E (1991). Glucose isotope, carbon recycling, and gluconeogenesis using
[U-l3C]glucose and mass isotopomer analysis. Biochemical Medicine and Metabolic Biology 45,
298-309.
Lindinger MI & Heigenhauser GJF (1988). Ion fluxes during tetanic stimulation in isolated perfused rat
hindlimb. American Journal ofPhysiology 254, RI17-RI26.
Lindinger MI & Heigenhauser GJF (1991). The role of ion fluxes in skeletal muscle fatigue. Canadian Journal
ofPhysiology and Pharmacology 69, 246-253.
Malloy CR, Sherry AD & Jeffrey FMH (1987). Carbon flux through citric acid cycle pathways in perfused
heart by l3C NMR spectroscopy. FEBS Letters 212, 58-62.
Malloy CR, Sherry AD & Jeffrey FMH (1990). Analysis of tricarboxylic acid cycle of the heart using l3C
isotope isomers. American Journal of Physiology 259, H987-H995.
Matheihssen NA, de Roos A, Doornbos J, Reiber JR, Waldman GJ & van der Wall EE (1993). Left ventricular
wall motion analysis in patients with acute myocardial infarction using magnetic resonance imaging.
Magnetic Resonance Imaging 11, 485-492.
Mizuno M, Secher NH & QuistorffB (1994). 31p_NMR spectroscopy, rsEMG, and histochemical fiber types
of human wrist flexor muscles. Journal ofApplied Physiology 76, 531-538.
Purcell E, Torrey H & Powel R (1946). Resonance absorption by nuclear magnetic moments in solids. Physics
Reviews 69, 37-38.
Sacks MS, Chuong CJ, Templeton GH & Peshock RM (1993). In vivo 3-D reconstruction and geometric
characterization of the right ventricular free wall. Annals ofBiomedical Engineering 21, 263-275.
Saltin B, Sjegaard G, Gaffney FA & Rowell LB (1981). Potassium, lactate, and water fluxes in human
quadriceps muscle during static contractions. Circulation Research (Supplement I) 48, 18-24.
Sreter FA (1963). Cell water, sodium, and potassium in stimulated red and white mammalian muscles.
American Journal of Physiology 205, 1295-1298.
Walker TE, Han CH, Kollman VH, London RE & MatwiyoffNA (1982). l3C nuclear magnetic resonance
studies of the biosynthesis by microbacterium ammoniaphilum of L-glutamate selectively enriched
with carbon-l3. Journal ofBiological Chemistry 257,1189-1195.
SECTION V
The Case for Segmental Motor Mechanisms

Neural mechanisms at the spinal level are explored in this section. The dominant
question is to determine if the control system for motoneuronal discharge is intrinsically
wise. If so, what are the segmental constraints on its actions, and what peripheral inputs
contribute to its operation?
First, in Chapter 16, Binder-Macleod addresses the potential role in fatigue of muscle
wisdom and the catch-like property. For muscle wisdom, evidence is presented that during
voluntary, fatiguing contractions subtle slowing in the discharge frequency of motoneurons
optimizes force and thus minimizes fatigue. An important distinction is made between the
force optimization observed during voluntary vs. imposedcontractions (by electrical stimu-
lation). However, because some patterns of motor unit discharge in exercise do not lead to
the expected slowing of muscle relaxation times, alternative strategies may also be consid-
ered wise. The catch-like property is an example of how a subtle change in the pattern of
imposed electrical stimulation can produce more force in both control and fatiguing contrac-
tions. Establishment of the boundary conditions for such stimulation patterns and the extent
to which catch-like effects are observable in voluntary contractions are issues of current
interest. However, imposed stimulation is an end in itself when designing patterns of
functional electrical stimulation to reduce the fatigue of patients with motor impairments.
Windhorst and Boorman (Chapter 17) review the segmental machinery for reflexly
changing motor unit behavior. They point out that little is known about the strategies used
by the eNS to mitigate against the effects of fatigue, although there is evidence that fatigue
may disturb the normal patterns of motor unit recruitment. Potentiation of motor unit force
and feedforward and feedback signals acting on premotoneuronal networks can optimize
force. As well as the descending inputs to motoneurons, there are many interneurons carrying
information from low- and high-threshold mechanoreceptors and chemoreceptors in the
muscles. Similarly, inputs from Golgi tendon organs and recurrent inhibition seem likely to
be important in the moment-to-moment regulation of motoneuron discharge rate, although
their individual effects are not simple to differentiate in vivo. Furthermore, the effects of
presynaptic inhibition are strong and likely to be distributed differently to modulate the reflex
effectiveness of inputs from spindle group Ia and II afferents and tendon organ (group Ib)
afferents. Ultimately, it will help to have more estimates of the net behavior and gain of the
various reflex paths during muscle fatigue and information on the interactions between
segmental mechanisms and pattern generators for locomotion.
The fusimotor system activates muscle spindle afferents during voluntary contrac-
tions under relatively isometric conditions. Hence, it could modulate the homonymous
excitation provided to motoneurons by muscle spindle afferents. Removal of the reflex
226 The Case for Segmental Motor Mechanisms

facilitation (by partial or complete nerve blocks) reduces the maximal firing rates of human
motoneurons. In Chapter 18, Hagbarth and Macefield explore the evidence that this
disfacilitation ofmotoneurons abolishes the progressive decline in motoneuronal firing rate
during maximal voluntary isometric contractions. A surprising conclusion from any analysis
of the role of the fusimotor system in human muscle fatigue is that it fails to act as a
sufficiently powerful servo to offset the effects of fatigue in the extrafusal muscle fibers.
Such a role had previously been suggested as one of the prime functions of this system. In
the initial seconds of a strong voluntary contraction, the declining muscle spindle input and
increasing presynaptic inhibition will contribute to the early decline in firing rate of
motoneurons.
In Chapter 19 (Garland & Kaufman), the properties of small-diameter afferents
innervating muscle are reviewed and various lines of evidence are presented for a reflex
inhibition of motoneurons during fatiguing isometric exercise. There is no consensus on the
extent to which a reduction in discharge rate during maximal voluntary contractions is due
to reflex inhibition, or the exact spinal and supraspinal sites involved in the inhibition.
Indirect evidence favors a role for recurrent inhibition. It would be helpful to know how
motoneurons discharge would be affected if descending supraspinal drives were stimulated
maximally; or, alternatively, how motoneurons would behave in the absence of feedback
from of small-diameter afferents. Unfortunately, the exact effects of muscle fatigue on the
discharge of unmyelinated (group IV) muscle afferents are not yet available, but it is likely
that those sensitive to metabolic factors are activated later in a strong sustained voluntary
contraction.
Importantly, this section reveals areas in which data from human and animal experi-
ments are either discrepant or inconclusive. Further work could well focus on these
discrepancies along with the changes in afferent properties with fatigue, the central (particu-
larly segmental) effects of these changes, and the extent to which supraspinal and other
descending drives can modify the segmental behavior.
 16

VARIABLE-FREQUENCY STIMULATION
PATTERNS FOR THE OPTIMIZATION OF
FORCE DURING MUSCLE FATIGUE
Muscle Wisdom and the Catch-Like Property

s. A. Binder-Macleod
Department of Physical Therapy
University of Delaware
Newark, Delaware 19716

ABSTRACT

Muscle wisdom is the process whereby the activation rates of motor units are modulated by
the central nervous system to optimize the force during sustained voluntary contractions.
During maximal voluntary contractions the activation rates decline as the muscle fatigues.
No similar decline has been observed during submaximal contractions. Subsequent chapters
explore the potential mechanisms for muscle wisdom. In this chapter, a historical background
on the development of ideas on muscle wisdom is first presented. Next, artificial wisdom,
the procedure used to optimize force during an electrically imposed tetanus by progressively
reducing the stimulation frequency as the muscle fatigues, is discussed. Finally, recent
studies are described in which fatigue was delayed and reduced by the use of variable-fre-
quency stimulus trains that elicit the catch-like property of muscle.

INTRODUCTION

It has long been known that the activation frequency of skeletal muscle affects the
generation, maintenance, and decline (fatigue) of force. In the control (unfatigued) state,
there is a sigmoidal relationship between the stimulation frequency and the peak force
produced by both whole muscles (Cooper & Eccles, 1930) and single motor units (Kernell
et aI., 1983). Low frequencies produce unfused twitches; intermediate frequencies produce
a steep, near-linear relationship between force and frequency, and relatively high frequencies
are needed to produce maximal forces (see also Kernell, Chapter 9). In contrast to high-fre-
quency stimulation, low-frequency stimulation produces relatively less fatigue during sus-
tained or intermittent contractions (Bigland-Ritchie et aI., 1979; Jones et aI., 1979;
Binder-Macleod et aI., 1995). Interestingly, the discharge rate of a motor unit during
voluntary contractions exhibits short-term (i.e., pulse to pulse) and long-term variations (Le.,

227
228 s. A. Binder-Macleod
changes over seconds to minutes) (Marsden etai., 1971; Bigland-Ritchie etai., 1983a; Maton
& Gamet, 1989). Variations in the motor unit activation rate have been shown to maximize
force (Burke et ai., 1976; Binder-Macleod & Clamann, 1989; Bevan et ai., 1992) and
minimize fatigue (Marsden et ai., 1976, Jones et ai., 1979; Binder-Macleod & Barker, 1983;
1991) during imposed electrical activation of muscle. The remainder of this chapter explores
the ways in which the modulation in activation rate of motor units can be used to maximize
force during both voluntary and imposed contractions.

HISTORICAL BACKGROUND ON MUSCLE WISDOM

At the 1968 meeting of the Physiological Society, Marsden, Merton and Meadows
coined the term muscle wisdom. This term describes the process whereby the activation rates
of the motor units are modulated by the central nervous system (CNS) to optimize the force
during sustained contractions (Marsden et ai., 1976, 1983; for review, see: Enoka & Stuart,
1992; Stuart & Callister, 1993). The experimental evidence for their conclusions can be
directly linked to a series of experiments conducted by these investigators, and their
contemporaries, studying both imposed electrical activation of the human skeletal muscle
and the activation pattern of motor units during fatiguing voluntary contractions. In 1954
Merton showed that, for the human adductor pollicis muscle, the force decline during a
sustained maximal voluntary contraction was due to failure within the muscle (cf., Bigland-
Ritchie et aI., 1978). Merton (1954) also showed for the first time that voluntary contractions
were able to produce comparable forces to those produced by maximal tetani. These were
important findings because they suggested that decreased central drive was not the cause of
fatigue, but rather, that the CNS was able to maximally activate muscle. The following year
Ncess and Storm-Mathisen (1955), also studying the human adductor pollicis muscle, showed
that a stimulation train with a pulse frequency that elicited a near-maximal tetanic contraction
(50 pps) produced a greater rate of decline in force than a maximal voluntary contraction.
Ncess and Storm-Mathisen (1955) also showed that higher rates of stimulation (200 pps),
while producing slightly greater initial forces, were less effective at maintaining force. In
1971, Marsden and colleagues showed that the compound action potential discharge rate
(i.e., activation rate) of motor units during prolonged, maximal, voluntary contractions
declined over time, beginning at -60-100 pps initially and slowing to -20 pps by 30 s.
In 1976 Marsden, Merton and Meadows reported that an artificial wisdom procedure,
where the frequency used to elicit a tetanus is progressively reduced as the muscle fatigues,
produced a force comparable to that obtained during sustained, voluntary, maximal contrac-
tions. They also noted that the fatigue experienced under ischemic conditions was related to
the number of motor impulses, rather than the duration or force of contraction. Marsden and
colleagues (1976) concluded that, based on their observations and those of Ncess and
Storm-Mathisen, to obtain the "most prolonged and powerful voluntary contractions the
activation rate of the motor units should slow down consistent with maintaining a fully fused
tetanus as the contraction proceeds." Dietz (1978), also an early proponent of the notion that
the decline in motor unit activation rate during a maximal voluntary contraction actually
helped to optimize force output (i.e., muscle wisdom) suggested that the decline of discharge
frequency "took an optimal course because higher as well as lower discharge rates would
reduce the force development."
Finally, in an often cited 1983 publication, Marsden and colleagues developed more
fully their theory of muscle wisdom. They argued that the force developed during a sustained
contraction depends on "a compromise between activation of tension generation and a
concomitant aggravation of activation failure." During the onset of a contraction high
activation rates are needed to produce maximal force; however, maintenance of these high
Variable-Frequency Stimulation Patterns 229

rates results in rapid fatigue. In addition, they noted that during a prolonged contraction the
"time scale of the contractile mechanism slows down, so that the rate of innervation needed
to activate it optimally falls progressively as the contraction proceeds." The authors argued
that activation failure produces the need for a decreased activation rate while contractile
slowing allows the activation rate to decline while still eliciting the best out of the muscle.
While most of the results reported by Marsden and colleagues have been supported
by more recent studies, caution must be advised in generalizing their findings (cf., Stuart &
Callister, 1993). It should be noted that most of their results were based on the study of just
one to three, highly trained subjects (i.e., the authors). Also, many of their observations were
made on motor units in adductor pollicis that received an aberrant innervation from the
median nerve (Marsden et aI., 1971). Furthermore, to allow the recording of the EMG of
single motor units during maximal efforts, the authors used chemical or ischemic blocking
of the ulnar nerve at the elbow to reduce the number of active motor units. Although these
conditions may have biased their results, the observations and conclusions drawn by Marsden
and colleagues have served as an important impetus for the study of muscle fatigue.

REDUCTION IN THE ACTIVATION RATE OF MOTOR UNITS

DURING VOLUNTARY CONTRACTIONS

Since the original work by Marsden and colleagues (1971), several investigators have
found similar declines in the activation rate of motor units during maximal voluntary
contractions. Bigland-Ritchie and colleagues (1983a) have reported that during a -60 s
isometric MVC of the human adductor pollicis brevis, the mean activation rate of motor
units fell by 50%, from -27 to -15 Hz. During this time, there was no decline in the muscle's
compound action potential (M-wave) elicited by a single-shock stimulus to the muscle nerve.
This suggested that the decline in motor unit activation rate was attributable to a correspond-
ing decline in motoneuron discharge rate rather than a progressive block of neuromuscular
conduction (also see Thomas et aI., 1989; cf. Bellemare & Garzaniti, 1988). Grimby and
colleagues (1981), who reported comparable declines in human short big toe extensor and
anterior tibialis muscles, have suggested that this progressive decline in motor unit activation
rate may protect peripheral neuromuscular propagation and hence elicit more force over time
than if continuous high frequency activation was maintained. Furthermore, Bigland-Ritchie
and colleagues (1981; 1983a; 1983b), and Woods and colleagues (1987) have argued that
the decline in motor unit activation rate may not only serve to reduce fatigue but may also
allow more effective modulation of the force of voluntary contractions by rate coding during
fatigue; that is, as the contractile (including force-relaxation) speed declines, the activation
frequency required for full tetanic force declines. If there was not a parallel decline in the
motoneuronal firing rate, the activation rate of motor units would become supra-tetanic, and
rate coding as a means of force modulation would be ineffective (Bigland-Ritchie, 1981; her
Fig. 6). Interestingly, during submaximal, fatiguing contractions the motor units activation
rate has been shown to increase (Bigland-Ritchie et aI., 1986; Maton & Gamet, 1989;
Dorfman et aI., 1990) or remain stable (Maton & Gamet, 1989).

A CAVEAT ON MUSCLE WISDOM: EFFECTS OF FATIGUE ON

THE NORMALIZED FORCE-STIMULUS FREQUENCY RELATION

Muscle wisdom proposes that force is optimized during fatiguing contractions. If

correct, such optimization should be reflected in the profile of the normalized relation
230 s. A. Binder-Macleod
between force and activation rate (T-jrelation). As reviewed in Sawczuk et aI., Chapter 8
and Kernell, Chapter 9, the T-frelation is subject to alteration by fatigue. For example, the
T-frelation is influenced by fatigue-induced contractile slowing, effects brought on by low-
vs. high-frequency stimulation, and motor unit type.

Contractile Slowing
Such slowing with fatigue is well documented (Fitts, 1994). It appears to be primarily
the result of an increase in force-relaxation time of the muscle (Bigland-Ritchie et aI., 1992).
The longer contraction time allows a greater summation of forces at subfusion stimulus
frequencies and produces a shift in the T-frelation to the left (for further explanation see
Sawczuk et aI., Chapter 8). That is, the steep portion of the T-j relation is shifted toward
lower rates. However, the assumption that fatigue must result in a concurrent shift in the T-f
relation toward lower frequencies has been challenged (for review see Enoka & Stuart, 1992;
Fitts, 1994); the confounding variables include the relative presence of low-frequency
fatigue and the motor-unit type distribution of the muscle.

Low-Frequency Fatigue
This term refers to a long-lasting (min to hr) selective force reduction in response to
low-frequency stimulation (e.g., ~ 20 Hz) (Edwards et aI., 1977; 1981). It is seen following
the fatigue brought on by voluntary and imposed contractions and is thought to be due to an
impairment in excitation-contraction coupling (Edwards et aI., 1977; Jones, 1981). Low-fre-
quency fatigue results in an attenuation of the twitch force, a shift in the steep portion of T-j
relation toward higher frequencies, and little change in the frequencies needed to produce a
fused contraction or a maximal tetanus.
Most studies on human muscle have observed significant low-frequency fatigue and
a shift in the T-frelation to higher stimulus rates (Edwards et aI., 1977; Cooper et aI., 1988;
Jones et ai., 1989; Stokes et ai., 1989; Binder-Macleod & McDermond, 1992; cf., Thomas
et aI., 1991). In contrast, in vitro work on the frog semitendinosus (Thompson et aI., 1992)
and rat diaphragm muscles (Metzger & Fitts, 1987) have shown much greater attenuation of
forces at the higher rates of stimulation (i.e., > 60 pps) than at the lower rates of stimulation
and a shift in the T-frelation to the left. These contrasting results can probably be explained
by differences in the experimental protocols used to produce the fatigue (e.g., the stimulation
frequency used to fatigue the muscle), the conditions of the preparation during the testing
(e.g., muscle temperature and blood supply), and the motor-unit- and fiber-type distribution
of the muscles.

Motor-Unit Type
Studies on fast-twitch single motor units in the cat (Powers & Binder, 1991) have
shown short-term potentiation in response to imposed low-frequency (20 Hz) stimulation of
type FR units immediately after fatiguing stimulation. This effect was not present in FI and
FF motor units. In addition, a longer-lasting depression in maximal tension and a delayed
onset (after ~ 30 min) depression of low-frequency responses (i.e., low-frequency fatigue)
was seen in all types of fast motor units. These changes resulted in a shift to the right in their
T-frelation, beginning ~ 30 min after the fatiguing protocol. In contrast, FR units first showed
a shift to the left in the T-frelation immediately after fatiguing protocol, and ~ 30 min later,
a shift to the right after fatiguing stimulation. Furthermore both Dubose and colleagues
(1987) and Gordon and colleagues (1990) have shown that type S motor units in cat hindlimb
muscles are far less susceptible than FF units to a fatigue-induced slowing in their force-re-
Variable-Frequency Stimulation Patterns 231

laxation times. These results, together with those of Powers and Binder (1991), infer that
muscle wisdom effects are more likely to be a feature of high- rather than low-force
contractions. This possibility is consistent with the finding that during fatiguing, submaximal
voluntary contractions, activation rates of human motor units have been shown to increase
(Bigland-Ritchie et aI., 1986; Maton & Gamet, 1989; Dorfman et aI., 1990) or remain stable
(Maton & Gamet, 1989). This increase in discharge rate, along with additional recruitment
(Fallentin et aI., 1993; Garland et aI., 1994) is needed to maintain the targeted force (generally
20%-30% of the maximal voluntary contraction) as the force generating ability of the
fatiguing muscle declines (Bigland-Ritchie et aI., 1986). On a functional level, any shifts in
the T-:frelation that occurred with fatigue were insufficient to alleviate the need for this
increase in motor unit activation rate.

BOUNDARY CONDITIONS FOR THE USE OF MUSCLE WISDOM

Though a number of studies have recently explored the mechanisms underlying

muscle wisdom (for review see Windhorst & Boorman, Chapter 17; Hagbarth & Macefield,
Chapter 18; and Garland & Kaufman, Chapter 19), as recently noted by Stuart and Callister
(1993), the boundary conditions of muscle wisdom remain relatively unexplored. To date,
all reports that have shown a decline in the motor unit activation rates have used sustained,
maximal, isometric contractions (Marsden et aI., 1971; Grimby et aI., 1981; Bigland-Ritchie
et aI., 1983a). Generally, no such decline in activation rate has been seen with submaximal
contractions (Bigland-Ritchie et aI., 1986; Maton & Gamet, 1989, Dorfman et aI., 1990;
Garland et aI., 1994). The relative amount of slowing in activation rate for the different types
of motor units has remained relatively unexplored (cf, Powers & Binder, 1991). Perhaps the
lack of decline in activation rate during submaximal contractions may be related to differ-
ences in the types of motor units recruited during maximal versus submaximal contractions.
As noted by Enoka and Stuart (1992), the muscle wisdom theme needs to be extended to
dynamic conditions to determine if the results observed during isometric conditions hold
true for a wide variety of natural movements. Similarly, the study of the changes in activation
rate during fatiguing intermittent and rhythmic contractions are also worthy of further study
(see Enoka & Stuart, 1992). Despite the need for further experiments to explore the boundary
conditions for muscle wisdom, the case is compelling for force optimization by variable-fre-
quency stimulation during evoked fatiguing contractions, particularly when the effects of
such stimulation bring on the catch-like property ofmusc1e.

REDUCTION IN THE ACTIVATION RATE OF MOTOR UNITS

DURING EVOKED CONTRACTIONS

Artificial Wisdom

In 1976, Marsden and colleagues coined this term to describe the procedure used to
optimize force during an electrically elicited tetanus by progressively reducing the stimula-
tion frequency as the muscle fatigues. Using this approach they, and others, have shown that
the forces produced by the imposed stimulation can approach those achieved during
sustained, voluntary, maximal contractions (see Fig. 1) (Marsden et aI., 1976, 1983; Jones
et aI., 1979). More recently, Binder-Macleod and Guerin (1990) showed that reducing the
stimulation frequency from 60 to 30 pps as the muscle fatigued evoked more force from the
human quadriceps femoris muscle during intermittent, submaximal contractions than that
232 s. A. Binder-Macleod

10
C,
.x
5
...
CD
()

If
0
0 45 90
Time (s)

Figure 1. Effects of artificial wisdom on the force output of human adductor pollicis muscle. Traces represent
the forces produced by a sustained voluntary contraction (a, bold, irregular line); evoked responses to 100 pps
(b), 50 pps (c), and 35 pps (d) constant-frequency trains; and the response evoked by a progressive reduction
in stimulation frequency (e). In e (artificial wisdom), stimuli were delivered at 60 pps for 8 s, 45 pps for 17 s,
30 pps for 15 s, and 20 pps for 55 s. At the onset of stimulation, the 100 pps train evoked the most force and
the 35 pps train the least. Throughout the contraction, the artificial wisdom procedure produced forces that
closely matched the voluntary contraction (modified from data presented by Marsden et aI., 1983; their Fig. 4).

achieved by maintaining a constant frequency of electrical stimulation (60 pps). To explain

the greater maintenance of force, Marsden and colleagues showed that the fatigue decline
was related to the number of motor impulses, rather than the duration or force of contraction.
More recently reports, have shown that the pattern of stimulation (e.g., the train frequency)
and the force of contraction both affect the rate of fatigue (Garland et aI., 1988; Binder-
Macleod et aI., 1995). Nevertheless, it appears clear that progressively reducing the stimu-
lation frequency as the muscle fatigues is more effective at maintaining the force than a
constant, high-rate of stimulation.

The Catch-Like Property of Skeletal Muscle

This term describes the force augmentation that occurs when an initial, brief,
high-frequency burst (containing -2-4 pulses) of stimuli is included at the beginning of an
imposed subtetanic train of pulses (Burke et aI., 1970, 1976; Binder-Macleod & C1amann,
1989). Burke and colleagues were the flrst to demonstrate a catch-like property in mammal-
ian motor units (Burke et aI., 1970, 1976; cf. Wilson & Larimer, 1968). The term catch-like
was used to distinguish this phenomenon from the true catch property of molluscan muscle
in which force production does not require continued stimulation of the muscle (Burke et
aI., 1976). A catch-like property has been observed in single motor units (Burke et aI., 1970,
1976; Zajac & Young, 1980a; Bevan et aI., 1992) as well as whole human (Binder-Macleod
& Barker, 1991) and animal muscles (Binder-Macleod & Barrish, 1992).
The catch-like property is a fundamental property ofthe muscle; the tension enhance-
ment is not due to the activation of additional muscle flbers (see Burke et aI., 1970; Bevan
et aI., 1992). Increased release of Ca2+ from the sarcoplasmic reticulum has been proposed
as a potential mechanism for the greater force (i.e., the sum of the forces produced by the
burst are greater than the sum of forces produced by an equivalent number of twitches)
(Duchateau & Hainaut, 1986b). This greater summation could explain the force augmenta-
tion seen with the catch-like property. However, as shown by Duchateau and Hainaut (1986a,
1986b), the presence of positive-staircase potentiation (i.e., a progressive increase in force
elicited in response to identical stimuli) can eliminate this greater summation. Although
positive-staircase potentiation reduces the augmentation associated with an imposed vari-
able-frequency stimulus train that exploits the catch-like property of muscle (i.e., an
optimized train) (Burke et aI., 1976; Bevan et aI., 1992), highly potentiated motor units
Variable-Frequency Stimulation Patterns 233

A 200
constant-frequency

150

~
II) 100
~
50

B 200

150

Figure 2. Characteristic evoked forces to

optimized and constant-frequency trains
from human quadriceps femoris muscle in
the control (nonfatigued) (A) and fatigued 50
(B) states. Both the optimized (C) and con-
stant-frequency (D) trains contained six
pulses. The optimized train contained an
initial interpulse interval of 5 ms and four 80 160 240 320 400 480 560 640
subsequent intervals =80 ms; the constant- Time (ms)
frequency train had all interpulse intervals
= 80 ms. Due to marked slowing in the rate
of tension development with the constant- C ~II__~~__~~___________
frequency train as the muscle fatigued and
the catch-like property of skeletal muscle,
the optimized train produced greater force
augmentation when the muscle was fa-
tigued than in the control state. D

(Burke et aI., 1976; Bevan et aI., 1992) and whole muscle (Binder-Macleod & Barker; 1991;
Binder-Macleod & Barrish, 1992) can still show significant augmentation. Thus, although
increased release of Ca2+ from the sarcoplasmic reticulum may contribute to the catch-like
property under certain circumstances (see Burke et aI., 1976), significant force augmentation
can be observed in the apparent absence of augmented Ca2+ release.
Increased muscle stiffness may also explain the force augmentation seen with the
catch-like property. Parmiggiani and Stein (1981) have likened the series elastic component
of the muscle to an elastic band at slack length. The slack in the series elastic component
must be taken up before force is produced. The stiffer the muscle the more substantial the
development of tension in response to each stimulus pulse. The initial, high-frequency burst
of the optimized train rapidly increases the force and muscle stiffness, thereby reducing the
slack in the series elasticity that will be encountered by subsequent stimuli (see Fig. 2). Thus,
the force augmentation due to the catch-like property may be explained largely by the ability
of the initial burst to increase muscle stiffness and reduce the time-dependent rate of tension
development (Cooper & Eccles, 1930; Binder-Macleod & Barrish, 1992).
234 s. A. Binder-Macleod
FACTORS AFFECTING THE AMOUNT OF FORCE
AUGMENTATION SEEN WITH CATCH-INDUCING TRAINS

Activation Pattern of the Optimized Train

To identify the pattern of interpulse intervals that evokes the greatest force-time
integral from single motor units or whole muscle most investigators have built pulse trains
one pulse at a time (Stein & Parmiggiani, 1979; Zajac & Young, 1980a; Parmiggiani & Stein,
1981; Duchateau & Hainaut, 1986a; Hennig & Lemo, 1987). Muscles were first stimulated
with two-pulse trains to determine the interpulse interval that produced the maximal force.
Next, three-pulse trains were tested. The first interpu1se interval was set at the duration that
was optimal for the two-pulse train and the duration of the third interpulse interval was
varied. This process was continued until a train with the desired number of optimized
interpulse intervals was completed. Surprisingly, although both fast and slow motor units
and muscles have been studied, most studies have shown that the optimal train contained
one or two short-duration (5-10 ms) intervals initially. In contrast, the optimal sequence after
these initial, short-durations intervals has been shown to vary as a function of the twitch
contraction time of the muscle (Burke et aI., 1976; Zajac & Young, 1980a). For cat medial
gastrocnemius motor units, these optimal interpulse intervals corresponded to a repetitive
train (i.e., all interval durations were equal) with their durations equal to -1.8 x twitch
contraction time of the muscle (Burke et aI., 1976; Zajac & Young, 1980a). Recently, we
have found for whole muscle that the constant-frequency portion of the optimized train had
durations of -1.0 x twitch contraction time for rat gastrocnemius muscle and 1.2 x twitch
contraction time for human quadriceps femoris muscle (see Figs. 3 and 4).

Motor Unit Type

Though not studied systematically, Burke and colleagues (1976) demonstrated that
slow-twitch motor units showed greater augmentation than fast-twitch units (Burke et aI.,
1976; their Figs. 6-7). Recently, we compared the forces produced with optimized and
constant-frequency trains in fast (gastrocnemius) and slow (soleus) rat muscles (Binder-
Macleod & Barrish, 1992; Binder-Macleod & Landis, 1994) and found results consistent
with those of Burke and colleagues. In the control condition (non-fatigued, highly potenti-
ated state) the soleus muscle produced a -17% greater force-time integral when stimulated
with the optimized trains than when stimulated with the constant-frequency trains (Binder-
Macleod & Barrish, 1992). In contrast, the gastrocnemius muscle showed no augmentation
when stimulated under similar conditions (Binder-Macleod & Landis, 1994) (see Fig. 3A).
However, as noted below, the gastrocnemius muscle did show augmentation with the
optimized train when the muscle was fatigued.

Activation History of the Motor Unit

Burke and colleagues (1976) showed that during isometric contractions of single
motor units in cat hindlimb muscles, the catch-like force augmentation was greater for
unpotentiated than potentiated units (potentiation is here defined as the increase in the
force due to prior activation; see also Kernell, Chapter 9). Bevan and colleagues (1992)
recently studied the effects of optimized stimulus trains on fast-twitch fatigable motor
units from cat tibialis posterior muscles during fatiguing contractions. The interpulse
intervals of the constant-frequency trains were 1.8 x the twitch contraction time of the
units, which resulted in a stimulus frequency range of 19-37 Hz. The optimized trains
Variable-Frequency Stimulation Patterns 235

-
...
-
...
A B
1.4 •• optimized 0.75 •• •• optimized
constant-frequency •• constant-frequency
0.7
••
*
0.65 **
0.6

0.55

0.9 -I----+---+---t---t-----1 0.5 +----t---t----+-----l----1

10 20 30 40 50 60 10 20 30 40 50 60
Stimulus interpulse interval (ms) Stimulus interpulse interval (ms)

Figure 3. Effects of optimized stimulus trains on the force output (force-time integral) of potentiated and
non-fatigued (A) and fatigued (8) rat gastrocnemius muscles. In this and Fig. 4, stimulus interpulse interval
(X-axis) refers to the constant-frequency stimuli of the full constant-frequency trains and the last four intervals
of the optimized trains. In the control, non-fatigued state, force was augmented when the interpulse interval
of the constant-frequency component of the optimized train was <: 2 x the twitch contraction time of the muscle
(<: 50 ms), which was not changed appreciably by fatigue. When the muscle was fatigued, force was augmented
for the optimized train when its later interpulse intervals were <: 25 ms. Comparisons were made at each
stimulus interpulse interval, using a paired t-test (* = p ~ 0.05; ** = P ~ 0.001).

began with a high-frequency burst consisting of three pulses with an interpulse interval
of 10 ms (100 Hz) followed by a constant interpulse-interval train that was adjusted so
the total train had the same number of pulses, duration, and average frequency as the
constant-frequency train for each muscle. They found that at the onset of stimulation,
(i.e., prior to fatigue and before the muscle was fully potentiated) the optimized trains
markedly increased the force-time integral produced by the muscle. After -30 s of
stimulation, when the muscle was maximally potentiated, the optimized trains produced
slightly less force than the constant-frequency trains. However, as the muscle began to
fatigue, the optimized trains again augmented the force compared with the constant-fre-
quency trains. Thus, it appears that fast-twitch fatigable motor units primarily show
augmentation with optimized trains when the muscle is either unpotentiated or fatigued.
We have observed that the effects of imposed optimized stimulus trains in human quad-
riceps femoris (contains -60% fast-twitch motor units) and whole rat gastrocnemius
(-100% fast-twitch motor units) muscles are similar to those reported by Bevan and
colleagues. In the non-fatigued, but highly potentiated state we observed little augmen-
tation for either muscle. However, as the muscles fatigued, the optimized trains were able
to significantly augment forces (Binder-Macleod & Barker, 1991; Binder-Macleod &
Baadte, 1992; Binder-Macleod & Landis, 1994).
Recently, we completed a series of studies that attempted to determine the effects on
the force of using various optimized trains, covering a wide range oftrain frequencies. These
studies on rat gastrocnemius and human quadriceps femoris muscles compared the force
produced by a variety of 6-pulse constant-frequency trains vs. optimized trains with the first
two pulses of the latter separated by a briefinterpulse interval (10 ms for rat muscle, and 5
ms for human muscle). The remaining four pulses within the optimized trains were separated
by interpulse intervals identical to those used in the constant-frequency trains. (see Figs. 3
and 4). Data were collected prior to fatigue, and during repetitive, fatiguing contractions (at
1 stimulus train/s). For rat gastrocnemius in the control (nonfatigued) state, the optimized
trains only augmented forces when the stimulus interpulse intervals of the last four intervals
in the train were 2: - 2 x the twitch contraction times of the muscles (i.e., 2: 50 ms; see Fig.
236 S. A. Binder-Macleod

A B
50
22 u*
,"",45 *** *** ***
21
'" ***
~40 20 ***

- -
'E ***
19
~ 35
.- .-
.S optimized 18 ** optimized
" 30
] 17 constant-frequency
§25
0
constant-frequency
16
~ 20 15
15 14~--+---~---r---+--~--~
0 W ~ W W 100 1W o 20 40 60 80 100 120
Stimulus interpulse interval (ms) Stimulus interpulse interval (ms)

Figure 4. Effects of optimized stimulus trains on the force output (force-time integral) of potentiated and
non-fatigued (A) and fatigued (B) human quadriceps femoris muscle. In the control, non-fatigued state, no
force augmentation was produced by the optimized trains at any of the frequencies tested (i.e., with their later,
constant-frequency intervals::; 1.5 x 80 rns, the twitch contraction time). When the muscle was fatigued, twitch
contraction times were reduced from - 80 ms to - 40 ms. Force augmentation was produced by the optimized
trains at all constant-frequency intervals <': 20 ms. The comparison is as in Fig. 3. (* = p::; 0.05; ** = p::; 0.01;
***= p::; 0.001).

3). The one constant-frequency train that had an interpulse interval < 25 ms produced greater
force than the corresponding optimizing train. Similarly, for the human quadriceps femoris,
in the control state, constant-frequency trains produced greater forces than optimized trains
for almost all interpulse intervals < - 80 ms (the muscles' twitch contraction time; note that
no measurement was made of stimulus interpulse intervals> 1.5 x this contraction time; see
Fig. 4).
In the fatigued state, very different results were observed. For the rat gastrocnemius,
the twitch contraction time (- 25 ms) showed little change when the muscle was fatigued,
though the optimized trains produced force augmentation for all trains with interpulse
intervals:e: 25 ms. The stimulus trains with 20-ms intervals revealed no significant difference
in the forces produced by the two train types. For the human quadriceps femoris, the twitch
contraction time decreased from -80 ms to -40 ms with fatigue, and for all trains with
interpulse intervals :e: 30 ms, the optimized trains produced significantly more force than the
constant-frequency trains. Interestingly, although previous studies have shown that catch-
inducing trains can augment forces from low-frequency subtetanic trains, these data show
that when muscles are fatigued catch-inducing trains can produce greater forces than even
the best constant-frequency train.

Anisometric Contractions
As noted by Enoka and Stuart (1992), the muscle wisdom theme needs to be extended
to dynamic conditions to determine if the results observed during isometric conditions hold
true for a wide variety of natural movements. Similarly, studies should examine the use of
variable-frequency trains that exploit the catch-like property of muscle. Unfortunately, there
is a dearth of information on both of these issues. Callister and colleagues (1992) reported
for a turtle hindlimb muscle that during shortening and lengthening contractions, optimized
stimulus trains produced slightly less force augmentation than during isometric contractions.
We have recently reported (Binder-Macleod & Lee, 1993) that in the potentiated, non-fa-
tigued state, optimized stimulus trains produced modest increases in the force-time integral
Variable-Frequency Stimulation Patterns 237

of the human quadriceps muscle during shortening contractions, and this augmentation
increased as the muscle became fatigued. In contrast, no augmentation was seen during
lengthening contractions in the control (non-fatigued) or fatigued state. Clearly, substantial
additional work is needed on anisometric contractions.

RELEVANCE OF THE CATCH-LIKE PROPERTY TO

VOLITIONAL CONTRACTIONS

High-frequency bursts (i.e., doublets or triplets) of motor unit EMG activity at the
onset of a train of activity have been reported to occur during volitional contractions in
human (Maton & Gamet, 1989) and animals (Zajac & Young, 1980b; Hennig & L0mo, 1985,
1987). In addition, motor unit activation rates during volitional contractions in human appear
to be sufficiently low to allow these high-frequency bursts to elicit a catch-like effect
(Bellemare et aI., 1983; Bigland Ritchie et aI., 1983a). Based on the appearance ofa doublet
during walking in high-decerebrate cat Zajac and Young (l980b) reasoned that, if during
normal locomotion there was a constraint on the number of pulses within a train to reduce
fatigue, then a motor unit should fIre with the activation pattern that optimizes the force.
However, several studies have shown that doublets occur infrequently during walking in
normal cats (Hoffer et aI., 1987) and that the incidence of doublets decreased as the treadmill
speed increased in both high-decrebrate (Zajac & Young, 1980b) and normal cats (Hoffer et
aI., 1987). In addition, doublets are seen in some subpopulations of motor units and not others
in the rat hindlimb (Hennig & L0mo, 1985, 1987). These data suggest that there must be
some functional advantages to utilizing the catch-like property for some motor unit types
and under some circumstances. Unfortunately, there are not enough data available to
determine which parameters determine when the catch-like property is used during volitional
activation of motor units (Bevan et aI., 1992; Binder-Macleod & Barrish, 1992).

CONCLUSION

In conclusion, it appears that selected patterns of variable-frequency stimulation

can optimize force output during sustained and prolonged intermittent contractions. These
stimulation patterns have been shown to be effective during both voluntary and imposed
high-force contractions. However, the boundary conditions for such patterns remain an
open issue. The stimulus patterns that exploit the catch-like property have been shown
to reduce and delay the fatigue produced by imposed contractions, with their role in
voluntary movement also an open issue. However, imposed stimulus patterns that exploit
the catch-like property of muscle have signifIcant implications for the use of functional
electrical stimulation in rehabilitation and to our overall understanding of segmental
motor control.

ACKNOWLEDGMENTS

The author would like to thank Dr. Lynn Snyder-Mackler for her criticism of a draft
of this chapter. The work was supported by United States Public Health Service (USPHS)
grant AR 441264. The author's attendance at the 1994 Bigland-Ritchie conference was
supported, in part, by AR 441264, and the University of Miami.
 238 S. A. Binder-Macleod

REFERENCES
Bellemare F & Garzaniti N (1988). Failure of neuromuscular propagation during human maximum voluntary
contraction. Journal ofApplied Physiology 64, 1084-1093.
Bellemare F, Woods JJ, Johansson R & Bigland-Ritchie B (1983). Motor unit discharge rates in maximal
voluntary contractions of three human muscles. Journal ofNeurophysiology 50, 13 80-1392.
Bevan L, Laouris Y & Reinking RM (1992). The effect of the stimulation pattern on the fatigue of single motor
units in adult cats. Journal of Physiology (London) 449, 85-108.
Bigland-Ritchie B (1981). EMG/force relations and fatigue of human voluntary contractions. Exercise Sport
Sciences Reviews 9,75-117.
Bigland-Ritchie B, Cafarelli E & Vallestad NK (1986). Fatigue of submaximal static contractions. Acta
Physiologica Scandinavica 128, 137-148.
Bigland-Ritchie B, Johansson R, Lippold DCJ, Smith S & Woods JJ (1983a). Changes in motoneuron fIring
rates during sustained maximal voluntary contractions. Journal ofPhysiology (London) 340, 335-346.
Bigland-Ritchie B, Johansson R, Lippold DCJ & Woods JJ (1983b). Contractile speed and EMG changes
during fatigue of sustained maximal voluntary contractions. Journal ofNeurophysiology 50, 313-324.
Bigland-Ritchie B, Jones DA, Hosking GP & Edwards RHT (1978). Central and peripheral fatigue in sustained
maximum voluntary contractions of human quadriceps muscle. Clinical Science and Molecular
Medicine 54, 609-614.
Bigland-Ritchie B, Jones DA & Woods JJ (1979). Excitation frequency and muscle fatigue: electrical responses
during human voluntary and stimulated contractions. Experimental Neurology 64, 414-427.
Bigland-Ritchie B, Thomas CK, Rice CL, Howarth N & Woods JJ (1992). Muscle temperature, contractile
speed, and motorneuron fIring rates during human voluntary contractions. Journal of Applied
Physiology 73, 2457-2461.
Binder-Macleod SA & Baadte SA (1992). IdentifIcation of optimal interpulse interval (IPI) patterns for
activation offatigued human quadriceps femoris muscle. Society for Neuroscience Abstracts 18, 1557.
Binder-Macleod SA & Barker CB (1991). Use of the catchlike property of human skeletal muscle to reduce
fatigue. Muscle & Nerve 14, 850-857.
Binder-Macleod SA & Barrish WJ (1992). Force response of rat soleus muscle to variable-frequency train
stimulation. Journal ofNeurophysiology 68, 1068-1078.
Binder-Macleod SA & Clamann HP (1989). Force-frequency relations of cat motor units during linearly
varying dynamic stimulation. Journal ofNeurophysiology 61, 208-217.
Binder-Macleod SA & Guerin T (1990). Preservation of force output through progressive reduction of
stimulation frequency in the human quadriceps femoris muscle. Physical Therapy 70,619-625.
Binder-Macleod SA, Halden EE & Jungles KA (1995). Effects of stimulation intensity on the physiological
responses of human motor units. Medicine and Science in Sports and Exercise 27,556-565.
Binder-Macleod SA & Landis LJ (1994). Effects of train frequency and fatigue state on the catchlike property
in the rat gastrocnemius muscle. Society for Neuroscience Abstracts 20, 1204.
Binder-Macleod SA & Lee SCK (1993). Catch1ike property of human muscle during shortening and length-
ening contractions. Society for Neuroscience Abstracts 19, 155.
Binder-Macleod SA & McDermond LR (1992). Changes in the force-frequency relationship of the human
quadriceps femoris muscle following electrically and voluntarily induced fatigue. Physical Therapy
72,95-104.
Burke RE, Rudomin P & Zajac FE (1970). Catch properties in single mammalian motor units. Science 168,
212-214.
Burke RE, Rudomin P & Zajac FE (1976). The effect of activation history on tension production by individual
muscle units. Brain Research 109, 515-529.
Callister RJ, Laidlaw DH, Reinking RM & Stuart DG (1992). The catch-like property of turtle muscle: effect
of shortening, lengthening, and fatiguing isometric contractions. Society for Neuroscience Abstracts
18, 1559.
Cooper RG, Edward RHT, Gibson H & Stokes MJ (1988). Human muscle fatigue: frequency dependence of
excitation and force generation. Journal ofPhysiology (London) 397, 585-599.
Cooper S & Eccles JC (1930). The isometric response of mammalian muscles. Journal ofPhYSiology (London)
69, 377-385.
Dietz V (1978) Analysis of the electrical muscle activity during maximal contraction and the influence of
ischaemia. Journal of the Neurological Sciences 37, 187-197.
Dorfman LJ, Howard IE & McGill KC (1990). Triphasic behavioral response of motor units to submaximal
fatiguing exercise. Muscle & Nerve 13, 621-628.
Variable-Frequency Stimulation Patterns 239

Dubose L, Schelhorn TB & Clamann HP (1987). Changes in contractile speed of cat motor units during activity.
Muscle & Nerve 10, 744-752.
Duchateau J & Hainaut K (1986a). Nonlinear summation of contractions in striated muscle. I. Twitch
potentiation in human muscle. Journal ofMuscle Research and Cell Motility 7,11-17.
Duchateau J & Hainaut K (1986b). Nonlinear summation of contractions in striated muscle. II. Potentiation
of intracellular Ca2+ movements in single barnacle muscle fibers. Journal of Muscle Research and
Cell Motility 7, 18-24.
Edwards RHT (1981). Human muscle function and fatigue. In: Porter R, Whelan J (eds.), Human Muscle
Fatigue, Physiological Mechanisms, pp. 1-18. London (Ciba Foundation symposium 82): Pitman
Medical.
Edwards RHT, Hill DK, Jones DA & Merton PA (1977). Fatigue of long duration in human skeletal muscle
after exercise. Journal ofPhysiology (London) 272, 769-778.
Enoka RM & Stuart DG (1992). Neurobiology of muscle fatigue. Journal of Applied Physiology 72,
1631-1648.
Fallentin N, Jorgensen K & Simonsen EB (1993). Motor unit recruitment during prolonged isometric
contractions. European Journal ofApplied Physiology & Occupational Physiology 67,335-341.
Fitts RH (1994). Cellular mechanisms of muscle fatigue. Physiological Reviews 74, 49-94.
Garland SJ, Enoka RM, Serrano LP & Robinson GA (1994). Behavior of motor units in human biceps brachii
during a submaximal fatiguing contraction. Journal ofApplied PhYSiology 76,2411-2419
Garland SJ, Gamer SH & McComas AJ (1988). Relationship between numbers and frequencies of stimuli in
human muscle fatigue. Journal ofApplied Physiology 65, 89-93.
Gordon DA, Enoka RM, Karst GM, & Stuart DG (1990). Force development and relaxation in single motor
units of adult cats during a standard fatigue test. Journal ofPhysiology (London) 421, 583-594.
Grimby L, Hannerz J & Hedman B (1981). The fatigue and voluntary discharge properties of single motor
units in man. Journal ofPhysiology (London) 316, 545-554.
Hennig R & L0mO T (1985). Firing patterns of motor units in normal rats. Nature 314, 164-166.
Hennig R & L0mO T (1987). Gradation of force output in normal fast and slow muscles of the rat. Acta
Physiologica Scandinavica 130, 133-142.
Hoffer JA, Sugano N, Loeb G, Marks WB, O'Donovan MJ, & Pratt CA (1987). Cat hindlimb motoneurons
during locomotion. II. Normal activity patterns. Journal ofNeurophysiology 57, 530-553.
Jones DA (1981). Muscle fatigue due to changes beyond the neuromuscular junction. In: Porter R, Whelan J
(eds.), Human Muscle Fatigue, PhYSiological Mechanisms, pp. 178-196. London (Ciba Foundation
symposium 82): Pitman Medical
Jones DA, Bigland-Ritchie B & Edwards RHT (1979). Excitation frequency and muscle fatigue: Mechanical
responses during voluntary and stimulated contractions. Experimental Neurology 64,401-413.
Jones DA, Newham DJ & Torgan C (1989). Mechanical influences on long-lasting human muscle fatigue and
delayed-onset pain. Journal ofPhysiology (London) 412, 415-427.
Kernell D, Eerbeek 0 & Verhey BA (1983). Relation between isometric force and stimulation rate in cat's
hindlimb motor units of different twitch contraction times. Experimental Brain Research 50, 220-227.
Marsden CD, Meadows JC & Merton PA (1971). Isolated single motor units in human muscle and their rate
of discharge during maximal voluntary effort (abstract). Journal of Physiology (London) 217,
12P-13P.
Marsden CD, Meadows JC & Merton PA (1976). Fatigue of human muscle in relation to the number and
frequency of motor impulses (abstract). Journal ofPhysiology (London) 258, 94P-95P.
Marsden CD, Meadows JC & Merton PA (1983). "Muscular wisdom" that minimizes fatigue during prolonged
effort in man: peak rates of motoneuron discharge and slowing of discharge during fatigue. Advances
in Neurology 39, 169-211.
Maton B & Gamet D (1989). The fatigability of two agonistic muscles in human isometric voluntary
submaximal contraction: an EMG study. European Journal ofApplied Physiology 58,369-374.
Merton PA (1954). Voluntary strength and fatigue. Journal ofPhysiology (London) 128, 553-564.
Metzger JM & Fitts RH (1987). Fatigue from and high- and low-frequency muscle stimulation: contractile
and biochemical alterations. Journal ofApplied Physiology 62, 2075-2082.
Nress K & Storm-Mathisen A (1955). Fatigue of sustained tetanic contradictions. Acta Physioiogica Scandi-
navica 34, 351-366.
Parmiggiani F & Stein RB (1981). Nonlinear summation of contractions in cat muscle. II. Later facilitation
and stiffness changes. Journal of General Physiology 78, 295-311.
Powers RK & Binder MD (1991). Summation of motor unit tensions in the tibialis posterior muscle of the cat
under isometric and nonisometric conditions. Journal ofNeurophysiology 66, 1838-1846.
240 S. A. Binder-Macleod

Stein RB & Parmiggiani F (1979). Optimal motor patterns for activating mammalian muscle. Brain Research
175,372-376.
Stokes MJ, Edwards RHT & Cooper RG (1989). Effect of low frequency fatigue on human muscle strength
and fatigability during subsequent stimulated activity. European Journal of Applied Physiology 59,
278-283.
Stuart DG & Callister RJ (1993). Afferent and spinal reflex aspects of muscle fatigue: Issues and speculations.
In: Sargent AJ, Kernell D (eds.), Neuromuscular Fatigue, pp. 169-180. Amsterdam: Royal Nether-
lands Academy of Arts and Sciences.
Thomas CK Bigland-Ritchie B & Johansson RS (1991). Force-frequency relationship of human thenar motor
units. Journal ofNeurophysiology 65, 1509-1516.
Thomas CK, Woods JJ & Bigland-Ritchie B (1989). Impulse propagation and muscle activation in long
maximal voluntary contractions. Journal ofApplied Physiology 67, 1835-1842.
Thompson LV, Balog EM, Riley DA & Fitts RH (1992). Muscle fatigue in frog semitendinosus: Alterations in
contractile function. American Journal ofPhysiology (Cell Physiology) 262, C 1500-C 1506.
Wilson DM & Larimer JL (1968). The catch properties of ordinary muscle. Proceedings of the National
Academy of Science, USA 61, 909-916.
Woods JJ, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal ofNeurophysiology 58, 125-137.
Zajac FE & Young JL(1980a). Properties of stimulus trains producing maximum tensiontime area per pulse
from single motor units in medial gastrocnemius muscle of the cat. Journal of Neurophysiology 43,
1206-1220.
Zajac FE & Young JL (1980b). Discharge properties of hindlimb motoneurons in decerebrate cats during
locomotion induced by mesencephalic stimulation. Journal of Neurophysiology 43,1221-1235.
 17

OVERVIEW: POTENTIAL ROLE OF

SEGMENTAL MOTOR CIRCUITRY IN
MUSCLE FATIGUE

U. Windhorst and G. Boorman

Departments of Clinical Neurosciences and Medical Physiology

The University of Calgary
Faculty of Medicine
3330 Hospital Drive N.W., Calgary, Alberta T2N 4Nl, Canada

ABSTRACT

This chapter reviews several mechanisms that the eNS may use to mitigate muscle fatigue,
including intrinsic motoneuron properties and feedback systems. The emphasis is on the
effects of sensory inputs on spinal cord interneurons including: Renshaw cells; Ib inhibitory
interneurons; interneurons mediating presynaptic inhibition; Ia inhibitory interneurons; and
interneuronal networks constituting central pattern generators for locomotion. This exercise
brings out how little is known about the operation of these circuits in dealing with muscle
fatigue.

INTRODUCTION
Muscle fatigue is a phenomenon of immense practical importance in normal indi-
viduals, particularly in physically demanding occupations and in athletic competition, and
in patients with various types of disorders affecting the CNS and/or peripheral neuromuscu-
lar apparatus. Whereas the first part of this statement is everyday experience, the latter is not
as evident because other phenomena and symptoms often supersede fatigue. For example,
in patients with upper motoneuron disease after stroke, multiple sclerosis and spinal cord
injury, leg muscle fatigability is enhanced, possibly resulting from conversion of many
fatigue-resistant motor units into fatigable ones after prolonged disuse (Lenman et aI., 1989;
Miller et aI., 1990), although there are reports to the contrary (for review: Enoka & Stuart,
1992). In the rare cases of mitochondrial myopathies leading to lactacidosis (Edwards et aI.,
1982), the predominant features are symptoms of severe exhaustion, weakness and muscle
fatigue during exercise. The lactacidosis should be a strong stimulus to high-threshold
muscle afferents which may reflexly alter motoneuron discharge. Muscle fatigue is also an
important limiting factor in situations in which functional electrical stimulation (FES) is
used to help restore function after neurological damage (Binder-Macleod & Barker 1991).

241
242 U. Windhorst and G. Boorman

Despite its ubiquity, little is yet known about the mechanisms contributing to muscle
fatigue and the means available to the CNS to mitigate its effects (Enoka & Stuart, 1992; Stuart
& Callister, 1993). This is due in only minor part to the lack of a unanimously accepted
definition of muscle fatigue (Clamann, 1990; Enoka & Stuart, 1992; Bigland-Ritchie, 1993).
Rather, the key problem is that multiple processes and mechanisms are involved at different
levels, from higher-order CNS structures to electrical and biochemical alterations within
muscle fibers (Clamann, 1990; Enoka & Stuart, 1992). Whatever muscle fatigue's underlying
causes, it becomes manifest as " ... any reduction in the force-generating capacity (measured
by the maximum voluntary contraction), regardless of the task performed" (Bigland-Ritchie &
Woods 1984). However, nature appears to have evolved a number of mechanisms for maintain-
ing force output at the optimum achievable at any time in the course of fatigue and thus
mitigating its effects. Some such mechanisms are intrinsic to the fatiguing muscle fibers
themselves, whereas others (like effects of sensory feedback) are located in the CNS. Thus,
study of these mechanisms not only reveals specifics about muscle fatigue, but provides a
window into the way the CNS optimizes the performance of its neuromuscular periphery. What
is now known about such mechanisms is orders of magnitude less than what has yet to be
discovered. For this reason, the present chapter deals with problems rather than solutions. It
focuses on the potential role of sensory feedback in adjusting static and dynamic properties of
spinal circuits which may be involved in optimizing force output during muscle fatigue.
Like the mechanisms giving rise to muscle fatigue, those counteracting it at the
segmental level are complex and multi-faceted. They include motor unit potentiation, the
diversity and recruitment order of the motor unit population, and the operation of premo-
toneuronal networks.

MOTOR UNIT POTENTIATION

The first protective mechanism resides in the muscle fibers. It involves potentiation, i.e.,
an increase in motor unit force output due to preceding activation (see Kemell, Chapter 9). This is
illustrated in Fig. 1, showing that injection oflong-Iasting depolarizing current into cat lumbar
u-motoneurons caused them to discharge, first at a high rate and then they adapted to lower rates.
Motoneurons labeled S unit (see below) did not adapt nearly as much as did FF unit
motoneurons. Similar differences in adaptation have been seen with extracellular current
application, even with repeated intermittent current pulses (Spielmann et aI., 1993). Further-
more, Fig. 1B shows that the average force declined much more in FF motor units than in S
units. Thirdly, the force ofFF units first declined steeply and then exhibited an intermediate
hump which probably resulted from force potentiation. The high initial firing rate of FF
motor units is well suited to potentiate their twitches after some seconds. Hence, motor units
may show both fatigue and potentiation in parallel (Gordon et ai., 1990; Powers & Binder,
1991; Bevan et aI., 1992; see also Thomas, Chapter 10), the latter transiently opposing the
former. In addition, selected firing patterns with initial short interspike intervals may
optimize force output in the face of fatigue (Binder-Macleod & Barker 1991; Bevan et ai.,
1992), a form of pattern optimization (see Binder-Macleod, Chapter 16)

MOTOR UNIT POPULATION: DIVERSITY OF MOTOR UNIT

TYPES AND RECRUITMENT ORDER

Implicit in Fig. I is the finding that there are different types of motor units, and that
they are differentially fatigable (type S vs. FR vs. FF; Burke et ai., 1973; Burke, 1981;
Overview: Potential Role of Segmental Motor Circuitry in Muscle Fatigue 243

A
100 5 unit
0'.
G) FF unit
iii 50
a:

o0 30 60

B
.,.
100
5 unit

G)
e
0 50
u.
FF unit

00 30 60
Time
Figure 1. A, plot of the firing rate (%) versus time (s) for first minute of motoneuron discharges produced by
depolarizing current of 5 nA above threshold for rhythmic firing. For consecutive seconds of discharge, mean
firing rates have been connected by straight lines. Firing rates given as percentage of rate during 2nd second
of discharge (first value plotted = 16.1 pulses per second (pps) for the S unit and 28.9 pps for the FF unit).
Twitch contraction time was 53 ms for the S unit and 25 ms for the FF unit. The standard fatigue index according
to Burke and colleagues (1973) was 0.97 for the S unit and 0.07 for the FF unit. D, contractile force produced
by the discharges in A. Force given as percentage of value for 2nd second of discharge. In nonfused
contractions, mean force was calculated over time periods of I s (reproduced from Kernell & Monster, 1982;
their Fig. 5, by permission of Springer-Verlag).

Kernell, 1993; for details see Kernell, Chapter 9; Thomas, Chapter 10). The type S motor
units have the longest duty periods throughout the day and are active in most muscle activities
of moderate strength and speed. This activity pattern is in keeping with the finding that their
discharge lasts longer during sustained depolarization (Kernell and Monster, 1982; Spiel-
mann et aI., 1993). Faster contracting, more fatigable motor units come into play during
stronger and faster contractions. Thus, the functionally ordered recruitment from type S to
FR to FF motor units of different size, usually referred to as Henneman's size principle
(Henneman, 1957; see also Binder & Mendell, 1990; Kernell, 1992; Sawczuk et aI., Chapter
8; Kernell, Chapter 9; Sargeant & Jones, Chapter 24; and Bigland-Ritchie, Chapter 26), is
another protective mechanism and may be called recruitment optimization (Kernell, 1993).
However, when strong contractions have to be sustained, the problem of high fatigability of
strong motor units has to be dealt with. Any fatigue-reducing mechanism should, therefore,
focus on fast-twitch motor units (FF in particular). In fact, at the single unit level, F (i.e., FR
+ FF) motor units exhibit faster and more extensive adaptation of discharge rate to maintained
currents. At the population level, the recruitment patterns found normally during non-fatigu-
ing contractions could conceivably be altered. Indeed, evidence is accumulating that recruit-
ment of motor units, which during non-fatigued increasing muscle contractions usually
occurs in the above orderly fashion, may change during fatigue, such that low-threshold
motor units may not be recruited (Enoka & Stuart, 1992), with the underlying mechanisms
244 U. Windhorst and G. Boorman

yet to be detennined. Presumably, both intrinsic motoneuron properties and synaptic inputs
have to be taken into account (Burke, 1981), requiring consideration of premotoneuronal
networks.

PREMOTONEURONAL NETWORKS

The third line of defense against fatigue resides in the action of premotoneuronal
networks, including five components: I) spinal tracts descending from higher motor centers;
2) neuromodulatory systems; 3) reflex circuits; 4) intrinsic spinal feedback circuits; and 5)
locomotor pattern generators
The first category includes a number of tracts mediating motor commands to
motoneurons.
The second category refers to descending systems from brainstem structures such as
the locus coeruleus and raphe nuclei, which modulate the excitability of many spinal neurons
including those of the central pattern generator (CPG).
The third category is the largest and most heterogeneous, including pathways from
a variety of muscle receptor afferents which exert effects on a-motoneurons, directly or most
often indirectly, via interneurons (see below). These include mono- and oligosynaptic
connections from muscle spindle group Ia and II afferents, the classical autogenetic di- and
trisynaptic pathways, as well as more complex connections from Golgi tendon organ (group
Ib) afferents, and various pathways from higher-threshold afferents which may sense some
aspects of muscle fatigue via mechanical and/or chemical stimuli (see Garland & Kaufman,
Chapter 19).
The fourth category primarily encompasses recurrent inhibition, involving Renshaw
cells which receive inputs from many species of muscle afferent.
Thefifth category encompasses interneuronal networks which constitute the CPG for
locomotion and allied movements. The CPG cyclically modulates reflex pathways and, in
tum, is influenced by muscle afferents and may thus playa role in fatigue-compensating
mechanisms during cyclic motor activities.
There is no clear delineation between the above circuits, requiring that we first
provide a global approach and then dissect out single mechanisms.

A GLOBAL APPROACH: NEGATIVE FORCE FEEDBACK

SYSTEMS

Negative feedback circuits have been designed both by nature and man to compensate
for various kinds of disturbances interfering with the execution of selected tasks. For example,
activation of a skeletal muscle by its motoneurons is co-modulated by feedback from receptors
monitoring muscle fiber length and force; that is, muscle spindles and Golgi tendon organs,
respectively. With qualifications to be discussed below, the afferents from these receptors can
have actions on synergistic and homonymous motoneurons that provide negative feedback
(Fig. 2). Under selected conditions, such actions allow compensation for disturbances, which
may arise either externally, internally, or both (Houk & Rymer, 1981). An example ofan internal
disturbance is muscle fatigue, during which event properties of the controlled plant (i.e.,
muscle) change in response to continuing activation (Houk et aI., 1970).
When the force-producing capability of muscle declines, a force feedback would be
well suited (Houk et aI., 1970). There are at least two parallel loops that could contribute to
force feedback. First and foremost, Golgi tendon organs are very sensitive force sensors, and
Overview: Potential Role of Segmental Motor Circuitry in Muscle Fatigue 245

Figure 2. Schematic diagram of selected spinal

reflex circuits involving group Ia and Ib affer-
ents. A pair of antagonist muscle groups are
shown for each circuit. Excitatory neurons are
represented by open circles and their synaptic
terminals by T-shaped lines. Inhibitory neurons
are represented by filled circles. Muscle spin-
dles are symbolized as a straight line with a coil
around its middle portion. They consist of spe-
cialized muscle fibers with a separate motor
innervation, as indicated by the thin line origi-
nating from the extensor y-motoneuron (labeled
Ye) in the spinal cord. Thick sensory nerve fibers,
denoted Iae. originate from the spiral sensory
ending on extensor muscle spindles and carry
information to the spinal cord where they con-
tact several types of neuron, with monosynaptic Renshaw cells _~--T
excitatory connections to a-motoneurons (a e) of
their own (homonymous) muscle and muscles of
similar (synergistic) function. When ankle ex-
tensor muscles are stretched, their Ia e axons are
excited and, in tum, excite homonymous/syner-
gistic motoneurons which induce contraction in
the stretched muscles, thereby opposing stretch
(negative feedback). Extensor Iae axons also
inhibit antagonist flexor motoneurons via recip-
rocal Ia inhibitory interneurons (Recip. Ia IN),
and vice versa (latter not illustrated). Renshaw
cells receive their main excitatory input from a
particular motoneuron pool (here extensor) and
inhibit homonymous and synergistic a-and less
frequently and strongly, y-motoneurons. Recip.
Ia INs are excited by homonymous and syner-
gistic Ia fibers, and other Renshaw cells (in
particular, those predominantly related to an-
tagonist motoneurons); mutual inhibition, not
shown. Ib afferents (here of flexor origin: Ib f )
from Golgi tendon organs are located at musculo-tendinous junctions. There axons di- or trisynaptically inhibit
homonymous and synergistic motoneurons (autogenetic inhibition.). The pattern ofIb afferent connections is
actually more complicated, involving alternative excitatory pathways to synergistic motoneurons (particularly
those of extensor muscles). These may be switched on during the stance phase oflocomotion, along with other
complex reflex effects (see text). a- and y-motoneurons, Recip. Ia INs, Renshaw cells and other interneurons
receive a variety of sensory and descending, excitatory and inhibitory inputs (upper oval labeled Inputs), as
symbolized by the dashed arrows. These inputs include afferents from high-threshold muscle afferents.

under certain circumstances (see below), their Ib afferents exert inhibitory effects on
homonymous and synergistic motoneurons according to the classical concept of autogenetic
inhibition (Fig. 2, left side). Thus, when force declines due to fatigue, Ib afferent excitation
should be reduced, and thereby also autogenetic inhibition. Descending motor commands
should then be disinhibited and muscle excitation enhanced. In addition, mono-and oligo-
synaptic excitation of homonymous and synergistic motor units from muscle spindle
afferents (Fig. 2, right side) might assist, because this loop also constitutes a negative
feedback system.
Muscle spindles monitor the length of their in-parallel skeletomotor muscle fibers.
Even when the whole muscle is in isometric contraction (i.e., at constant overall length), a
246 U. Windhorst and G. Boorman

decline in muscle force entails an increase in internal muscle length because muscle fibers
are connected to bones via in-series elastic elements which shorten with decreasing force.
Therefore, provided other factors like fusimotor innervation remain constant, spindles
should lengthen and their firing rate increase during muscle fatigue (see Hagbarth &
Macefield, Chapter 18). The net outcome should be increased motoneuron excitation. The
problem then is whether the loop gain of the system is large enough to reduce fatigue effects.
Recent measurements in man suggest this to be the case. Based on several assumptions and
using two different methods, Kirsch and Rymer (1992) estimated the open-loop gain in the
compound force feedback system to average between 1.3 and 4.6 (first method) or 8.3
(second method) and concluded that these unexpectedly high loop gains may significantly
reduce the susceptibility of the overall neuromuscular system to fatigue by 56-89%.
However, this simple model poses a number of problems which, in part, are related
to the precise nature ofthe fatiguing task, and, in part, to the interference of additional reflex
actions. Thus, during maximal isometric muscle contractions in man when no recruitment
of additional motor units is assumed to occur, motor unit discharge rates decline, contrary
to expectations based on negative feedback operation. Only in sustained or repeated
sub-maximal contractions can such recruitment be observed (Vl1Jllestad et aI., 1988; Fugle-
vand et aI., 1993). Furthermore, muscle spindle firing in man appears not to increase, but to
decline during fatiguing muscle contraction (see Hagbarth & Macefield, Chapter 18). This
may have detrimental effects on the gain in the autogenetic Ib afferent pathway because in
the cat about 40% of the Ib inhibitory interneurons with Ib input from extensor muscles are
also excited by spindle Ia afferents (Jankowska et aI., 1981). Therefore, since Ia input is
declining, these interneurons' sensitivity to Ib input is likely to decrease as well. The patterns
of changes in Ib afferent discharge during fatiguing muscle contractions in humans are not
yet known. Furthermore, the gain of force feedback can also be altered, at a number of sites,
by modulating influences arising, in particular, from high-threshold afferents in the fatiguing
muscle(s) (for details see Garland & Kaufman, Chapter 19). Finally, central effects ofIa and
Ib afferents are subject to presynaptic inhibition (see below), which might also change during
fatigue. In summary, the neuromuscular processes during fatigue are more complex than
captured in global feedback models, necessitating another approach.

MUSCLE WISDOM: RATE OPTIMIZATION AND ITS

MECHANISMS

Of increasing visibility and appeal, is the formulation of a hypothesis based on system

optimization, and termed muscle wisdom (for details, see Binder-Macleod, Chapter 16). The
idea is currently based on results using imposed (electrical) stimulation and voluntary contrac-
tions of human subjects. The issue of its generalization is still open (Enoka and Stuart, 1992)

Motor Unit Firing Rate Adaptation

During fatiguing maximal voluntary isometric muscle contractions, human motor
units reduce their firing rate over several tens of seconds from an initially high value
(Bigland-Ritchie et aI., 1983a,b; Marsden et aI., 1983; Bigland-Ritchie & Woods, 1984;
Gandevia et aI., 1990); for exceptions see Enoka and Stuart (1992). From simple control
considerations, this decrease is counter-intuitive because, in view of the monotonic (sigmoi-
dal) steady-state force-frequency relation (e.g., Clamann, 1990; Kernell, 1992; for further
details, see Kernell, Chapter 9), compensation for force reduction would seem to call for a
firing rate increase. Decrease in force-producing capability during continuing muscle exer-
Overview: Potential Role of Segmental Motor Circuitry in Muscle Fatigue 247

cise being inevitable, the only possibility is to minimize it. Why then would a decrease in
firing rate optimize force output?
During progressing fatigue, twitch contraction relaxation slows in whole muscle,
motor units and muscle fibers (Bigland-Ritchie et aI., 1983b; Dubose et aI., 1987; reviewed
in Bigland-Ritchie, 1993). This has several consequences. Contractile slowing reduces
fusion frequency and, hence, the need for high activation rates which are particularly
fatiguing (Marsden et aI., 1983; Binder-Macleod & Guerin, 1990; Clamann, 1990; Binder-
Macleod & Barker, 1991; reviewed in Gandevia, 1993). Thus, to maintain the optimal force
output achievable, activation rates should decrease over time (see also Botterman & Cope,
1988). Also, the reduced firing rate may help prevent failure of spike propagation at motor
axon branching points (see Fuglevand et aI., 1993; Fuglevand, Chapter 6).
Therefore, it has been hypothesized that the rate reduction is an adaptive optimizing
mechanism (i.e., muscle wisdom; Bigland-Ritchie et aI., 1983a; Marsden et aI., 1983; Enoka
& Stuart, 1992; see also Binder-Macleod, Chapter 16). An implicit corollary is that each
motor unit should display its particular rate reduction profile tailored to its individual
contractile and fatigue properties. Another corollary is that, since fatigability of muscle
depends on muscle length (Fitch & McComas, 1985) and other peripheral factors, the optimal
adjustment of motoneuron firing patterns would best be accomplished by taking account of
peripheral circumstances through muscle afferent feedback (see below).
What requires optimization for muscle wisdom to occur is a fatiguing motor unit's
discharge pattern, in terms of both its mean rate (frequency optimization; see Kernell 1993)
and, potentially, its variability (pattern optimization, for further details, see Binder-Macleod,
Chapter 16). Several possible mechanisms need to be considered, including: desceI).ding
commands, motoneuron properties, and reflex effects onto motoneurons from fatiguing
muscle(s); that is, as mediated by mono- and oligosynaptic excitatory connections, Renshaw
cells, Ib inhibitory interneurons, other inhibitory interneuronal systems, and interneurons
involved in CPGs.

Descending Motor Commands

Voluntary motoneuron activation declines with fatigue (McKenzie et aI., 1992;
Gandevia 1993; for further details, see Gandevia et aI., Chapter 20). This may affect
motoneuron discharge not only directly, but also indirectly via interneurons intercalated in
reflex and other pathways.

Motoneuron Properties
Motoneuron properties may contribute to late firing-rate adaptation (Fig. lA),
through a variety of potential mechanisms that are discussed in Sawczuk et aI., Chapter 8.
Appropriately, late adaptation is more prevalent in F-type than in S-type motoneurons (Fig.
1A; Kernell & Monster, 1982; Spielmann et aI., 1993). If these mechanisms were intrinsic
to the motoneuron and unalterable by external (synaptic-input) influences, late adaptation
could not be matched to mechanical and metabolic circumstances encountered by the motor
unit's muscle fibers. For instance, the strength and time course of fatigue depends on muscle
length, a peripheral variable (Fitch & McComas, 1985). However, at least after-hyperpolari-
zation is mutable: for example, by descending monoaminergic systems favoring plateau
potentials and perhaps other descending commands (see above), and during fictive locomo-
tion in decerebrate cats, where it is reduced and does not significantly affect motoneuron
discharge patterns (Brownstone et aI., 1992). Likewise it may not be an important feature of
motoneuronal discharge during natural movements, including maximal voluntary muscle
contractions (Stuart & Callister, 1993). But the possibility remains that sensory feedback
248 U. Windhorst and G. Boorman

from the fatiguing muscle alters after-hyperpolarization to contribute to appropriate mo-

toneuronal adaptation (Windhorst & Kokkoroyiannis, 1991).

Afferent Feedback from the Fatiguing Muscle

To match the discharge patterns of motor units to their changing contractile properties,
afferent feedback appears to be required (for details, see Garland & Kaufman, Chapter 19).
Bigland-Ritchie and colleagues (1986), Woods and colleagues (1987) and Garland and
colleagues (1988) working on humans, and Hayward and colleagues (1988) working on cats
provided indirect evidence that feedback from the fatiguing muscle may inhibit homonymous
and synergistic motoneurons and thus reducing their firing rate. There are two possible sources
of such feedback: large-diameter mechano-afferents, and small-diameter polymodal afferents.
The first group includes afferent feedback to excite motoneurons mono- or oligosy-
naptically, in particular, Ia afferents from primary muscle spindle endings (Fig. 2) and group
II afferents from secondary muscle afferents. Furthermore, group Ib afferents from Golgi
tendon organs whose effects are mediated via Ib inhibitory and other interneurons (Fig. 2).
The second group requires the mediation ofinterneurons, including: 1) Renshaw cells; 2) Ib
inhibitory interneurons receiving input from Ib afferents; 3) interneurons mediating presy-
naptic inhibition; 4) Ia interneurons mediating reciprocal inhibition; 5) interneurons medi-
ating polysynaptic effects from muscle spindle afferents; and 6) interneurons establishing
private pathways from small-diameter muscle afferents to motoneurons.

Large-Diameter Muscle Afferents

Large-diameter afferents from fatiguing muscles may act in two different ways to
reduce motoneuron firing rate during fatigue. Human muscle spindle afferents reduce their
mean discharge rate in the course of fatiguing muscle contractions, possibly due to a
reduction in fusimotor drive to the spindles (Macefield et ai., 1991; for details, see Hagbarth
& Macefield, Chapter 18). The ensuing disfacilitation of homonymous and synergistic
motoneurons might contribute to the decline of their firing rate. Group II afferents from
muscle spindles may add to these effects. The effect of muscle fatigue on Golgi tendon organ
responsiveness is not clear, nor is their reflex effect (reviewed in Enoka & Stuart, 1992;
Stuart & Callister, 1993). The changes in mean discharge rate, particularly of Ia and Ib
afferents, have been hypothesized to be supplemented by much subtler means of signaling
reflected in their precise temporal patterns of discharge.
Due to their high sensitivity to small muscle fiber length changes, spindle Ia afferents
exhibit prominent transient responses, usually including rate reductions, to twitch contractions
even of single motor units (Windhorst, 1988; Windhorst & Kokkoroyiannis, 1991). Golgi
tendon organs respond with firing rate increases to single motor unit contractions (for review,
see Windhorst, 1988). Large-diameter muscle afferents would thus appear optimally suited to
monitor the change in motor unit contractile properties during fatigue (decay in amplitude and
slowing of contraction), as hypothesized by Windhorst & Kokkoryiannis (1991). Surprisingly,
however, even if large-diameter afferents changed their discharge pattern in response to
changing contractile properties, this does not appear to significantly affect motoneuron firing
rates. Alterations in contractile muscle properties brought about by changes in muscle tempera-
ture and length do not affect motoneuron firing rates (Bigland-Ritchie, 1993).

Small-Diameter Muscle Afferents

Group III-IV (and some group II non-spindle) polymodal afferents may monitor the
mechanical state of the muscle and be sensitized by metabolic substances released by fatiguing
Overview: Potential Role of Segmental Motor Circuitry in Muscle Fatigue 249

muscle fibers (Mense, 1986; for further details, see Garland & Kaufinan, Chapter 19). One
hypothesis proposes that these afferents are primarily responsible for the inhibition exerted onto
homonymous and synergistic motoneurons from fatiguing muscles (see Bigland-Ritchie,
1993). Indeed, acute animal experiments have demonstrated that selected afferents in these
groups were excited by products of muscle metabolism, such as K+, H+, lactic acid and
arachidonic acid (Rotto & Kaufinann, 1988; Sinoway et aI., 1993; see also Garland & Kaufman,
Chapter 19). Other acute animal experiments have shown that afferents in these groups were
excited after forceful, long-sustained fatigue-inducing contractions (Hayward et aI., 1991). It
was suggested (Hayward et aI., 1988; 1991) that these afferents could inhibit homonymous and
synergistic motoneurons via Ib inhibitory interneurons or Renshaw cells which receive
excitatory input from high-threshold muscle afferents (see Windhorst, 1988; Jankowska, 1992).
The above reflex mechanisms are not mutually exclusive, but may predominantly act
at different times during an ongoing fatiguing contraction. In fact, the high-threshold muscle
afferents studied by Hayward and colleagues (1991) were excited at relatively long latency
after the onset of sustained contraction. Initially, the sensitivity of these afferents to
mechanical stimuli was actually reduced. This led Gandevia (1993) to suggest that the
decline in motoneuronal firing rate over the first few seconds of sustained contractions may
primarily reflect a reduction in muscle spindle facilitation, whereas the inhibition from
small-diameter muscle afferents may occur subsequently.

SPINAL INTERNEURONS: INTEGRATIVE CENTERS

In an anatomical but only limited functional sense, motoneurons are the final common
path to muscle. Rather, interneurons and propriospinal neurons comprise premotoneuronal
networks which perform most of the integrative computation. Convergence of descending
motor commands and sensory afferent information on to these networks enables them to
modulate, up-date and adapt motor commands to motoneurons according to the prevailing
state of the peripheral mechanical apparatus, and vice versa. This appears to be a general
principle, applying to all interneurons identified to date, including Renshaw cells, Ib
inhibitory interneurons, reciprocal la inhibitory interneurons, interneurons mediating presy-
naptic inhibition, etc. (Jankowska, 1992). As an example, Fig. 4 (below) depicts the
convergence of inputs onto the traditional Ib interneuron.
Hayward and colleagues (1991) emphasized two types of interneurons for potentially
mediating inhibitory reflex effects onto motoneurons from small-diameter muscle afferents,
Renshaw cells and Ib inhibitory interneurons. These authors gave priority to the Renshaw cell
pathway because they felt that this pathway might be involved in specifically regulating
motoneuron discharge rate (Hayward et aI., 1991; see also Windhorst & Kokkoroyiannis, 1991).

RECURRENT INHIBITION

Hayward and colleagues (1991) presumed that in the course of ongoing fatiguing
contractions, the Renshaw cell burst discharge after each motoneuron spike is strengthened
and prolonged by high-threshold muscle afferent feedback, so as to prolong motoneuron
interspike intervals. This presumption requires that a motoneuron's action potential can
influence the timing of its own next spike via recurrent inhibition; that is, via a high-gain
recurrent inhibitory feedback loop from each motoneuron back onto itself. This assumption
is by no means self-evident. Each motoneuron excites many Renshaw cells, each of which,
in tum, receives input from many motoneurons and distributes its output to many mo-
toneurons. As such, the response of Renshaw cells to asynchronous input from many
250 U. Windhorst and G. Boorman

motoneurons should be a somewhat erratic discharge without producing a prolonged burst

to a single motoneuron's individual action potentials, as proposed by Hayward and col-
leagues (1991). This problem of convergence-divergence is potentially obviated by focusing
and/or synchronization. For focusing, recurrent inhibition should be strongest between
adjacent motoneurons (Windhorst, 1989; Windhorst & Kokkoroyiannis, 1991), and there is
experimental evidence that neighboring motoneurons are coupled via stronger recurrent
inhibition than more remote motoneurons (Hamm et aI., 1987). Moreover, individual
motoneurons are subject to significant self-inhibition (van Keulen, 1981). For synchroniza-
tion to playa role, small subsets of adjacent motoneurons should discharge synchronously;
and therefore, increase the strength and duration of the burst discharges of local Renshaw
cells. These possibilities are not mutually exclusive. Provided the above prerequisites are
met, the mechanism suggested by Hayward and colleagues (1991) would also be appealing
insofar as fast fatigable (FF) motoneurons provide the strongest input to Renshaw cells
(Hultborn et aI., 1988; Windhorst, 1988). Their motoneuron discharges could thus produce
the Renshaw cell burst discharges required to delay each of the subsequent spikes, with
previously recruited smaller motoneurons continuing to provide background excitation to
the fatiguing muscle.
Ifrecurrent inhibition contributes to the decline of motoneuron rate during sustained
maximal contractions, it could do so by virtue of its own intrinsic dynamics and/or by a
change in its gain brought on by descending commands and/or sensory feedback from
fatiguing muscle. To address the first possibility, Windhorst and Kokkoroyiannis (1991)
recorded the discharges of cat Renshaw cells to motor axon stimulation at time-varying rates
simulating motoneuron adaptation during fatigue (Fig. 3A). Typically, during such a test, the
relation between motoneuron and Renshaw cell discharge rates (Fig. 3B) was nonlinear.
Recurrent inhibition was strong over the initial few hundred milliseconds and could con-
ceivably have contributed to the early adaptation of motoneuron firing rate. Subsequently,
recurrent inhibition was relatively depressed for several seconds, but, thereafter, became
stronger again. This latter intrinsic increase in gain could have contributed to the prolonged
decline in motoneuron discharge rate. More generally, this peculiar input-output relationship
shows the need to obtain quantitative data, if precise regulation of motoneuron discharge
patterns by interneurons is to be understood.
Renshaw cells receive additional input from various sources (Windhorst 1988). For
example, activation of a motoneuron pool by descending tracts is often associated with an
inhibition of Renshaw cells. Thus, if during maintained fatiguing muscle contractions,
descending motor activation declines (McKenzie et aI., 1992; Gandevia et aI., 1993;
Gandevia, 1993; for details see Gandevia et aI., Chapter 20), the concomitant reduction in
descending inhibition of Renshaw cells may increase recurrent inhibition, which, in tum,
may contribute to a decline in motoneuron firing rate. Also, with excitation of small-diameter
muscle afferents, the dynamically changing strength of recurrent inhibition could be appro-
priately augmented during fatigue so as to optimally adapt motoneuron discharge, in line
with a suggestion of Hayward and colleagues (1991). However, there is little relevant data
on these possibilities. Kukulka and colleagues (1986), using an indirect H-reflex method in
humans, provided some evidence that recurrent inhibition and/or summation of motoneuron
after-hyperpolarizations increased over the first 30 s of maximal sustained triceps surae
contractions. How this change comes about, which afferents are involved and how they affect
the Renshaw cell dynamics shown in Fig. 3 are not yet known. In principle, afferent effects
onto Renshaw cells could be exerted via two mechanisms. Firstly, the mean discharge rate
(bias) of Renshaw cells could be elevated; and secondly, their dynamic response to changing
motoneuronal input (see Fig. 3) could be altered. Alternatively, or in addition, motoneuron
after-hyperpolarization could be modulated (see above).
Overview: Potential Role of Segmental Motor Circuitry in Muscle Fatigue 251

Figure 3. Response of a Renshaw cell to time-

60 300
varying motor axon stimulation simulating A
adapted motoneuron firing. In an anesthetized cat MN RC
with dorsal roots L6-S I cut, the biceps posterior
muscle nerve was stimulated with a pulse train of 40 200
42-43 s duration. It started at a high initial rate (ca.
45 pps), and decayed exponentially with a fast (10 RC
ms) and a slow (20 s) time constant towards a final

I
20 100
rate of 16 pps. This train was repeated 10 times
every 75 s, and the cycle histogram (bin width 200 MN
ms) displaying the average stimulus rate over time
<Il
is shown as the lower curve labeled MN in part A. -I!! 0 O+--+-------.r-------.------J~
The cycle-averaged firing rate elicited in the Ren- CI
-5 o 15 3D 45
shaw cell (RC) is shown by the upper curve la- ·cc Time from cycle start (s)
beled RC (same bin width). The wriggled thin line ...
:;::

o 300
in part B is a plot derived from the cycle histo- B
grams in part A of instantaneous RC discharge
rate vs. stimulus rate with time progressing as
indicated by the arrow. The curved part of this 200
dynamic input-output relationship covers the first
8-10 s. The straight line has been drawn through
the origin, and the later part of this relationship
corresponding to times beyond 8-10 s (time direc-
tion indicated by arrow). During this later phase
within the cycle, the RC's discharge rate was close
to proportional to the stimulus rate; that is, the
simulated motor axon activation rate. The slope 20 40 60
of this line represents the input-output gain of the Stimulus rate (pps)
system transforming motor axon activation rate
into RC discharge rate. As compared to the later period, this gain undergoes dynamic changes during the first
few seconds (up to lOs), starting from an initially high to a declining gain which then increases again to settle
to a static value represented by the slope of the straight line.

EFFECTS FROM Ib AFFERENTS

Golgi tendon organ Ib afferents from one muscle exert reflex effects that spread
widely to many motor nuclei of a limb and may be inhibitory or excitatory, this effect in part
depending on movement context (Jami, 1992; Jankowska, 1992; see also below). Thus, there
are excitatory and inhibitory Ib interneurons.
The responses of Ib interneurons to muscle contraction before and during muscle
fatigue are neither well known, nor easily predicted. About half of them receive excitatory
input from muscle spindle Ia afferents (see above). They also receive input from high-thresh-
old muscle afferents (Jankowska, 1992). As such, firing rates of Ib interneurons might
decrease due to decreasing inputs from Ia and Ib afferents during ongoing fatiguing muscle
contraction (Fig. 4), but increase due to augmented inputs from high-threshold muscle
afferents. The balance of these effects in determining mean discharge rates is hard to predict,
as is modulation of the gain of Ib interneurons in response to decrementing and slowing
contractions of individual motor units (see above). Therefore, the responses of Ib in-
terneurons need to be determined experimentally. Another important issue is that the different
convergent inputs (cf. Fig. 4) by no means distribute homogeneously to each single neuron,
entailing a fractionation of the interneuron pool (Harrison & Jankowska, 1985). Such a
functional diffentiation could principally be used to re-distribute reflex effects to different
motor unit combinations depending on peripheral circumstances; for example, inhomogene-
ous distribution of fatigue in muscles with different motor unit compositions. But again, this
252 U. Windhorst and G. Boorman

Figure 4. Convergence of segmental afferent and

spinally descending inputs onto lumbar group Ib
inhibitory interneurons. Note that, as mentioned
before, only about 40% of cat Ib inhibitory in-
terneurons receive excitatory Ia afferent input
(Jankowska et aI., 1981). There are, therefore, sub-
groupii ~ populations ofIb interneurons with different input
patterns (Jankowska 1992). Dashed arrowheads,
HTA • mixed excitatory and inhibitory effects. Dashed
arrowshafts, oligo- or polysynaptic pathways. Ab-

c"~7 ~r- t\
breviations: CS, corticospinal; Cut, cutaneous;
HTA, high-threshold muscle and joint afferents;
d ReS
ProS, propriospinal; d ReS, dorsal reticulospinal;
-<
Join! '\..- maReS
rna ReS, monaminergic reticulospinal; RuS,
rubrospinal; VS, vestibulospinal (modified from
ProS RuS Baldissera et aI., 1981; Harrison & Jankowska,
CS 1985).

has yet to be studied. Similar functional fractionation may apply to other interneuron pools,
such as reciprocal Ia inhibitory interneurons (less pronounced: E. lankowska, personal
communication) and interneurons mediating presynaptic inhibition.

PRESYNAPTIC INHIBITION

An additional complication making predictions difficult results from the fact that the
synaptic efficacy of spindle group Ia and II, and tendon organ Ib afferents can be altered via
presynaptic inhibition elicited by activity in the same groups of afferents, by high-threshold
muscle afferents, by signals in descending tracts, and by the ePGs (see below) (Jankowska,
1992). Since descending signals and muscle afferent feedback change during fatiguing
muscle contractions, so too should signal transmission from some afferent classes to spinal
networks. Presynaptic inhibition generally has a long duration of the order of a few hundreds
of milliseconds. One of the major advantages ofpre-vs. postsynaptic inhibition is its potential
for separate control of subsets of synaptic inputs to a neuron, but this possibility is not to be
always exploited. For example, GABAergic interneurons can simultaneously exert presy-
naptic inhibition of primary afferents and postsynaptic inhibition ofmotoneurons (Rudomin,
1990; lankowska, 1992); therefore, irrespective of their effect on synaptic transmission,
last-order presynaptic inhibitory interneurons provide for yet another postsynaptic inhibitory
pathway to motoneurons from muscle afferents affected by muscle fatigue.
The pattern of presynaptic inhibition is complex. Group Ia and II muscle spindle
afferents and group Ib Golgi tendon organ afferents appear to be presynaptically inhibited
by separate populations of interneurons with slightly different input patterns (Rudomin,
1990; lankowska, 1992). As with Ib inhibitory interneurons, this arrangement may generate
functional fractionation that could be used to advantage in re-shaping spatio-temporal motor
activity during fatigue.

RECIPROCAL INHIBITION

Thus far, emphasis has been placed on afferent effects from fatiguing muscle onto
synergistic motoneurons. However, reciprocal inhibition of antagonists (Fig. 2) will be
affected as well, because reciprocal Ia inhibitory interneurons receive virtually the same
Overview: Potential Role of Segmental Motor Circuitry in Muscle Fatigue 253

convergent inputs as do Ib inhibitory interneurons (Fig. 4), particularly from Ia afferent input
and high-threshold muscle afferent input. This may not matter when only agonist muscle
groups are undergoing fatiguing contractions as in most of the experimental paradigms
discussed above. However, when fatigue of alternating rhythmic movements such as loco-
motion is considered, there is no information about reciprocal inhibition, or other networks,
including those generating the basic rhythmic pattern.

LOCOMOTION

It is everyday knowledge that muscle fatigue and its attendant problems arise during
rhythmic movements, such as locomotion and mastication. Their basic patterns are generated
actively by the CNS by way ofCPGs. Even the isolated lumbosacral spinal cord is capable,
upon proper pharmacological pretreatment, to produce quite elaborate motor output patterns
which are quite akin to normal ones. This applies even if movement and hence sensory
afferent feedback is abolished by paralysis (fictive locomotion; see below). Despite intense
efforts, the network constituting each CPG has not yet been identified in mammals and,
therefore, no data are available on the effects of sensory feedback from fatiguing muscle.
Motoneurons, Renshaw cells and reciprocal Ia inhibitory interneurons (Fig. 2) are
presumably included in the CPG, their interactions including mutual inhibition and many
other control components found in lower animals' CPGs (Getting, 1989); cf., Pratt & Jordan,
1987; Gelfand et aI., 1988). The discharge patterns of Renshaw cells and Ia inhibitory
interneurons are rhythmically modulated during locomotion, and they are also subject to
effects from high-threshold muscle afferents, which are responsive to muscle fatigue (for
details, see Garland & Kaufman, Chapter 19). In 1912, Graham-Brown postulated that two
tonically excited half-centers of interneurons, each driving flexors or extensors at the same
joint, would produce the alternating locomotor pattern by reciprocally inhibiting each other.
Candidates for Graham-Brown's half-centers might be the interneurons which mediate the
flexion reflex response to noxious and other low-threshold stimuli (Jankowska et aI., 1967).
Like the half-centers, these interneurons mutually inhibit each other Jankowska et aI.,
1967a,b and, undoubtedly, receive input from muscle afferents excited during muscle
fatigue. If these interneurons can be demonstrated to be part of CPGs, then the CPGs,
themselves, should be directly affected by muscle fatigue.
Changes in environmental conditions, errors of posture and movement, and altera-
tions in internal system properties such as fatigue require that centrally generated patterns
be adapted on a short-term basis. CPGs and sensory feedback must interact rapidly, with
sensory information modulating CPG activity and, vice versa, with descending controls
superimposed, as well.
Sensory feedback plays important roles in timing locomotor activity and the transi-
tion between different phases of rhythmic movement, including reinforcement of ongoing
motor output and adjustment to environmental conditions (Pearson, 1993). For example, in
the presence of sensory feedback, cat locomotion is more normal and stable, less fatigable,
and more easily adjustable to a wider range of speeds (Grillner, 1985). But the mechanisms
for these effects are not known. Nonetheless, these effects must change during fatigue, either
by a primary change in discharge patterns, or by modulatory effects on intercalated in-
terneurons.
In the phase switch from extension to flexion, it is likely that cat mid-lumbar
interneurons are implicated. These receive monosynaptic excitatory input from group II
muscle spindle (and other) afferents from hip muscles (Jankowska, 1992). Another set of
interneurons which may be involved in switching from extension to flexion are those
mediating reciprocal inhibition. During human walking, the ankle dorsi flexors are stretched
254 U. Windhorst and G. Boorman

in the late stance phase; (i.e., between heel off and toe off), when their Ia muscle spindle
afferents are excited and thereby help activate their own motoneurons as well as reciprocally
inhibit the antagonist soleus motoneurons (Capaday et al., 1990). A third interneuronal
system of potential importance involves those mediating reflex effects from Golgi tendon
organ afferents. Whereas in quiescent preparations, extensor Ib afferents exert widespread
inhibition on leg extensor motoneurons (Baldissera et al., 1981), this effect is reduced or
abolished during locomotion. Instead, extensor Ib afferents activate extensor motoneurons
and inhibit ipsilateral flexor motoneurons, effects that vary in size according to the phase of
the locomotor step cycle (Pearson & Collins, 1993; Gossard et al., 1994). Such declining
extensor force and Ib discharge in the late stance phase may release flexors from inhibition
and facilitate initiation of the swing phase (Pearson, 1993). Since muscle fatigue affects all
the above afferent systems, it should disturb locomotor timing, as, for example, when fatigue
makes walking less stable and secure.
The locomotor evidence suggests that its CPGs influence the sign and strength of
various reflexes as a means of selecting sensory information that is appropriate for the phase
of the particular step (Dietz, 1992; Pearson, 1993). Substantial evidence is on hand that
reflexes can be modulated in size, or even reversed in effect, as dependent on the phase of
the step. (Forssberg et al., 1977; Yang & Stein, 1990). Such phase-dependent reflex control
and reversal is illustrated in Fig. 5. It has three features at the segmental level (i.e.,
suprasegmental effects are not shown).
1. The first effect (presynaptic inhibition, 1 in Fig. 5) occurs at sites immediately
after the entrance of sensory afferents into the spinal cord. Its strength varies as
the locomotor phase (Dubuc et al., 1988; Duenas et al., 1990; Gossard & Rossig-
nol, 1990). Such phase control is evident in even the presynaptic inhibitory
modulation of the monosynaptic Ia afferent-motoneuronal connection.
2. The CPG may modulate any premotor interneuron intercalated between primary
afferents and motoneurons (e.g., first and last order, 2 and 3, respectively in Fig.
5), including Renshaw cells and reciprocal Ia inhibitory interneurons. Since the
latter receive convergent input from many sensory afferents, transmission of this
information to motoneurons is phasically gated. Furthermore, autogenetic Ib
inhibition from extensor muscles may be depressed and replaced with excitation
during locomotion. It is also appropriate to consider y-motoneurons as premo-
toneurons, insofar as they integrate many segmental afferent and descending
influences, and transmit such influences via muscle spindles and their afferents to
u-motoneurons and back onto themselves.
3. Motoneurons and interneurons receive locomotor drive potentials from the CPG.
During the hyperpolarizing phases of these potentials, which are caused in part
by inhibitory inputs from the CPG, excitatory sensory inputs are effectively
shunted. Thus, the motoneurons themselves contribute to locomotor reflex modu-
lation (4 in Fig. 5).
In summary, reflex modulation plays an important role in normal locomotor activity
and the mechanisms involved are strongly affected by muscle fatigue.

SUMMARY ON THE SPINAL SEGMENTAL REGULATION OF

MOTOR UNIT FIRING RATE

The CNS appears to have mechanisms available for mitigating the effects of muscle
fatigue, at lease in the short term. Many of them shape the firing patterns of motor units in
Overview: Potential Role of Segmental Motor Circuitry in Muscle Fatigue 255

Figure 5. Scheme of the possible sites of modulation of

reflex transmission by the locomotor CPO. The sites in-
clude: 1, terminals of primary sensory afferents (for presy-
naptic inhibition); 2, inhibition of the first-order
interneuron; 3, postsynaptic excitation and inhibition of the
last-order interneurons which connect monosynaptically
(site 4) with motoneurons. Effects of muscle afferents on the
CPO are also indicated (modified from Sillar, 1991; his Sensory
Fig. 1). Muscle
input

order to optimize their force output. It would make physiological sense that these mecha-
nisms primarily target fast-fatigable motor units (Stuart & Callister, 1993). However, these
mechanisms are multifarious as to their location and principles of operation, and they
understandably cooperate in a complex way. To umavel these mechanisms is an enormous
task for experimenters. Of special immediate concern is to measure signal transmission in
spinal interneurons and motoneurons and the modulating effects exerted by those of the
muscle afferents that are affected by muscle fatigue. The elements and networks requiring
further detailed study include: 1) a-motoneurons; 2) y-motoneurons; 3) recurrent inhibition
via Renshaw cells; 4) Ib inhibitory interneurons; 5) reciprocal Ia inhibitory interneurons; 6)
interneurons mediating presynaptic inhibition; and 7) other interneurons whose effects on
motoneurons, such as those possibly involved in locomotor pattern generation. This list is
by no means exhaustive. As such work is undertaken, theoretical concepts and computer
models will need to be developed in parallel, in order to integrate experimental data.

ACKNOWLEDGMENTS

This work was supported by the Medical Research Council of Canada (u. W.), and
the Alberta Heritage Foundation for Medical Research, Canada. G.B. held a postdoctoral
fellowship from the Alberta Paraplegic Foundation, Canada. Attendance of U. W. at the 1994
Bigland-Ritchie conference was supported, in part, by the University of Miami.

REFERENCES

Baldissera F, Hultborn H & Illert M (1981). Integration in spinal neuronal systems. In: Brooks VB (ed.),
Handbook of Physiology, Vol. II, Part 1, The Nervous System, pp. 509-595. Bethesda: American
Physiological Society.
Bevan L, Laouris Y, Reinking RM & Stuart DO (1992). The effect of the stimulation pattern on the fatigue of
single motor units in adult cats. Journal of Physiology (London) 449, 85-108.
 256 U. Windhorst and G. Boorman

Bigland-Ritchie BR (1993). Regulation of motorneuron fIring rates in fatigue. In: Sargeant AJ, Kernell D
(eds.), Neuromuscular Fatigue, pp 147-155. Amsterdam: Royal Netherlands Academy of Arts and
Sciences.
Bigland-Ritchie B, Dawson NJ, Johansson RS & Lippold OCJ (1986). Reflex origin for the slowing of
motoneurone fIring rates in fatigue of human voluntary contractions. Journal o/Physiology (London)
379,451-459.
Bigland-Ritchie B, Johansson RS, Lippold OCJ & Woods JJ (1983a). Changes in motoneurone fIring rates
during sustained maximal voluntary contractions. Journal 0/ Physiology (London) 340, 335-346.
Bigland-Ritchie B, Johansson RS, Lippold OCJ & Woods JJ (l983b). Contractile speed and EMG changes
during fatigue of sustained maximal voluntary contractions. Journal o/Neurophysiology 50, 313-324.
Bigland-Ritchie B & Woods JJ (1984). Changes in muscle contractile properties and neural control during
human muscular fatigue. Muscle & Nerve 7, 691-699.
Binder MD & Mendell LM (eds.) (1990). The Segmental Motor System. New York: Oxford University Press
Binder-Macleod SA & Barker CB (1991). Use of a catch-like property of human skeletal muscle to reduce
fatigue. Muscle & Nerve 14, 850-857.
Binder-Macleod SA & Guerin T (1990). Preservation of force output through progressive reduction of
stimulation frequency in human quadriceps femoris muscle. Physical Therapy 70. 619-625.
Bottennan BR & Cope TC (1988). Motor-unit stimulation patterns during fatiguing contractions of constant
tension. Journal o/Neurophysiology 60, 1198-1214.
Brownstone RM, Jordan LM, Kriellaars DJ, Noga BR & Shefchyk SJ (1992). On the regulation of repetitive
fIring in lumbar motoneurones during fIctive locomotion in the cat. Experimental Brain Research 90,
441-455.
Burke RE (1981). Motor units: anatomy, physiology, and functional organization. In: Brooks VB (ed.),
Handbook 0/ Physiology, Vol. IL Part 1, The Nervous System, pp 354-422. Bethesda: American
Physiological Society.
Burke RE, Levine DN, Tsairis P & Zajac FE (1973). Physiological types and histochemical profIles in motor
units of the cat gastrocnemius. Journal 0/ Physiology (London) 234, 723-748.
Capaday C, Cody FWJ & Stein RB (1990). Reciprocal inhibition of soleus motor output in humans during
walking and voluntary tonic activity. Journal o/Neurophysiology 64,607- 616.
Clamann HP (1990). Changes that occur in motor units during activity. In: Binder MD, Mendell LM (eds.),
The Segmental Motor System, pp. 239-257. New York: Oxford University Press.
Dietz V (1992). Human neuronal control of automatic functional movements: interaction between central
programs and afferent input. PhYSiological Reviews 72, 33-69.
Dubose L, Schelhorn TB & Clamann HP (1987). Changes in contractile speed of cat motor units during activity.
Muscle & Nerve 10, 744-752.
Dubuc R, Cabelguen J-M & Rossignol S (1988). Rhythmic fluctuations of dorsal root potentials and antidromic
discharges of primary afferents during fIctive locomotion in the cat. Journal o/Neurophysiology 60,
2014-2036.
Dueiias SH, Loeb GE & Marks WB (1990). Monosynaptic and dorsal root reflexes during locomotion in nonnal
and thalamic cats. Journal o/Neurophysiology 63, 1467-1476.
Edwards RHT, Wiles CM, Gohil K, Krywawych S & Jones DA (1982) Energy metabolism in human myopathy.
In: Schotland DL (ed.), Disorders o/the Motor Unit, pp. 715-726. New York: Wiley.
Enoka RM & Stuart DG (1992). Neurobiology of muscle fatigue. Journal 0/ Applied Physiology 72,
1631-1648.
Fitch S & McComas A (1985). Influence of human muscle length on fatigue. Journal 0/Physiology (London)
362,205-213.
Forssberg H, Grillner S & Rossignol S (1977). Phasic gain control of reflexes from the dorsum of the paw
during spinal locomotion. Brain Research 132,121-139.
Fuglevand AJ, Zackowski KM, Huey KA & Enoka RM (1993). Impainnent of neuromuscular propagation
during human fatiguing contractions at submaximal forces. Journal 0/ PhYSiology (London) 260,
549-572.
Gandevia SC (1993). Central and peripheral components to human isometric muscle fatigue. In: Sargeant AJ,
Kernell D (eds.), Neuromuscular Fatigue, pp. 156-164. Amsterdam: Royal Netherlands Academy of
Arts and Sciences.
Gandevia SC, MacefIeld G, Burke D & McKenzie DK (1990). Voluntary activation of human motor axons in
the absence of muscle afferent feedback. The control of the deafferented hand. Brain 113, 1563-1581.
Gandevia SC, MacefIeld VG, Bigland-Ritchie B, Gonnan R & Burke D (1993). Motoneuronal output and
gradation of effort in attempts to contract acutely paralyzed leg muscles in man. Journal o/Physiology
(London) 474, 411-427.
Overview: Potential Role of Segmental Motor Circnitry in Muscle Fatigue 257

Garland SJ, Gamer SH & McComas AJ (1988). Reduced voluntary electromyographic activity after fatiguing
stimulation ofhurnan muscle. Journal of Physiology (London) 401, 547-556.
Gelfand 1M, Orlovsky GN & Shik ML (1988). Locomotion and scratching in tetrapods. In: Cohen AH,
Rossignol S, Grillner S (eds.), Neural Control of Rhythmic Movements in Vertebrates, pp 167-199.
New York: Wiley.
Getting PA (1989). Emerging principles governing the operation of neural networks. Annual Review of
Neuroscience 12, 185-204.
Gordon DA, Enoka RM & Stuart DG (1990). Motor-unit force potentiation in adult cats during a standard
fatigue test. Journal of Physiology (London) 421, 569-582.
Gossard J-P, Brownstone RM, Barajon I & Hultborn H (1994). Transmission in a locomotor-related group Ib
pathway from hindlimb extensor muscles in the cat. Experimental Brain Research 98, 213-228.
Gossard J-P & Rossignol S (1990). Phase-dependent modulation of dorsal root potentials evoked by peripheral
nerve stimulation during fictive locomotion in the cat. Brain Research 537, 1-13.
Graham-Brown T (1912). The factors in rhythmic activity of the nervous system. Proceedings Royal Society
London B 85, 278-289.
Grillner S (1985). Neural control of vertebrate locomotion - central mechanisms and reflex interaction with
special reference to the cat. In: Barnes WJP, Gladden MH (eds.), Feedback and Motor Control in
Invertebrates and Vertebrates, pp. 35-56. London: Croom Helm.
Hamm TM, Sasaki SI, Stuart DG, Windhorst U & Yuan C-U (1987). Distribution of single-axon recurrent
inhibitory post-synaptic potentials in a single spinal motor nucleus in the cat. Journal of Physiology
(London) 388, 653-664.
Harrison PJ & Jankowska E (1985). Sources of input to interneurones mediating group I non- reciprocal
inhibition of motoneurones in the cat. Journal of Physiology (London) 361, 379-40 I.
Hayward L, Breitbach D & Rymer WZ (1988). Increased inhibitory effects on close synergists during muscle
fatigue in the decerebrate cat. Brain Research 440, 199-203.
Hayward L, Wesselmann U & Rymer WZ (1991). Effects of muscle fatigue on mechanically sensitive afferents
of slow conduction velocity in the cat triceps surae. Journal of Neurophysiology 65, 360-370.
Henneman E (1957). Relation between size of neurons and their susceptibility to discharge. Science 126,
1345-1346.
Houk JC & Rymer WZ (1981). Neural control of muscle length and tension. In: Brooks VB (ed.), Handbook
of Physiology, Vol II, Part 1, The Nervous System, pp. 257-323. Bethesda: American Physiological
Society.
Houk JC, Singer JJ & Goldman MR (1970). An evaluation oflength and force feedback to soleus muscles of
decerebrate cats. Journal ofNeurophysiology 33, 784-811.
Jami L (1992). Golgi tendon organs in mammalian skeletal muscle: functional properties and central actions.
Physiological Reviews 72, 623-666.
Jankowska E (1992). Interneuronal relay in spinal pathways from proprioceptors. Progress in Neurobiology
38,335-378.
Jankowska E, Johanisson T & Lipski J (1981). Common interneurons in reflex pathways from group Ia and
Ib afferents of ankle extensors in the cat. Journal of Physiology (London) 310, 381-402.
Jankowska E, Jukes MGM, Lund S & Lundberg A(1967a). The effect of DOPA on the spinal cord. 5. Reciprocal
organization of pathways transmitting excitatory action to alpha motoneurones of flexors and
extensors. Acta Physiologica Scandinavica 70, 369-388.
Jankowska E, Jukes MGM, Lund S & Lundberg A (1967b). The effect of DOPA on the spinal cord. 6.
Half-centre organization of interneurones transmitting effects from the flexor reflex afferents. Acta
Physiologica Scandinavica 70, 389-402.
Kernell D (1992). Organized variability in the neuromuscular system: a survey of task- related adaptations.
Archives Italiennes de Biologie 130, 19-66.
Kernell D (1993) Neuromuscular fatigue and the differentiation of motoneurone and muscle unit properties.
In: Sargeant AJ, Kernell D (eds.), Neuromuscular Fatigue, pp 139-146. Amsterdam: Royal Nether-
lands Academy of Arts and Sciences.
Kernell D & Monster AW (1982). Motoneurone properties and motor fatigue. An intracellular study of
gastrocnemius motoneurones of the cat. Experimental Brain Research 46, 197-204.
Kirsch RF & Rymer WZ (1992). Neural compensation for fatigue-induced changes in muscle stiffness during
perturbations of elbow angle in human. Journal ofNeurophysiology 68, 449-470.
Kukulka CG, Moore MA & Russell AG (1986). Changes in human a-motoneuron excitability during sustained
maximum isometric contractions. Neuroscience Letters 68, 327-333.
Lenman AJR, Tulley FM, Vrbova G, Dimitrijevic MR & Towle JA (1989). Muscle fatigue in some neurological
disorders. Muscle & Nerve 12, 938-942.
258 U. Windhorst and G. Boorman

Macefield G, Hagbartb K-E, Gorman R, Gandevia SC & Burke D (1991). Decline in spindle support to
u-motoneurones during sustained voluntary contractions. Journal 0/ Physiology (London) 440,
497-512.
Marsden CD, Meadows JC & Merton PA (1983). "Muscular wisdom" that minimizes fatigue during prolonged
effort in man: peak rates of motoneuron discharge and slowing of discharge during fatigue. In:
Desmedt JE (ed.), Motor Control Mechanisms in Health and Disease, pp. 169-211. New York: Raven
Press.
McKenzie DK, Bigland-Ritchie B, Gorman RB & Gandevia SC (1992). Central and peripheral fatigue of
human diaphragm and limb muscles assessed by twitch interpolation. Journal 0/Physiology (London)
454, 643-656.
Mense S (1986). Slowly conducting afferent fibers from deep tissues - neurobiological properties and central
nervous actions. In: Ottoson D (ed.), Progress in Sensory Physiology, Vol. 6, pp. l39-219. Berlin:
Springer-Verlag.
Miller RG, Green AT, Moussavi RS, Carson PJ & Weiner MW (1990) Excessive muscular fatigue in patients
with spastic paraparesis. Neurology 40, 1271-1274
Pearson KG (1993) Common principles of motor control in vertebrates and invertebrates. Annual Review 0/
Neuroscience 16, 265-297.
Pearson KG & Collins DF (1993). Reversal of the influence of group Ib afferents from plantaris on activity in
medial gastrocnemius muscle during locomotor activity. Journal 0/Neurophysiology 70, 1009-1017.
Powers RK & Binder MD (1991). Effects oflow-frequency stimulation on the tension- frequency relations of
fast-twitch motor units in the cat. Journal o/Neurophysiology 66, 905-918.
Pratt CA & Jordan LM (1987). Ia inhibitory interneurons and Renshaw cells as contributors to the spinal
mechanisms of fictive locomotion. Journal o/Neurophysiology 57,56-71.
Rotto DM & Kaufmaun MP (1988). Effect of metabolic products of muscular contraction on discharge of
group III and IV afferents. Journal 0/Applied Physiology 64, 2306-23l3.
Rudomin P (1990). Presynaptic inhibition of muscle spindle and tendon organ afferents in the mammalian
spinal cord. Trends in Neurosciences 13, 499-505.
Sillar KT (1991) Spinal pattern generation and sensory gating mechanisms. Current Opinion in Neurobiology
1,583-589.
Sinoway LI, Hill, JM, Pickar, JG & Kaufman, MP (1993). Effects of contraction and lactic acid on the discharge
of group III muscle afferents in cats. Journal o/Neurophysiology 69, 1053-1059.
Spielmaun JM, 'Laouris Y, Nordstrom MA, Robinson GA, Reinking RM & Stuart DG (1993). Adaptation of
cat motoneurons to sustained and intermittent extracellular activation. Journal 0/ Physiology (Lon-
don) 464, 75-120.
Stuart DG & Callister RJ (1993). Afferent and spinal reflex aspects of muscle fatigue: issues and speculations.
In: Sargeant AJ, Kernell D (eds), Neuromuscular Fatigue, pp. 169- 180. Amsterdam: Royal Nether-
lands Academy of Arts and Sciences.
Vl:lllestad NK, Sejersted OM, Bahr R, Woods JJ & Bigland-Ritchie B (1988). Motor drive and metabolic
responses during repeated submaximal voluntary contractions in man. Journal 0/Applied PhYSiology
64,1421-1427.
Windhorst U (1988) How Brain-like is the Spinal Cord? Interacting Cell Assemblies in the Spinal Cord. Berlin:
Springer-Verlag.
Windhorst U & Kokkoroyiannis T (1991). Interaction of recurrent inhibitory and muscle spindle afferent
feedback during muscle fatigue. Neuroscience 43, 249-259.
Woods JJ, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal o/Neurophysiology 58, 125-l37.
Yang JF & Stein RB (1990). Phase-dependent reflex reversal in human leg muscles during walking. Journal
0/Neurophysiology 63, 11 09-1117.
 18

THE FUSIMOTOR SYSTEM

Its Role in Fatigue

K-E. Hagbarth' and V. G. Macefield2

, Department of Clinical Neurophysiology

University Hospital
S-751 85 Uppsala
Sweden
2 Prince of Wales Medical Research Institute
Randwick, Sydney 2031
Australia

ABSTRACT

Several lines of evidence point to an important role of the fusimotor system in the "muscle-
wisdom" phenomenon during peripheral fatigue of some human voluntary contractions: I)
muscle afferents provide a net amplification of skeletomotor output, with the only known
afferent species capable of this being the muscle spindle; 2) muscle spindle firing rates
decline during constant-force voluntary contractions, so fusimotor support to skeletomotor
output decreases; 3) this waning support can be offset by application of high-frequency
vibration to the fatiguing muscle, which excites spindle endings; and 4) the progressive
decline in motor unit firing rates during maximal voluntary contractions is abolished by
blocking muscle afferent inputs, and it is argued that, at least in the initial stages of a
contraction, this must be due to a progressive withdrawal of spindle support.

INTRODUCTION

The functional role of the fusimotor system in muscle fatigue has been the subject
of debate for several decades. This is understandable given the potent excitatory inputs
that spindle afferents provide to alpha motoneurons and hence to force production. Yet
the roles of muscle spindles in motor control have varied with the fashions of the time,
and so fatigue theories have had to adapt to encompass the prevailing views on fusimotor
function. Therefore, before dealing with the issue of muscular fatigue it is worth reviewing
the shifts in these basic hypotheses and the controversies that still exist in this field of
research.

259
260 K-E. Hagbarth and V. G. Macefield

BASIC HYPOTHESES ON FUSIMOTOR INVOLVEMENT IN

MOTOR CONTROL

The lengthfollow-up servo theory from the 1950's was largely based on experiments
performed on decerebrate cats. In this theory, an important role was attributed to the gamma
motoneurons, regarded at the time as the initial elements in a servo-controlled system which,
via the muscle spindle endings and their stretch reflex afferents, drive the skeletomotor
output to achieve desired muscle length in voluntary contractions (Eldred et aI., 1953;
Merton, 1953). Although this was an attractive theory, subsequent human-based studies and
experiments in awake and alert animals led to its dismissal (see Matthews, 1981). Partly as
a result of microneurographic studies of muscle spindle activity in man, it has become clear
that the skeletomotor output during voluntary contractions is not servo-driven, but rather
servo-assisted by fusimotor-induced activity in muscle spindle afferents. This widely ac-
cepted servo-assistance theory implies that during voluntary contractions, the fusimotor-bi-
ased spindle endings provide not only a tonic autogenetic reflex support to the skeletomotor
output (combined with a tonic inhibitory influence on antagonistic motoneurons) but by
virtue of the high dynamic sensitivity of primary endings to stretch, the fusimotor-biased
spindle endings can also exert a reflex influence on the precise timing of motor unit
discharges and, thereby, provide reflex compensation for unexpected perturbations.
Much interest has been directed to factors which may influence the efficacy of this
servo-assistance and the gain of the stretch reflex. Such factors may include 1) the relative
strength of the fusimotor output; 2) the extent to which spindle unloading during concentric
contractions counteracts the fusimotor-induced internal activation of spindle endings; 3) the
extent to which the history of movements and contractions affects the inherent slackness of
intrafusal muscle fibers; 4) variations in pre-motoneuronal excitability of segmental and
supraspinal stretch reflex arcs; and 5) the level of excitability of alpha motoneurons.
Studies in alert cats indicate that the participation of the fusimotor system in motor
performance is task-dependent; i.e., fusimotor outputs mayor may not be accompanied by
concurrent skeletomotor outputs, depending on the motor-set as determined by the task to
be performed (Loeb & Hoffer, 1981; Prochazka & Wand, 1981). By contrast, human
microneurography studies have so far not shown any consistent signs of such task-dependent
dissociations between skeletomotor and fusimotor outputs (Vallbo et aI., 1979; Burke, 1981;
Hagbarth, 1993; but see Aniss et aI., 1990). As judged by these studies, factors 4) and 5)
must be awarded the primary roles in task-dependent variations in fusimotor-driven servo
assistance of skeletomotor output.

HYPOTHESES ON FUSIMOTOR INVOLVEMENT IN MUSCLE

FATIGUE

As exemplified by the above discussion, theories based on animal experiments do

not always agree with those derived from studies in humans. Such discrepancies are
particularly evident when considering theories dealing with involvement of the fusimotor
system in muscle fatigue. Conflicting interpretations may partly be explained by the diversity
of definitions of muscle fatigue, partly by the variety of experimental paradigms. If muscle
fatigue is defined simply as " ... any reduction in the force- or velocity-generating capacity
of a muscle that is alleviated by rest" (Gandevia et aI., 1992), then, strictly speaking, fatigue
sets in at the start of a contraction, not just at the end. In the following, we will, therefore,
deal not only with that later stages of fatigue, evidenced by a reduced force output, but also
Fusimotor System 261

with the earlier stages when a gradually increasing effort is required to maintain a constant
force.
While the length follow-up servo theory was still in vogue, the fusimotor system was
envisaged as the initiator of a powerful servo which provided stretch reflex compensation
for a decline in force during fatigue of extrafusal muscle fibers (Merton, 1954). That was
before it was realized that, rather than an increase, a decrease in motor unit firing rates is
actually required to optimize the force output of fatiguing muscles (Bigland-Ritchie et aI.,
1979; Marsden et aI., 1983; Bigland-Ritchie et aI., 1983; Binder-Macleod & Guerin, 1990).
There is now convincing evidence that as extrafusal contractile fatigue develops during a
sustained maximal voluntary contraction (MVC), there is a concurrent decline in motor unit
firing rates (Grimby et aI., 1981; Marsden et aI., 1983; Bigland-Ritchie et aI., 1983;
Bigland-Ritchie & Woods, 1984; Bigland-Ritchie et aI., 1986; Woods et aI., 1987; Gandevia
et aI., 1990). This decline has been regarded as an expression of muscle wisdom, since it
apparently serves to ensure appropriate economical activation offatiguing muscle (Marsden
et aI., 1983): as twitch relaxation times increase during fatigue, the fusion frequency
decreases such that lower firing rates are required for individual motor units to generate their
maximal force (Bigland-Ritchie & Woods, 1984). In sustained submaximal voluntary
contractions when increasing effort is required to maintain a desired force, the whole muscle
EMG activity increases even though some motor units may show declines in firing rate
(Bigland-Ritchie et aI., 1986). The increase in EMG, which reflects progressive recruitment
of new motor units, typically continues until the desired force can no longer be maintained,
at which point motor unit firing rates decline largely in parallel with the force even though
signs of motor unit rotation can be seen (Person, 1974). While recognizing that a reduced
voluntary drive will necessarily contribute to a reduction in motor unit firing rates during a
sustained voluntary contraction, the subsequent discussion will consider only segmental
contributions to muscle fatigue; the phenomenon of central fatigue will be addressed in later
Chapters 20-23.
The mechanisms underlying muscle wisdom are controversial, and have been dis-
cussed by Windhorst and Boorman, Chapter 17). Here, it suffices to say that the two leading
explanations for the phenomenon are based on a changing reflex input to the motoneuron
pool:
1. As a fatiguing contraction proceeds, there is a gradual dis facilitation of alpha
motoneurons due to a progressive withdrawal offusimotor-driven spindle support
to the skeletomotor output. This theory derives its main support from experiments
on conscious humans trying to maintain a desired voluntary force. The most direct
evidence comes from microneurographic recordings of impulse activity in iden-
tified muscle afferents and motor axons in human subjects. There are distinct
advantages to studying muscle fatigue in humans: activation of muscle volition-
ally will access segmental (and suprasegmental) reflex circuitry in a physiological
manner (e.g., Hultbom et aI., 1987; Baldissera & Pierrot-Deseilligny, 1989;
Meunier & Pierrot-Deseilligny, 1989; Burke et aI., 1992), and motoneurons will
fire with an intrinsic variability that may in itself optimize the production of force
(see Bevan et aI., 1992).
2. The alpha motoneurons are inhibited by group III and IV muscle afferents,
activated by the accumulation of metabolites in the fatiguing muscle. The most
direct evidence for this theory comes from experiments on reduced animals, in
which muscle fatigue is induced not volitionally but by electrical nerve stimula-
tion (Kniftki et aI., 1981; Cleland et aI., 1982). This idea has received indirect
support from experiments in human subjects (Bigland-Ritchie et aI., 1986;
Kukulka et aI., 1986; Woods et aI., 1987; Garland et aI., 1988; Garland &
262 K-E. Hagbarth and V. G. Macefield

McComas, 1990; Garland, 1991; see Garland & Kaufman, Chapter 19). However,
the studies supporting the alternative disfacilitation theory stand on their own,
and the two theories need not be considered mutually exclusive. Both sources of
peripheral input to the motoneuron pool, large-diameter as well as small-diameter
intramuscular afferents, could contribute to the decline in motoneuronal discharge
during muscular fatigue, but their relative contributions may well be expected to
differ during the development of fatigue.

MUSCLE AFFERENT FEEDBACK PROVIDES A NET

AMPLIFICATION OF SKELETOMOTOR OUTPUT

Irrefutable evidence indicates that muscle afferent feedback causes a net facilitation
of volitionally generated skeletomotor output in human subjects:
1. Peak motor unit firing rates during brief MVC's of nonfatigued muscles are
reduced by partial nerve blocks following perineural injection oflocal anesthetic,
which preferentially blocks small-diameter axons before effects on large-diameter
fibers can be observed (Fig. I). This was taken as evidence that normally activity
in such fibers (efferent and/or afferent) has a net facilitatory effect on skeletomotor
output (Hagbarth et ai., 1986; Bongiovanni et ai., 1990; Bongiovanni & Hagbarth,
1990). It was also found that the decline in motor unit firing rates normally seen
during sustained MVCs was absent during the partial nerve blocks, indicating that
with intact innervation the net facilitatory effect of small-fiber activity gradually
fades away as muscle fatigue develops. Muscle vibration, known to be a potent
stimulus for primary spindle endings, was used to test the hypothesis that the
normal fatigue-induced decline in skeletomotor output primarily depends on a
decline in fusimotor-driven spindle activity. This hypothesis was supported by the
findings that short periods « lOs) of muscle vibration counteracted both the
fatigue-induced declines in motor unit firing rates during sustained MVCs and the
reductions in peak motor unit firing rates induced by partial nerve blocks.

A motor unit 1

40 [
Hz 20
~.,,~. "
-::""-'-"'~--'7-"t;t'o-A:i.~-------------'
••,.., ~~
to'" .. .• '. 't •..,.~. '". ,.v/• lie .. , . I.... ..
o . . . .o'f
.o.o
• •••• ,'1 ....... . , . -... : ••••

3 min 3.5 min 4 min into block

.-----,
2s

B motor unit 2
, ..' ,
40 [ •
.....;,.;....",..:'~:"'"
, (i' •.•
.,..,.:,
• .....'1,..... ('.
....,
'I •••
••
- ..........;. ___
o .:. . . - . . . ,.---. ". .
'
Hz 20 ---.-~:.....-''''\...J.,~--~_._ ;-..,.,.,.!..~_
..0 ....
v.~.-,\ I .1, •••,~.
:.. I, ,:
I '¥~.:
:.:: .... :Ie\.' .'-: ..

4 min 4.5 min 5 min into block

,----,
2s
Figure 1. Instantaneous frequency plots of two single motor units recorded from tibialis anterior during
maximal voluntary contractions performed during the development of paralysis following injection of local
anesthetic around the peroneal nerve (modified from Hagbarth et aI., 1986).
Fusimotor System 263

1.

0.8 deafferented

Cumulative 0.6
probability
0.4 motoneurons

0.2

10 20 30 40 50
Mean discharge frequency (Hz)
Figure 2. Firing rate distributions for tibialis anterior motoneurons from minimal to maximal levels of
voluntary drive. Deafferented motoneurons, data from 39 single motor axons recorded proximal to a complete
block of the peroneal nerve. Intact motoneurons, data from 595 single motor units. At all levels of voluntary
drive, firing rates were lower in the absence of muscle afferent feedback (modified from MacefIeld et aI.,
1993).

However, longer periods of high-frequency muscle vibration had opposite effects,

accentuating the fatigue-related declines in skeletomotor output. This effect was
tentatively ascribed to Ia presynaptic inhibition or exhaustion of Ia excitatory
synapses (Bongiovanni & Hagbarth, 1990).
2. Efferent activity in human motor axons has been recorded from the ulnar and peroneal
nerves during acute deafferentation of the target muscles by complete anesthetic block
of nerve conduction distal to the recording site (Gandevia et aI., 1990; Gandevia et aI.,
1992; Gandevia et aI., 1993; Macefield et aI., 1993). In the absence of muscle afferent
feedback, subj ects were able to recruit single motoneurons and to grade their discharge
frequency, but in attempted maximal efforts the firing rates of single motor axons were
lower than those of normally innervated motor units recorded in separate experiments.
Firing rates were reduced by approximately one-third at all levels of volitional drive,
from minimal to maximal efforts (Macefie1d et aI., 1993). The overall range offiring
rates expressed by intact and deafferented motoneurons was the same, indicating that
muscle afferent feedback provides a net amplification of skeletomotor output from the
spinal cord without affecting the dynamic range over which the alpha motoneurons
discharge vo1itionally. This is illustrated in Fig. 2. These experiments also showed that
in the complete absence of muscle afferent feedback, individual motoneurons did not
show the progressive decline ·in firing rates exhibited by normally-innervated motor
units during sustained maximal efforts, supporting the reflexogenic nature of the
decline in motoneuron discharge frequency during muscle fatigue. These results, and
those outlined in the preceding paragraph, are consistent with the view that the reflex
facilitatory influence provided by muscle afferents declines as fatigue develops
(Gandevia et al., 1990; Gandevia et aI., 1992; Gandevia et aI., 1993; Macefield et aI.,
1993).

MUSCLE SPINDLE AFFERENT FEEDBACK DOES DECLINE

DURING FATIGUE

Direct proof that muscle spindle feedback does decline during a constant-force
voluntary contraction came from microneurographic recordings of the discharge of single
264 K-E. Hagbarth and V. G. Macefield

N 25.0 >. 1.0

::s A
0
c::
OJ
G
c:
22.5 ~0.9
OJ .....
as 20.0
:::>
1;'00.8
.t: lil
.d
OJ
~ 17.5 .~ 0.7
.<: "0
0 OJ
.~
"0 15.0 .~ 0.6
~
e
12.5 0.5

~
10.0 0.4

7.5 0.3

5.0 0.2

2.5 0.1

0 0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
time (s) time (s)

Figure 3. Decline in discharge frequency of 11 muscle spindles in tibialis anterior as a function of initial firing
rate at the onset of an isometric voluntary contraction. The data were divided into units with initial frequencies
above 15 Hz (closed symbols) and below 15 Hz (open symbols). The data are expressed as absolute changes
inA and relative changes in B (from MacefIeld et ai., 1991).

muscle spindle afferents from tibialis anterior and extensor digitorum longus (Macefield et
aI., 1991). The discharge of most muscle spindle afferents decreased by some 50% during
contractions of the parent muscles sustained for at least 60 s, when increasing effort was
required to maintain the isometric force at the target level. This reduction began within the
first few seconds of the contraction, even before the target force had been reached, and
continued progressively while the surface EMG activity of the muscle gradually increased.
The decline in spindle discharge frequency was more prominent with contractions at higher
force levels, when firing rates were higher, and for contractions performed later in the
experiment after many sustained contractions. Significantly, the decline in firing rates of
single motor units during maximal voluntary contractions is also greater for units discharging
at higher initial frequencies (Bigland-Ritchie et aI., 1983). The decline in spindle discharge
during a contraction is shown in Fig. 3. For technical reasons, it has so far not been possible
to record muscle spindle activity in man during maximal voluntary contractions, but the
experiments described in the preceding paragraph indicate that fusimotor support is equally
important in maximal as well as submaximal voluntary contractions.
These studies indicate that muscle wisdom can at least partly be explained by a
progressive withdrawal of tonic support via the fusimotor loop, but the mechanisms respon-
sible for this waning input remain conjectural. Possibly, muscle ischemia with accompanying
accumulation of metabolites produces failure in excitation-contraction coupling or failure
in the contractile capacity of intrafusal muscle fibers. The observed decline in afferent
spindle discharge might also be explained by a declining fusimotor output or by adaptation
processes in muscle spindle receptors. In addition, central processes such as presynaptic
inhibition of Ia terminals may contribute to a reduction in fusimotor assistance as muscle
fatigue develops during sustained voluntary contractions.
Fusimotor System 265

EFFECTS OF MUSCLE FATIGUE ON SPINDLE-MEDIATED

REFLEX ASSISTANCE

Seemingly conflicting reports have been presented concerning fatigue-induced

changes in the efficacy of stretch and unloading reflexes (Darling & Hayes, 1983; Hunter &
Kearney, 1983; Hakkinen & Komi, 1983; Windhorst et aI., 1986; Kirsch & Rymer, 1987,
1992; Balestra et aI., 1992; Duchateau & Hainaut, 1993). One possible explanation for these
discrepancies is that some investigators have been mainly concerned with reflex changes
occurring during submaximal constant-force contractions, when there is a gradual overall
increase in EMG activity, whereas others have paid more attention to the later stages of
fatigue when skeletomotor output declines. Considering the principle of automatic gain
compensation (Matthews, 1986), by which the relative amplitude of reflex modulation
matches the ongoing level ofEMG, it is to be expected that with the increase in EMG activity
during sustained, submaximal constant-force contractions, there will be a corresponding
increase in EMG recorded in phasic stretch and unloading reflexes. Whereas, reductions of
reflex EMG responses are likely to occur later when the overall decline in motor unit firing
rates can no longer be prevented. In agreement with this principle, some investigators have
observed early enhancements and late reductions of reflex responses to muscle perturbations
during sustained isometric contractions (Marsden et aI., 1983; Darling & Hayes, 1983).
In order to avoid variations in reflex efficacy which are due to variations in the
pre-existing level of skeletomotor activity, it is essential that this level is similar in control
(non-fatigued) conditions as when the reflexes are tested during the different stages of
fatigue. In recent experiments where this criterion was fulfilled, both stretch and unloading
reflexes (especially their short-latency components) were found to be reduced at a stage of
fatigue just before the desired force could no longer be maintained. At this stage, there was
also a marked increase in the EMG activity of antagonistic muscles, a finding interpreted as
a sign of reduced Ia reciprocal inhibition (Hagbarth, in preparation).

DIRECT RECORDINGS FROM FUSIMOTOR NEURONS DURING

MUSCLE FATIGUE

Due to technical difficulties, there have been no recordings from human fusimotor
neurons during the development of muscle fatigue, the most direct recordings being from
the resultant changes in muscle spindle discharge (Mace field et aI., 1991). However, there
have been a few recent studies in which the fatigue-related changes in discharge behavior
of single fusimotor neurons have been recorded from experimental animals. Interestingly,
though, whereas the human studies referred to above indicate that the support via the
fusimotor loop declines during fatiguing contractions, data derived from experiments on
reduced animals point in an opposite direction. Based on their findings when studying
fatigue-induced changes in fusimotor output and afferent spindle activity in decerebrate or
spinal cats, Ljubisavljevic and Anastasijevic (1995) arrived at the following hypothesis: " ..
. the same muscle afferent discharge, elicited in groups III and IV afferents by muscle fatigue
itself would at the same time inhibit skeletomotor neurons by direct reflex action and lend
them support through the gamma loop". In the following, we review the evidence leading
up to this hypothesis.
Animal experiments have provided convincing evidence that known products of
muscle metabolism can increase the discharge rate in group III and IV intramuscular afferents
(Kumazawa & Mizumura, 1977; Kniftki et aI., 1978; Mense & Stahnke, 1983; Kaufman et
aI., 1983; Rotto & Kaufman, 1988; Lagier-Tessonnier et aI., 1993; Sinoway et aI., 1993).
266 K-E. Hagbarth and V. G. Macefield

However, not all group III and IV afferents are stimulated during fatiguing contractions:
small-diameter mechanoreceptive afferents have been shown to exhibit a decline in their
discharge rate during the initial phase of fatigue, although chemosensitive afferents become
spontaneously active late in a maintained contraction (Hayward et ai., 1991; Sinoway et ai.,
1993). This would strongly suggest that any contribution small diameter intramuscular
afferents may make to reflexogenic muscle fatigue would be towards the end of a contraction,
rather than during the initial part. Their contribution might also depend on the ischemic
condition of a muscle, such that metabolites are not washed out from the intramuscular
environment. A logical consequence of this is that these afferents may not be involved in the
generation of fatigue associated with submaximal contractions, in which the vascular supply
of the contracting muscle is not compromised.
Several investigators have reported that in reduced animals, activity evoked in group
III and IV muscle afferents has an autogenetic excitatory influence on fusimotor neurons
(Appelberg et ai., 1983; Jovanovic et ai., 1990; Ljubisavljevic et ai., 1992a,b). In their studies
on decerebrate or spinal cats, Ljubisavljevic and coworkers (1992a,b) found that an increase
in discharge rate of fusimotor neurons develops in parallel with muscle fatigue during
long-lasting muscle contractions elicited by tetanic stimulation of muscle nerves or ventral
roots. The bearing of these results on human muscle fatigue is difficult to evaluate since
segmental reflex machineries are not likely to operate in a similar way in decerebrate or
spinal cats as in healthy human beings. Reported differences in fatigue-induced fusimotor
responses when comparing spinal with decerebrate preparations or preparations with or
without wide limb denervations (Anastasijevic et aI., 1995) provide good examples of how
reflex responses can vary with the state of excitability in segmental reflex arcs and in
internuncial networks. Moreover, in cats anaesthetized with chloralose, the excitability in
central neural networks may well be different from that in conscious human subjects (cf.
Johansson et ai, 1993). Other matters of concern in experiments of this type are the
difficulties associated with distinguishing both between fatigue-related and non fatigue-re-
lated reflex responses and between nociceptive and non-nociceptive reactions.

CONFOUNDING EFFECTS OF THIXOTROPIC CHANGES IN

MUSCLE SPINDLE BEHAVIOR

Further difficulties arise when interpreting changes in spindle afferent actiVIty

resulting from long-lasting tetanic stimulation of muscle nerves or ventral roots. In particular,
it may be hard to distinguish between fatigue-induced changes and changes due to thixotropic
(actomyosin cross-bridge dependent) after-effects in extrafusal and intrafusal muscle fibers.
In a study performed by Nelson and Hutton (1985) on the triceps surae in anaesthetized cats,
sustained contractions of these muscles (elicited by 7x threshold electrical stimulation of
ventral roots) were followed by a significant increase in resting discharge and dynamic
sensitivity of both group Ia and group II spindle afferents. In addition, resting muscle force
and passive peak muscle stiffness were consistently higher following contraction. These
post-tetanic effects were believed to depend on increased inherent stiffness ofintrafusal and
extrafusal muscle fibers, due to persisting actomyosin bonds. It was also proposed that
accompanying adjustments in resting discharge and dynamic sensitivity of spindle endings
can provide reflex compensation for fatigue-induced force declines (cf. Ribot-Ciscar et ai.,
1991). However, as claimed by Proske and colleagues (Gregory et ai., 1987; Gregory et aI.,
1990; Polus et ai., 1991), the thixotropic after-effects of contractions may be either an
enhancement or a reduction of inherent muscle stiffness, dependent on whether a muscle has
previously been held at a short length or been exposed to stretch. Recent studies in man
Fusimotor System 267

confirm this claim (Hagbarth et aI., 1995). The history-dependent changes in muscle spindle
behavior may well account for some of the apparently conflicting reports on involvement of
the fusimotor system in muscle fatigue.

CONCLUDING REMARKS

To summarize, it is questionable whether studies performed on decerebrate or spinal

cats can ever shed light on the behavior of fusimotor neurons during fatiguing contractions
in man. Further microneurographic studies on conscious human subjects are needed to
elucidate this issue, particularly with respect to the possible contributions of group III and
IV afferents to fatigue. Results obtained with this technique so far indicate that the amplifi-
cation of skeletomotor output via the fusimotor loop declines as muscle fatigue develops. It
remains to be determined whether this declining support is the main cause of the muscle
wisdom phenomenon, but the parallel time courses of the changes in spindle discharge and
motor unit firing lend strong support to this idea. As pointed out by Gandevia and colleagues
(1990), it is likely that waning excitation via the fusimotor loop (due to a reduction in muscle
spindle input and increase in pre-synaptic inhibition) is most prominent in the early phases
of fatigue, whereas progressive reflex inhibition mediated by small diameter afferents
contributes more later in prolonged contractions. Certainly all available experimental
evidence supports this view. Evidence of active inhibition of alpha motoneurons during
fatigue could come from knowing the firing rates of single motor units in tibialis anterior
towards the end of a sustained maximal voluntary contraction; only MVCs of lOs duration
have been studied in this muscle so far (Bigland-Ritchie et aI., 1992). If these firing rates
tum out to be lower than the known peak firing rates of deafferented motoneurons supplying
this muscle (Macefield et aI., 1993) then active inhibition must be responsible. Conversely,
if peak firing rates are not lower then the decline in firing rates could be explained fully by
a progressive dis facilitation of alpha motoneurons resulting from withdrawal of fusimotor
support. Observed enhancements in the efficacy of stretch and unloading reflexes during
muscle fatigue can be explained in terms of automatic gain compensation (Matthews, 1986)
but do not exclude the possibility that there actually is a reduction in the spindle-mediated
neural messages impinging on the motoneuron pool, such that motor output becomes more
alpha-dominated as fatigue processes develop.

ACKNOWLEDGMENTS

The authors' laboratories are supported by the Medical Research Council of Sweden
and the National Health & Medical Research Council (NH & MRC) of Australia. Attendance
ofV.G.M. at the 1994 Bigland-Ritchie conference was supported, in part, by the NH & MRC,
the American Physical Therapy Association (Section on Research), and the Muscular
Dystrophy Association (USA).

REFERENCES
Anastasijevic R, Jovanovic K & Ljubisavljevic M (1995). Fusimotor responses to fatiguing contractions ofa
muscle in non-denervated cat hindlimb. In: Stuart DG, Gantchev GN, Gurfinkel VS, Wiesendanger
M (eds.), Motor Control VII, pp. 00-00. Tucson: Motor Control Press. In press
Aniss AM, Diener HC, Hore J, Burke D & Gandevia SC (1990). Reflex activation of muscle spindles in human
pretibial muscles during standing. Journal ofNeurophysiology 64,671-679.
268 K-E. Hagbarth and V. G. Macefield

Appelberg B, Hulliger M, Johansson H & Sojka P (1983). Actions on ga=a-motoneurones elicited by

electrical stimulation of group III muscle afferents in the hind limb of the cat. Journal of Physiology
(London) 335, 275-292.
Baldissera F & Pierrot-Deseilligny E (1989). Facilitation of transmission in the pathway ornon-monosynaptic
Ia excitation to wrist flexor motoneurones at the onset of voluntary movement in man. Experimental
Brain Research 74, 437-439.
Balestra C, Duchateau J & Hainaut K (1992). Effects of fatigue on the stretch reflex in a human muscle.
Electroencephalography and Clinical Neurophysiology 85, 46-52.
Bevan L, Laouris Y, Reinking RM & Stuart DG (1992). The effect of the stimulation pattern on the fatigue of
single motor units in adult cats. Journal of Physiology (London) 449, 85-108.
Bigland-Ritchie B, Cafarelli E & Vellestad NK (1986). Fatigue of submaximal static contractions. Acta
Physiologica Scandinavia Supplement 556, 137-148.
Bigland-Ritchie B, Johansson R, Lippold DCJ, Smith S & Woods JJ (1983). Changes in motoneurone firing
rates during sustained maximal voluntary contractions. Journal ofPhysiology (London) 340, 335-346.
Bigland-Ritchie B, Jones DA & Woods JJ (1979). Excitation frequency and muscle fatigue: electrical responses
during voluntary and stimulated contractions. Experimental Neurology 64, 414-427.
Bigland-Ritchie B & Woods JJ (1984). Changes in muscle contractile properties and neural control during
human muscular fatigue. Muscle & Nerve 7,691-699.
Bigland-Ritchie BR, Dawson NJ, Johansson RS & Lippold DC (1986). Reflex origin for the slowing of
motoneurone firing rates in fatigue of human voluntary contractions. Journal ofPhysiology (London)
379,451-459.
Bigland-Ritchie BR, Furbush FH, Gandevia SC & Thomas CK (1992). Voluntary discharge frequencies of
human motoneurons at different muscle lengths. Muscle & Nerve 15,130-137.
Binder-Macleod SA & Guerin T (1990). Preservation of force output through progressive reduction of
stimulation frequency in human quadriceps femoris muscle. Physical Therapeutics 70, 619-625.
Bongiovanni LG & Hagbarth KE (1990). Tonic vibration reflexes elicited during fatigue from maximal
voluntary contractions in man. Journal of Physiology (London) 423, 1-14.
Bongiovanni LG, Hagbarth KE & Stjernberg L (1990). Prolonged muscle vibration reducing motor output in
maximal voluntary contractions in man. Journal of Physiology (London) 423, 15-26.
Burke D (1981). The activity of human muscle spindle endings during normal motor behavior. In: Porter R
(ed.), Neurophysiology IV. International Review ofPhysiology, pp. 91-126. Baltimore, MD: Univer-
sity Park Press.
Burke D, Gracies JM, Meunier S & Pierrot-Deseilligny E (1992). Changes in presynaptic inhibition of afferents
to propriospinal-like neurones in man during voluntary contractions. Journal ofPhysiology (London)
449, 673-687.
Cleland C, Rymer W & Edwards F (1982). Force-sensitive interneurones in the spinal cord of the cat. Science
217,652-655.
Darling WG & Hayes KC (1983). Human servo responses to load disturbances in fatigued muscle. Brain
Research 267, 345-351.
Duchateau J & Hainaut K (1993). Behaviour of short and long latency reflexes in fatigued human muscles.
Journal of Physiology (London) 471, 787-799.
Eldred E, Granit R & Merton PA (1953). Supraspinal control of the muscle spindles and its significance.
Journal of Physiology (London) 122,498-523.
Enoka RM & Stuart DG (1992). Neurobiology of muscle fatigue. Journal of Applied Physiology 72,
1631-1648.
Gandevia SC, Burke D, Macefield G & McKenzie DK (1992). Human motor output, muscle fatigue and muscle
afferent feedback. Proceedings of the Australian Physiological and Pharmacological Society 23,
59-67.
Gandevia SC, Macefield G, Burke D & McKenzie DK (1990). Voluntary activation of human motor axons in
the absence of muscle afferent feedback. The control of the deafferented hand. Brain 113, 1563-1581.
Gandevia SC, Macefield VG, Bigland-Ritchie B, Gorman R & Burke D (1993). Motoneuronal output and
gradation of effort in attempts to contract acutely paralyzed leg muscles in man. Journal ofPhysiology
(London) 474, 411-427.
Garland 8J (1991). Role of small diameter afferents in reflex inhibition during human muscle fatigue. Journal
ofPhysiology (London) 435, 547-558.
Garland 8J, Garner 8H & McComas AI (1988). Reduced voluntary electromyographic activity after fatiguing
stimulation of human muscle. Journal ofPhYSiology (London) 401, 547-556.
Garland SJ & McComas AJ (1990). Reflex inhibition of human soleus muscle during fatigue. Journal of
Physiology (London) 429,17-27.
Fusimotor System 269

Gregory JE, Mark RF, Morgan DL, Patak A, Polus B & Proske U (1990). Effects of muscle history on the
stretch reflex in cat and man. Journal of Physiology (London) 424, 93-107.
Gregory JE, Morgan DL & Proske U (1987). Changes in size of the stretch reflex of cat and man attributed to
aftereffects in muscle spindles. Journal of Neurophysiology 58, 628-640.
Grimby L, Hannerz J & Hedman B (1981). The fatigue and voluntary discharge properties of single motor
units in man. Journal of Physiology (London) 316, 545-554.
Hagbarth KE (1993). Microneurography and applications to issues of motor control: Fifth Annual Stuart Reiner
Memorial Lecture. Muscle & Nerve 16, 693-705.
Hagbarth KE, Kunesch EJ, Nordin M, Schmidt R & Wallin EU (1986). Gamma loop contributing to maximal
voluntary contractions in man. Journal of Physiology (London) 380, 575-591.
Hagbarth KE, Nordin M & Bongiovanni LG (1995). After-effects on stiffness and stretch reflexes of human
finger flexor muscles attributed to muscle thixotropy. Journal ofPhysiology (London) 482.1, 215-223.
Hayward L, Wesselmann U & Rymer WZ (1991). Effects of muscle fatigue on mechanically sensitive afferents
of slow conduction velocity in the cat triceps surae. Journal of Neurophysiology 65, 360-370.
Hultborn H, Meunier S, Pierrot-Deseilligny E & Shindo M (1987). Changes in presynaptic inhibition of Ia
fibres at the onset of voluntary contraction in man. Journal of Physiology (London) 389, 757-772.
Hunter IW & Kearney RE (1983). Invariance of ankle dynamic sensitivity during fatiguing muscle contrac-
tions. Journal of Biomechanics 16, 985-991.
Hiikkinen K & Komi PV (1983). Electromyographic and mechanical characteristics of human skeletal muscle
during fatigue under voluntary and reflex conditions. Electroencephalography and Clinical Neuro-
physiology 55, 436-444.
Johansson H, Djupsjobacka M & Sjolander P (1993). Influences on the gamma-muscle spindle system from
muscle afferents stimulated by KCl and lactic acid. Neuroscience Research 16, 49-57.
Jovanovic K, Anastasijevic R & Vuco J (1990). Reflex effects on gamma fusimotor neurones of chemically
induced discharges in small-diameter muscle afferents in decerebrate cats. Brain Research 521, 89-94.
Kaufman MP, Longhurst JC, Rybicki KJ, Wallach JH & Mitchell JH (1983). Effects of static muscular
contraction on impulse activity in groups III and IV afferents in cats. Journal ofApplied Physiology
55, 105-112.
Kirsch RF & Rymer WZ (1987). Neural compensation for muscular fatigue: evidence for significant force
regulation in man. Journal of Neurophysiology 57,1893-1910.
Kirsch RF & Rymer WZ (1992). Neural compensation for fatigue-induced changes in muscle stiffness during
perturbations of elbow angle in human. Journal ofNeurophysiology 68, 449-470.
Kniffki KD, Mense S & Schmidt RF (1978). Responses of group IV afferent units from muscle to stretch,
contraction and chemical stimulation. Experimental Brain Research 31, 511-522.
Kniffki KD, Shomburg ED & Steffens, H (1981). Synaptic effects from chemically activated fine muscle
afferents upon alpha motoneurones in decerebrate and spinal cats. Brain Research 206, 361-370.
Kukulka CG, Moore MA & Russell AG (1986). Changes in human alpha-motoneuron excitability during
sustained maximum isometric contractions. Neuroscience Letters 68,327-333.
Kumazawa T & Mizumura K (1977). Thin-fibre receptors responding to mechanical, chemical and thermal
stimuli in the skeletal muscle of the dog. Journal of Physiology (London) 273, 179-194.
Lagier-Tessonnier F, Balzamo E & Jammes Y (1993). Comparative effects of ischemia and acute hypoxemia
on muscle afferents from tibialis anterior in cats. Muscle & Nerve 16, 135-141.
Ljubisavljevic M, Jovanovic K & Anastasijevic R (1992a). Changes in discharge rate of cat hamstring
fusimotor neurones during fatiguing contractions of triceps surae muscles. Brain Research 579,
246-252.
Ljubisavljevic M, Jovanovic K & Anastasijevic R (1992b). Changes in discharge rate offusimotor neurones
provoked by fatiguing contractions of cat triceps surae muscles. Journal ofPhysiology (London) 445,
499-513.
Ljubisavljevic M & Anastasijevic R (1995). Muscle spindle afferent discharges during fatiguing muscle
contractions in decerebrate cats. In: Stuart DG, Gantchev GN, Gurfinkel VS, Wiesendanger M (eds.),
Motor Control VII, pp. 00-00. Tucson AZ: Motor Control Press. In press
Loeb GE & Hoffer JA (1981). Muscle spindle function during normal and disturbed locomotion. In: Taylor A,
Prochazka A (eds.), Muscle Receptors and Movement, pp. 219-228. London: Macmillan.
Macefield G, Hagbarth K-E, Gorman R, Gandevia SC & Burke D (1991). Decline in spindle support to alpha
motoneurones during sustained voluntary contractions. Journal of Physiology (London) 440, 497-
512.
Macefield VG, Gandevia SC, Bigland-Ritchie B, Gorman R & Burke D (1993). The firing rates of human
motoneurones voluntarily activated in the absence of muscle afferent feedback. Journal ofPhysiology
(London) 474, 429-443.
270 K-E. Hagbarth and V. G. Macefield

Marsden CD, Meadows JC & Merton PA (1983) "Muscular wisdom" that minimized fatigue during prolonged
effort in man: peak rates of motoneuron discharge and slowing of discharge during fatigue. In:
Desmedt JE (ed.), Motor Control Mechanisms in Health and Disease, pp. 169-211. New York: Raven
Press.
Matthews PB (1986). Observations on the automatic compensation of reflex gain on varying the pre-existing
level of motor discharge in man. Journal of Physiology (London) 374, 73-90.
Matthews PBC (1981). Evolving views on the internal operation and functional role of the muscle spindle.
Journal of Physiology (London) 320, 1-30.
Mense S & Stahnke M (1983). Responses in muscle afferent fibres of slow conduction velocity to contractions
and ischaemia in the cat. Journal of Physiology (London) 342, 383-397.
Merton PA(1953). Speculations on the servo-control of movement. In: Wolstenholme, GEW (ed.), The Spinal
Cord, pp. 247-307. London: Churchill.
Merton PA (1954). Voluntary strength and fatigue. Journal of Physiology (London) 123, 553-564.
Meunier S & Pierrot-Deseilligny E (1989). Gating of the afferent volley of the monosynaptic stretch reflex
during movement in man. Journal of Physiology (London) 419, 753-763.
Nelson DL & Hutton RS (1985). Dynamic and static stretch responses in muscle spindle receptors in fatigued
muscle. Medicine and Science in Sports Exercise 17, 445-450.
Person RS (1974). Rhythmic activity in a group of human motoneurones during voluntary contraction of a
muscle. Electroencephalography and Clinical Neurophysiology 36, 585-595.
Polus BI, Patak A, Gregory JE & Proske U (1991). Effect of muscle length on phasic stretch reflexes in humans
and cats. Journal ofNeurophysiology 66, 613-622.
Prochazka A & Wand P (1981). Independence of fusimotor and skeletomotor systems during voluntary
movements. In: Taylor A, Prochazka A (eds.), Muscle Receptors and Movement, pp. 229-243. London:
Macmillan.
Ribot-Ciscar E, Tardy-Gervet MF, Vedel JP & Roll JP (1991). Post-contraction changes in human muscle
spindle resting discharge and stretch sensitivity. Experimental Brain Research 86, 673-678.
Rotto DM & Kaufman MP (1988). Effect of metabolic products of muscular contraction on discharge of group
III and IV afferents. Journal of Applied Physiology 64, 2306-2313.
Sinoway 11, Hill JM, Pickar JG & Kaufman MP (1993). Effects of contraction and lactic acid on the discharge
of group III muscle afferents in cats. Journal of Neurophysiology 69, 1053-1059.
Vallbo AB, Hagbarth K-E, Torebjork HE & Wallin BG (1979). Somatosensory, proprioceptive and sympathetic
activity in human peripheral nerves. Physiological Reviews 59, 919-957.
Windhorst U, Christakos CN, Koehler W, Hamm TM, Enoka RM & Stuart DG (1986). Amplitude reduction
of motor unit twitches during repetitive activation is accompanied by relative increase of hyperpo-
larizing membrane potential trajectories in homonymous alpha-motoneurons. Brain Research 398,
181-184.
Woods JJ, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal ofNeurophysiology 58, 125-137.
 19

ROLE OF MUSCLE AFFERENTS IN THE

INHIBITION OF MOTONEURONS DURING
FATIGUE

S.1. Garland l and M. P. Kaufman2

1 Department of Physical Therapy

Elborn College, University of Western Ontario
Ontario, Canada N6G IHI
2 Division of Cardiovascular Medicine, Departments of Internal Medicine
and Human Physiology
University of California
Davis, California 95616

ABSTRACT

In conscious humans, fatiguing muscular contractions are accompanied by a decrease in the

discharge rate of alpha motoneurons. The association between alpha motoneuron discharge
rate and the generation of force by skeletal muscle has been called "muscle wisdom"
(Marsden et ai., 1983). Its purpose is believed to ensure that central neural drive to skeletal
muscle, which is fatigued, matches that needed to generate the required force. In addition,
muscle wisdom may be one mechanism that functions either to decrease or to postpone
central neural fatigue (Enoka & Stuart, 1992). Bigland-Ritchie and colleagues (1986) have
suggested that a reflex arising from fatigued skeletal muscle is responsible, at least in part,
for muscle wisdom. This chapter has two purposes. The first is to evaluate the evidence that
a reflex arising from fatigued skeletal muscle causes muscle wisdom, and the second is to
examine the discharge properties of muscle afferents to determine which ones are most likely
to initiate reflexly this phenomenon.

INTRODUCTION

Before discussing the role played by muscle afferents in reflexly inhibiting mo-
toneuron discharge during fatigue, we need to provide some basic information about the
sensory innervation of limb skeletal muscles, the discharge properties of the afferents and
in selected circumstances, the locations of their endings. Limb muscles are innervated by
five types of sensory nerves. These have been classified as group I through IV, with the first
group having two subtypes, la and lb. The classification scheme is based, in part, on the

271
272 S. J. Garland and M. P. Kaufman

diameter of the afferent fibers. Specifically, the thicker the fiber (i.e., the more myelin it has
surrounding it), the faster it conducts impulses (Hunt, 1954; Boyd & Davy, 1968).
Group Ia and Ib muscle afferents are thickly myelinated and conduct impulses
between 72 and 120 mls in cats and dogs. The receptors of group Ia afferents are primary
muscle spindle endings and those of group Ib afferents are Golgi tendon organs. Primary
spindle afferents (i.e., group Ia) are located in parallel with the skeletal muscle fibers that
they innervate. Consequently, the discharge of primary spindle afferents is increased by
stretch (i.e., lengthening) of the muscle and is decreased by contraction (i.e., shortening).
Golgi tendon organs, in contrast, are situated in series with the muscle fibers they innervate;
consequently, the discharge of these afferents is increased by both contraction and muscle
stretch.
Group II afferents, also called secondary spindle endings, conduct impulses between
31 and 71 mls in cats and dogs. The receptors of these afferents (i.e., the spindles), like their
group Ia counterparts, are situated in parallel with the muscle fibers that they innervate.
Although secondary spindle afferents (i.e., group II) are stimulated by muscle stretch, they
are not capable of signaling the rate at which this occurs. In contrast, primary spindle
afferents (i.e., group Ia) are capable of signaling the rate of stretch; consequently, they are
dynamically sensitive. Muscular contraction decreases the discharge rates of group II spindle
afferents. There are also group II afferent fibers, which are not innervated by muscle spindles.
These group II fibers, unlike group II spindle afferents, are stimulated by contraction.
Group III afferents, also called A8 fibers, are thinly myelinated and conduct impulses
between 2.5 and 30 mls in cats and dogs. Their receptors are free nerve endings. Group IV
afferents, also called C-fibers, are unmyelinated and conduct impulses at less than 2.5 mls. Like
their group III counterparts, the receptors of group IV afferents are free nerve endings.
Nevertheless, the endings innervating both group III and IV afferents are almost entirely
surrounded by Schwann cells. A small portion of the ending is bare, and this area is thought to
be the site of action for the stimuli that activate them. Moreover, the bare areas of the nerve
ending contain mitochondria and other structures whose presence is consistent with the notion
that these areas are indeed the receptive part of the sensory neuron (Andres et aI., 1985).
Serial reconstruction of electron micrographs have unmasked the locations of the
endings of group III and IV afferents in the calcaneal (i.e., Achilles) tendon ofthe cat (Andres
et aI., 1985). Five locations were found for the endings of group III afferents. Two of these
locations were in vessels (venules and lymphatics), two were in the connective tissue of the
peritenoneum extemum and intemum, and the other was in the endoneurium. The locations
of the endings of group IV afferents showed a marked topographic relationship to the blood
and lymphatic vessels ofthe tendon. Moreover, group IV endings, but not group III, contain
granulated vesicles, the content of which is unknown. These granulated vesicles may contain
neuropeptides, whose release via the axon reflex evokes vasodilation. Obviously, the
locations of the group IV endings in blood vessels is ideal to cause this effect. In addition,
preliminary description of the locations of the endings of group III and IV afferents in the
triceps surae muscles of the cat appear to parallel closely the locations of these endings in
the calcaneal tendon (von During & Andres, 1990). In both the tendon and the muscles, the
diameters of the group III endings were found to be larger than those of the group IV endings.
Also, group III endings had more mitochondria and had a more distinct receptor matrix than
did group IV endings (Andres et aI, 1985; von During & Andres, 1990).

HUMAN STUDIES

During a fatiguing maximal voluntary contraction (MVC), electromyographic

(EMG) activity has been shown to decline roughly in parallel with the loss of force
Muscle Afferents and Inhibition of Motoneurons during Fatigue 273

(Bigland-Ritchie, 1981). The origin of the declining EMG is largely central to the neuromus-
cular junction (Woods et aI., 1987; Garland et aI., 1988) and has been attributed to reflex
inhibition of the motoneuron pool (Bigland-Ritchie et aI., 1986; Garland & McComas, 1990).
The existence of a reflex whereby motoneurons are inhibited by sensory input from the
fatigued muscle is not limited to sustained MVCs (Bigland-Ritchie et aI., 1986) but also has
been demonstrated in fatiguing intermittent submaximal contractions performed under
ischemic conditions (Garland & McComas, 1990). This reflex is thought to match the
motoneuron output to the functional status of the muscle fibers during fatigue.
The contribution of different sized afferents to reflex inhibition during fatigue remains
unresolved. Two hypotheses have been forwarded: 1) input from small diameter afferents (i.e.,
group III and IV) responsive to chemical or mechanical stimuli during fatigue serves to decrease
motoneuron activity (Bigland-Ritchie, 1986); and 2) a reduction in discharge of muscle spindle
afferents (i.e., group Ia and II) results in disfacilitation (i.e., reduced excitation of motoneurons
by a decrease in fusimotor-driven feedback from muscle spindles) of motoneurons during
fatigue (Macefield et aI., 1991). A third hypothesis, that is unrelated to muscle afferents, is that
motoneuron discharge decreases as a result of intrinsic motoneuron adaptation. Support for this
hypothesis has come from cat motoneurons subjected to constant intracellular or extracellular
current injection (Kernell & Monster, 1982; Spielmann et aI., 1993). This mechanism is less
likely in voluntary contractions, especially in intermittent or submaximal contractions, because
the motoneuron does not experience constant input.

Small Diameter Afferents

The role of small diameter afferents in reflex inhibition has been established in human
muscle fatigue. When fatigue was induced following compressive blockade oflarge diameter
afferents, the EMG during MVCs declined (Garland, 1991); this decline was comparable to
that found in a previous study in which the sensory input was unaltered (Garland et aI., 1988).
The nature of the sensory input, i.e. mechanical or chemical, has not been demonstrated,
although the latter may be more likely than the former.
Bigland-Ritchie and colleagues (1986) found that EMG associated with MVCs
following fatigue remained depressed for 2 or 3 min while the limb was rendered ischemic.
In that study, the EMG demonstrated near-full recovery in 3 min if the blood supply was
intact; this suggested a chemical stimulus. Release of an ischemic cuff after fatigue is
associated with a reduction in muscle pain (Garland et aI., 1988), an effect presumably due
to wash-out of metabolites. Moreover, many of the chemosensitive afferents are known to
be nociceptive. If, however, fatigue is induced by low-force submaximal intermittent
contractions performed while the muscles are freely perfused, motor units showed a constant
or an increased discharge rate (Garland et aI., 1988). In addition, Marsden and colleagues
(1983) were unable to demonstrate any change in motor unit firing rates by slowing muscle
contraction through cooling. Similarly, Bigland-Ritchie and colleagues (1992) found no
difference in motoneuron discharge rates when the twitch contraction was slowed by
changing muscle length. Hence, in the absence of chemical changes associated with fatigue,
the motoneuron discharge was unaffected.

Large-Diameter Afferents
Feedback from large-diameter afferents has been demonstrated to influence motor
unit discharge rates in non-fatigued muscle. Hagbarth and coworkers (1986) recorded
motoneuron discharge rates following blockade of fusimotor efferent activity during brief
nonfatiguing MVCs. Under these conditions, irregular and reduced motoneuron discharge
was evident; this effect could be reversed by muscle vibration, which strongly activates Ia
274 S. J. Garland and M. P. Kaufman

spindle afferents (e.g., Burke et aI., 1976). Local anesthetic has been used to block all muscle
afferents in subjects attempting sustained MVCs (Gandevia et aI., 1990; Macefield et aI.,
1993). Under these conditions, subjects also demonstrated lower than normal motoneuron
discharge rates yet the motoneuron discharge rates did not show the decline during sustained
maximal voluntary effort evident in normally innervated muscle. This provides further
support for the role of muscle afferents (but not specifically large-diameter afferents) in
modulating motoneuron discharge during fatigue.
The only direct measure of spindle activity in humans has indicated disfacilitation
from muscle spindles evoked by low-force submaximal contractions ofless than one min
duration (Macefield et aI., 1991). However, the amount offatigue in this situation would be
minimal. Bongiovanni and Hagbarth (1990) found that vibration was able to increase the
discharge rate ofmotoneurons during sustained MVCs for only a short time, after which the
motoneuron discharge rate declined despite continued vibration. In that experiment, there
was no verification (i.e., through superimposed electrical stimulation) that subjects were able
to activate fully the musculature. Thus the vibration could have simply increased the
motoneuron firing rates by supplementing declining descending central drive, rather than by
removing any disfacilitation arising from the muscle spindles. Further, the proprioceptive
ability of subjects remained unchanged following fatiguing MVCs (Sharpe & Miles, 1993).
It is possible that any fatigue-induced reductions in muscle spindle discharge may have been
compensated for by concurrent changes in central processing of the sensory input, thereby
rendering the proprioceptive ability unchanged. Thus, although a decline in muscle spindle
input would result in disfacilitation of the motoneuron pool, there remains some uncertainty
as to its relative importance during fatigue.
Studies that demonstrated a decrease in the muscle's response to stretch following
fatigue, evoked by repetitive stretch-shortening cycles, do not enable a definitive distinction
between the type of afferent that may be involved (Balestra et aI., 1992). The stretch reflex
could have been reduced by decreased sensitivity of the muscle spindles or alternatively
decreased alpha motoneuron excitability resulting from inhibition mediated by small diame-
ter afferents. Others have demonstrated an increase in the stretch reflex evoked by mechani-
cal tendon tap following fatigue induced by stretch-shortening exercise (Hartobagyi et aI.,
1991) or by submaximal sustained isometric contractions (Hakkinen & Komi, 1983). In these
studies, the effect of muscle fiber membrane hyperpolarization, the phenomenon accounting
for transient M-wave potentiation during fatigue (Hicks & McComas, 1989), could explain
the increased amplitude of the response to stretch without any effect on the stretch reflex,
per se. The aforementioned studies have not included the controls necessary to interpret the
data in terms of spindle sensitivity during fatigue.

ANIMAL STUDIES

Effects of Fatigue on the Discharge of Muscle Afferents

Experimental paradigms examining the effects of fatigue on afferent discharge in
anesthetized animals require some explanation. Obviously, the essence of the paradigm is to
measure an afferent's response to some maneuver before and during muscle fatigue. The
specific criterion used to define fatigue differs among investigators, but the one used by the
Hayward and colleagues (1991) may prove informative. These investigators defined fatigue
as a reduction in force output by the contracting gastrocnemius muscles to 30% of that
developed by the initial contraction, which was induced by electrically stimulating the
muscle nerve using 25 Hz at 1.3 times motor threshold. This stimulation protocol produced
a submaximal contraction.
Muscle Afferents and Inhibition of Motoneurons during Fatigue 275

Muscle spindles, if responsible for causing fatigue-induced reflex inhibition of alpha

motoneuron discharge, should display a reduction in their responses to a maneuver (such as
tendon stretch or contraction) during fatigue as compared with their resJ10nses to the maneuver
before fatigue. The animal literature provides little support for this hypothesis. For example,
although spindle firing has been shown to decrease temporarily at the onset of fatigue
(Windhorst & Kokkoroyiannis, 1991), this recovers quickly. As the contraction period pro-
gressed, fatigue enhanced rather than reduced the responses of muscle spindles to contraction
in either chloralose or barbiturate anesthetized cats (Nelson & Hutton, 1985; Hayward et ai.,
1991). It is also possible that muscle spindle afferent feedback could combine with recurrent
inhibition from Renshaw cells as proposed by Windhorst and Kokkoroyiannis (1991).
Golgi tendon organs, if responsible for causing fatigue-induced reflex inhibition of
alpha motoneuron discharge, should display an increase in their responses to a maneuver
such as tendon stretch during fatigue as compared with their responses to stretch before
fatigue. This hypothesis has not been supported by the animal literature. Fatigue has been
shown to decrease the responses of Golgi tendon organs to stretch (Hutton & Nelson, 1986;
Thompson et aI, 1990); in another study, fatigue had little effect on the responses of tendon
organs to stretch (Hayward et aI., 1991).
Fatigue affects the discharge properties of group III and non-spindle group II afferents
in a manner consistent with the notion that they are responsible, in part, for reflexly
decreasing motoneuron discharge rates either directly or by Renshaw cell inhibition. Mus-
cular fatigue, induced in barbiturate anesthetized cats, has been shown to increase the
spontaneous discharge rate of group II non-spindle and group III afferents as well as to
increase their responses to tendon stretch, to surface pressure (i.e., probing their receptive
fields) and in a few instances to contraction (Hayward et ai., 1991). Unfortunately, the effect
of fatigue on the discharge properties of group IV afferents has not been determined.

Discharge Properties Of Small Diameter Afferents

The stimulus to the group III and non-spindle group II afferents causing their
fatigue-induced sensitization to mechanical stimuli, such as tendon stretch and surface
pressure, is likely to be metabolic in nature. Many group III as well as group IV muscle
afferents have been shown to be stimulated by intraarterial injection of metabolic products
of muscular contraction. These products include bradykinin (Mense & Schmidt, 1974;
Mense, 1977; Kaufman et ai., 1983), arachidonic acid and prostaglandin E2 (Mense, 1981;
Rotto & Kaufman, 1988), potassium (Kniftki et ai., 1978; Rybicki et ai., 1985), as well as
lactic acid (Thimm & Barm, 1987; Rotto & Kaufman, 1988; Sinoway etaI., 1993). Moreover,
intraarterial injection of these metabolic products has little, if any, direct effect on the
discharge of spindle afferents and Golgi tendon organs (e.g., Prochazka & Somjen, 1986).
Conceivably, these injections could stimulate indirectly these group Ia afferents by activating
gamma motoneurons (Jovanovic et aI., 1990; see Hagbarth & Macefield, Chapter 18).
The discharge pattern of group III and IV muscle afferents in response to a tetanic
contraction of 1-2 min might also provide a clue about the role played by these small diameter
afferents in reflexly inhibiting the discharge of motoneurons supplying fatigued muscles.
Group III afferents often respond vigorously at the onset of a contraction, but then decrease
their discharge rate as the tension developed by the tetanized muscle decreases; i.e., fatigue
(Kaufman et ai., 1983; Mense & Stahnke, 1983). Group IV afferents, on the other hand,
respond weakly, or sometimes not at all, at the onset of a brief tetanic contraction. After this
latent period, which usually lasts 10 to 30 s, group IV afferents respond strongly to
contraction and often maintain their response to contraction as the tension developed by the
tetanized muscle decreases (Kaufman et ai., 1983; Mense & Stahnke, 1983). Ischemia has
little effect on the responses of most group III afferents to tetanic contraction; in contrast,
276 s. J. Garland and M. P. Kaufman
ischemia markedly increased the responses to contraction of a substantial number of group
IVafferents (Kaufman et aI., 1983; Mense & Stahnke, 1983).
The above information might lead one to conclude that group III afferents appear to be
sensitive to mechanical events in the muscle, whereas group IV afferents appear to be sensitive
to metabolic events in the muscle. While there is some basis for such a conclusion, it is a gross
oversimplification. For example, the mechanical sensitivity of group III muscle afferents to
contraction is increased by the prior administration of bradykinin (Mense & Mayer, 1988) and
arachidonic acid (Rotto et aI., 1990). Moreover, both substances are produced by hindlimb
skeletal muscle when it contracts (Rotto et al., 1989; Stebbins et aI., 1990). In addition, the
mechanical sensitivity of group III afferents to contraction is decreased by administration of
dichloroacetate, a substance which decreased the working muscle's ability to produce lactic
acid (Sinoway et aI., 1993). These fmdings lead to the speculation that fatigue-induced
excitation of group III and IV afferents might be caused by the production of bradykinin, lactic
acid, and arachidonic acid or its metabolites in the working muscles.

SUMMARY

In humans, the inhibitory reflex effect ofboth large and small diameter muscle afferents
on motoneuron output during fatigue has been demonstrated. This reflex inhibition has been
demonstrated with isometric contractions with occluded blood flow utilizing a sustained MVC
(Bigland-Ritchie et aI., 1986) or submaximal contractions under ischemic conditions (Garland
& McComas et aI., 1990). Reflex inhibition is less evident in non-ischemic conditions (Garland
et aI., 1988). Disfacilitation of motoneurons has been supported by the depressed spindle
afferent discharge at the onset of sustained submaximal contractions (Macefield et aI., 1991).
Bongiovanni and Hagbarth (1990) found that the declining motoneuron discharge rates during
MVCs could be alleviated, albeit temporarily, with vibration. The actions oflarge- and small-
diameter muscle afferents may not be mutually exclusive. Rather, the two effects could occur
together, with spindle disfacilitation having a larger impact at the onset of fatigue and inhibition
from small-diameter afferents may dominate as fatigue progresses (Gandevia et aI., 1990).
In animals, the inhibitory reflex effect by small diameter muscle afferents on mo-
toneuron output during fatigue has been demonstrated (Hayward et aI., 1988). In addition, there
is substantial evidence that the stimulus to these small diameter afferents is metabolic in nature.
In contrast to the fmdings in humans, the findings in animals offer no support for an inhibitory
reflex effect by large diameter afferents on motoneuron output during fatigue. In a way, this is
surprising because the inhibition from group III and IV afferents may be exerted on both alpha
and gamma motoneurons. We can offer no explanation for the discrepant findings in humans
and in animals, other than the different periods of recording in each paradigm may have
prevented the mapping of the full fatigue-related response. Alternatively, the explanation may
be found between differences in the methods used to induce fatigue in humans and those used
to induce fatigue in animals. In any event, further work in this important area should focus on
this discrepancy and should seek to confirm and extend previous findings.

ACKNOWLEDGMENTS

The authors' laboratories are supported by the Natural Sciences and Engineering
Research Council (NSERC) of Canada (S.J. G.) and United States Public Health Service grant
HL 30710 (MP.K.). Their attendance at the 1994 Bigland-Ritchie conference was supported,
in part, by NSERC (8.J.G.), the Muscular Dystrophy Association (USA; MP.K.), and the
University of Miami (8.J.G.).
Muscle Afferents and Inhibition of Motoneurons during Fatigue 277

REFERENCES
Andres KH, von During M & Schmidt RF (1985). Sensory innervation of the Achilles tendon by group III and
IV afferent fibers. Anatomy and Embryology 172, 145-156.
Balestra C, Cuchateau J & Hainaut K (1992). Effects of fatigue on the stretch reflex in a human muscle.
Electroencephalography and Clinical Neurophysiology 85, 46-52.
Bigland-Ritchie B (1981). EMG and fatigue of human voluntary and stimulated contractions. In: Porter R,
Whelan J (eds.), Human Muscle Fatigue, Physiological Mechanisms, pp 130-156. London: Pitman
Medical.
Bigland-Ritchie B, Dawson NJ, Johansson RS & Lippold OCJ (1986). Reflex origin for the slowing of
motoneurone firing rates in fatigue of human voluntary contractions. Journal ofPhysiology (London)
379,451-459.
Bigland-Ritchie B, Furbush F, Gandevia SC & Thomas CK (1992). Voluntary discharge frequencies of human
motoneurons at different muscle lengths. Muscle & Nerve 15,130-137.
Bongiovanni LG & Hagbarth K-E (1990). Tonic vibration reflexes elicited during fatigue from maximal
voluntary contractions in man. Journal of Physiology (London) 423, 1-14.
Boyd IA & Davy MR (1968). Composition of Peripheral Nerves. Edinbugh: Livingston.
Burke D, Hagbarth K-E, Logstedt L & Wallin BG (1976). The response of human muscle spindle endings to
vibration of non-contracting muscles. Journal of Physiology (London) 261, 695-711.
Enoka RM & Stuart DG (1992). Neurobiology of muscle fatigue. Journal of Applied Physiology 72,
1631-1648.
Gandevia SC, Macefield G, Burke D & McKenzie DK (1990). Voluntary activation of human motor axons in
the absence of muscle afferent feedback. Brain 113, 1563-1581.
Garland SJ. (1991). Role of small diameter afferents in reflex inhibition during human muscle fatigue. Journal
of Physiology (London) 435, 547-558.
Garland SJ, Garner SH & McComas AJ (1988). Reduced voluntary electromyographic activity after fatiguing
stimulation of human muscle. Journal of Physiology (London) 408, 547-556.
Garland SJ & McComas AJ (1990). Reflex inhibition of human soleus muscle during fatigue. Journal of
Physiology (London) 429, 17-29.
Hagbarth K-E, Kunesch EJ, Nordin M, Schmidt R & Wallin EU (1986). Gamma loop contributing to maximal
voluntary contraction in man. Journal ofPhysiology (London) 380, 575-591.
Hakkinen K & Komi PV (1983). Electromyographic and mechanical characteristics of human skeletal muscle
during fatigue under voluntary and reflex conditions. Electroencephalography and Clinical Neuro-
physiology (Limerick) 55, 436-444.
Hayward L, Breitbach D & Rymer WZ (1988). Increased inhibitory effects on close synergists during muscle
fatigue in the decerebrate cat. Brain Research 440,199-203.
Hayward L, Wesse1mann U & Rymer WZ (1991). Effects of muscle fatigue on mechanically sensitive afferents
of slow conduction velocity in the cat triceps surae. Journal of Neurophysiology 65, 360-370.(Ab-
stract}
Hicks A & McComas AJ (1989). Increased sodium pump activity following repetitive stimulation of rat soleus
muscles. Journal of Physiology (London) 414, 337-349.
Hortobagyi T, Lambert NJ & Kroll WP (1991). Voluntary and reflex responses to fatigue with stretch-short-
ening exercise. Canadian Journal of Sport Sciences 16,142-150.
Hunt CC (1954). Relation of function to diameter in afferent fibers of muscle nerves. Journal of General
Physiology 38, 117-131.
Hutton RS & Nelson DL (1986). Stretch sensitivity ofGolgi tendon organs in fatigued gastrocnemius muscle.
Medicine and Science in Sports and Exercise 1, 69-74.
Jovanovic K, Anastasijevic R & Vuco J (1990). Reflex effects on gamma fusimotor neurones of chemically
induced discharges in small diameter muscle afferents in decerebrate cats. Brain Research 521, 89-94.
Kaufman MP, Longhurst JC, Rybicki KJ, Wallach JH & Mitchell JH (1983). Effects of static muscular
contraction on impulse activity of groups III and IV afferents in cats. Journal ofApplied Physiology
55, 105-112.
Kaufman MP, Rybicki KJ, Waldrop TG & Ordway GA (1984). Effect of ischemia on responses of group III
and IV afferents to contraction. Journal ofApplied Physiology 57, 644-650.
Kernell D & Monster AW (1982). Motoneurone properties and motor fatigue: an intracellular study of
gastrocnemius motoneurones of the cat. Experimental Brain Research 46, 197-204.
Kniffki K-D, Mense S & Schmidt RF (1978). Responses of group IV afferent units from skeletal muscle to
stretch, contraction and chemical stimuli. Experimental Brain Research 31, 511-522.
278 S. J. Garland and M. P. Kaufman

Macefield VG, Gandevia SC, Bigland-Ritchie B, Gorman RB & Burke D (1993). The firing rates on human
motoneurones voluntarily activated in the absence of muscle afferent feedback. Journal ofPhysiology
(London) 471, 429-443.
Macefield G, Hagbarth K-E, Gorman R, Gandevia SC & Burke D (1991). Decline in spindle support to
alpha-motoneurones during sustained voluntary contractions. Journal of Physiology (London) 440,
497-512.
Marsden CD, Meadows JC & Merton PA (1983). "Muscular wisdom" that minimizes fatigue during prolonged
effort in man: peak rates of motoneuron discharge and slowing of discharge during fatigue. In:
Desmedt JE (ed.), Motor Control Mechanisms in Health and Disease, pp 169-211. New York: Raven
Press.
Mense S (1977). Nervous outflow from skeletal muscle following chemical noxious stimulation. Journal of
Physiology (London) 267, 75-88.
Mense S (1981). Sensitization of group IV muscle receptors to bradykinin by 5-hydroxytryptamine and
prostaglandin E-2. Brain Research 225, 95-105.
Mense S & Meyer H (1988). Bradykinin-induced modulation of the response behavour of different types of
feline group III and IV muscle receptors. Journal of Physiology (London) 398, 49-63.
Mense S & Schmidt RF (1974). Activation of group IV afferent units from muscle by algesic agents. Brain
Research 72, 305-310.
Mense S & Stahnke M (1983). Responses in muscle afferent fibers of slow conduction velocity to contractions
and ischemia in the cat. Journal of Physiology (London) 342, 383-397.
Nelson DL & Hutton RS (1985). Dynamic and static stretch responses in muscle spindle receptors in fatigued
muscle. Medicine and Science in Sports and Exercise 17, 45-450 (Abstract)
Prochazka A & Somjen GG (1986). Insensitivity of cat muscle spindles to hyperkalaemia in the physiological
range. Journal ofPhysiology (London) 372, 26P.
Rotto DM & Kaufman MP (1988). Effects of metabolic products of muscular contraction on the discharge of
group III and IV afferents. Journal ofApplied PhYSiology 64, 2306-2313.
Rotto DM, Massey KD, Burton KP & Kaufman MP (1989). Static contraction increases arachidonic acid levels
in gastrocnemius muscles of cats. Journal ofApplied Physiology 66, 2721-2724.
Rotto DM, Schultz HD, Longhurst JC & Kaufman MP (1990). Sensitization of group III muscle afferents to
static contraction by products of arachidonic acid metabolism. Journal of Applied Physiology 68,
861-867.
Rybicki KJ, Waldrop TG & Kaufman MP (1985). Increasing gracilis interstitial potassium concentrations
stimulates group III and IV afferents. Journal ofApplied PhYSiology 58, 936-941.
Sharpe MH & Miles TS (1993). Position sense at the elbow after fatiguing contractions. Experimental Brain
Research 94, 179-182.
Sinoway LI, Hill JM, Pickar JG & Kaufman MP (1993). Effects of contraction and lactic acid on the discharge
of group III muscle afferents in cats. Journal of Neurophysiology 69, 1053-1059.
Spielmann JM, Laouris Y, Nordstrom MA, Robinson GA, Reinking RM & Stuart DG (1993). Adaptation of
cat motoneurons to sustained and intermittent extracellular activation. Journal of Physiology (Lon-
don) 464, 75-120.
Stebbins CL, Carretero OA, Mindroiu T & Longhurst JC (1990). Bradykinin release from contracting skeletal
muscle of the cat. Journal ofApplied PhYSiology 69, 1225-1230.
Thimm F & Baum K (1987). Response of chemosensitive nerve fibers of group III and IV to metabolic changes
in rat muscles. Pflugers Archiv 410, 143-152.
von During M & Andres KH (1990). Topography and ultrastructure of group III and IV nerve terminals of
cat's gastrocnemius-soleus muscle. In: Zenker W, Neuhuber WL (eds.), The Primary Afferent Neuron:
A Survey of Recent Morpho-Functional Aspects, New York: Plenum.
Windhorst U & Kokkoroyiannis T (1991). Interaction of recurrent inhibitory and muscle spindle afferent
feedback during muscle fatigue. Neuroscience 43,249-259.
Woods J, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal ofNeurophysiology 58, 125-137.
SECTION VI
The Case for Central Fatigue

This section explores the status of central fatigue, ranging from its definition and
measurement to evidence that serotonergic pathways within the CNS are involved.
Chapter 20 (Gandevia and colleagues) introduces a definition of central fatigue as
an exercise-induced decrease in muscle force that is attributable to a decline in motoneuronal
output. In isometric voluntary contractions in human subjects, the technique of twitch
interpolation has revealed that progressive central fatigue develops during a range of exercise
protocols. Its contribution to the reduction in force may vary in different tasks, perhaps being
less important during sustained maximal voluntary contractions than intermittent submaxi-
mal contractions.
It does not necessarily develop to an equal extent in all muscle groups either. The
available evidence suggests that the diaphragm, the principal inspiratory muscle, shows less
central fatigue than some limb muscles. Clearly this arrangement would have survival value.
For limb muscles, transcranial motor cortical stimulation has revealed that corticospinal
output is submaximal during the development of central fatigue.
To study the cognitive processes that may underlie or co-exist with central fatigue,
Chapter 21 (Popivanov and colleagues) reviews the evidence that premovement cerebral
potentials extracted from the EEG are altered during fatigue. Also, they demonstrate that
these processes can, in future, be evaluated by appropriate non-linear analysis of single trials.
This possibility provides a potential new opening in the study of central fatigue although the
simple effect of recruitment of more muscles during fatigue will need to be critically
addressed.
Next, the way in which fatigue distorts the sensation of force is discussed (Chapter
22, L. Jones). The observation that a suitcase held for long enoughfeels heavy underlies the
formal psychophysical finding that force is usually overestimated during muscle fatigue. It
is unlikely that this error is due purely to altered signals from Golgi tendon organs. The sense
of effort, which is related to the level of the motor command to drive motoneurons, is
enhanced during fatigue and it is often difficult for subj ects to dissociate this perception from
the sensation of the absolute force being generated.
In Chapter 23 (Newsholme & Blomstrand) changes in cerebral biochemistry that
might contribute to central fatigue are considered. There are many central transmitter systems
and it is unlikely that only one would be involved in all the cerebral manifestations offatigue.
Evidence is presented that the plasma ratio of tryptophan to branched-chain amino acids falls
during exercise and it is speculated that this may alter the discharge of some 5-hydroxytryp-
tamine neurons within the CNS. The biochemical conversion of tryptophan to 5-hy-
droxytryptamine is not likely to be easily saturated. Increased levels of5-hydroxytryptamine
280 CNS Issues: The Case for Central Fatigue

have been found in the brainstem and hypothalamus of rats after exhaustive exercise. Studies
in human subjects suggest that the manipulation of 5-hydroxytryptamine levels may alter
the perceived levels of fatigue and mental effort.
During muscle fatigue there are not only changes in the perception of performance
due to the altered feedback from the active muscles, but also altered humoral factors that
affect CNS performance. In addition, both reflex feedback and humoral factors will change
the performance of the segmental motor apparatus. While central fatigue has a specific
definition and is measurable using interpolated nerve stimuli, the processes within the CNS
likely to be directly involved in its production are still not well understood.
 20

CENTRAL FATIGUE
Critical Issues, Quantification and Practical Implications

S. C. Gandevia,I,3 G. M. Allen,1 and D. K. McKenzie 1,2,3

1Prince of Wales Medical Research Institute

2 Department of Respiratory Medicine
Prince of Wales Hospital
3Division of Medicine
University of New South Wales
Sydney, Australia

ABSTRACT

Central fatigue during exercise is the decrease in muscle force attributable to a decline in
motoneuronal output. Several methods have been used to assess central fatigue; however,
some are limited or not sensitive enough to detect failure in central drive. Central fatigue
develops during many forms of exercise. A number of mechanisms may contribute to its
development including an increased inhibition mediated by group III and IV muscle afferents
along with a decrease in muscle spindle facilitation. In some situations, motor cortical output
is shown to be suboptimal. A specific terminology for central fatigue is included.

HISTORICAL PERSPECTIVE

Philosophers and psychologists have long recognized that volition underlies the
limits of human motor performance. Late last century Mosso wrote perceptively about
muscle performance and the role played by the central nervous system (CNS) in a volume
entitled simply Fatigue. He measured fatigue in a simple repetitive task in which a weight
was moved as far as possible by flexors of a finger at a set rate and he realized that the role
played by the CNS could be deduced by electrical stimulation of the relevant muscles. In
the twentieth century however, physiologists have either ignored the importance of volition
by assuming that near maximal performance is achieved in an attempted maximal task, or
considered volition too difficult to measure or not worthy of study. For example, A. V. Hill
(1926) clearly recognized that for other than the elite athlete, sub-maximal voluntary drive
could limit performance: "If one took a patient from the hospital and made him work till he
could barely move, one could never be sure that he had driven himself to his limit . ... With
young athletic people one may be sure that they really have gone 'all out'. "

281
282 S. C. Gandevia et al.

This chapter sets out some historical and phenomenological problems in the study of
central fatigue, specifies appropriate terminology, and notes some, but not all, of the sites
where central fatigue occurs.

TERMINOLOGY

Given that muscle fatigue represents any reduction in force generating capacity of a
muscle (e.g., Gandevia, 1992), what should the best performance by a subject be termed?
Traditionally, the best trial from a number of maximal efforts is called the maximal voluntary
contraction, usually abbreviated to MVC. However, it is obvious that in some attempted
maximal efforts, performance can be less than maximal. Even when a trained subject receives
continuous feedback, and is able to eliminate any poor efforts, voluntary drive measured by
twitch interpolation (see below) ranges from 90-100% (maximal) in repeated brief maximal
isometric contractions of 2-3 s duration with the elbow flexor muscles (e.g., Allen et aI.,
1993a; their Fig. 1). In deference to this variability, Bigland-Ritchie and colleagues occa-
sionally ask subjects to produce not maximal but super efforts (Bigland-Ritchie et aI., 1978)!
Therefore, we use the term optimal force for that which would be obtained when the
motoneuron pool is activated maximally by volition, but retain MVC for the best contraction
observed on that occasion in the laboratory. We use the term maximal effort for every
contraction which the subject believes to be maximal, regardless of the force actually
achieved. Voluntary activation denotes the level ofmotoneuronal drive achieved voluntarily
under laboratory conditions with continuous feedback, loud exhortation and the ability to
disallow poor efforts. We reserve the term central fatigue for the reduction in voluntary
activation during a particular maneuver or exercise regimen (e.g., Bigland-Ritchie et aI.,
1978). Because a maximal voluntary contraction can usually be increased by motor nerve
stimulation (see below), the term maximal evocable force would indicate that voluntary force
which cannot be increased by interpolated peripheral nerve stimuli. Under this terminology,
optimal voluntary activation is the level of drive that produces maximal evocable force. This
terminology is given in Table 1.
Processes associated with central fatigue should theoretically be demonstrable not
simply at the level of the final common pathway (i.e., the motoneuron pool), but also at other
sites within the neuraxis. However, the more remote these sites are from the motoneurons

Table 1. Terminology for evaluation of human muscle fatigue

Tenn Definition
Muscle fatigue any exercise-induced reduction in the ability to exert muscle force or
power, regardless of whether or not the task can be sustained
Maximal voluntary contraction a voluntary contraction which a subject believes to be maximal that is
perfonned with continuous feedback and encouragement
Maximal evocable force the force in a maximal voluntary contraction which cannot be increased
by interpolated supramaximal motor-nerve stimuli
Voluntary activation the level of motoneuronal drive during a voluntary contraction
Optimal voluntary activation the level of motoneuronal activation which produces maximal eVQcable
force
Central fatigue a progressive exercise-induced reduction in voluntary activation of a
muscle (usually assessed in maximal voluntary contractions with twitch
interpolation)
These definitions are adapted from those used to evaluate supraspinal influences on muscle fatigue (Gandevia
et aI., 1995).
Central Fatigue 283

the more difficult it is to establish a causal relation to central fatigue. Motor cortical neurons
show a variety of patterns of discharge during movement but some decline during a sustained
contraction (see Phillips & Porter, 1977; Porter & Lemon, 1993). Indirect evidence from
transcranial cortical stimulation might indicate a reduced role of motor cortical input to
motoneuronal discharge during a sustained effort (Brouwer et aI., 1989). Muscle spindle
afferent discharge diminishes progressively during strong isometric contractions although it
is not technically feasible to measure this decline during maximal efforts (Macefield et aI.,
1991). Processes associated with central fatigue at the segmental level will include inhibitory
feedback from Golgi tendon organs and group III and IV afferents.

MEASUREMENT OF MAXIMAL VOLUNTARY DRIVE:

EXPERIMENTAL APPROACHES

There are few techniques to assess the degree to which voluntary isometric muscle
performance approaches the optimum. The simplest is to measure the maximal muscle force
and/or EMG activity during a voluntary task, but these are inadequate measures because the
true maximal output is unknown. Optimal EMG and force for most muscle groups are likely
to be slightly greater than actually observed, and this would lead to an unknown degree of
overestimation of voluntary activation. When muscles are not fatigued this overestimation
is probably small in normal subjects, but when measurements are made in patients during
exercise, it may be much larger. Indeed, the wide normal range of maximal voluntary strength
for all muscle groups means that weakness is difficult to establish with a measurement at
one point during the evolution of a disease. Also, voluntary strength within the normal range
may lead to the false conclusion that voluntary activation is adequate. An example of the
latter error comes from estimation of maximal inspiratory muscle strength in patients with
moderately severe asthma. Strength is within the normal range (e.g., McKenzie & Gandevia,
1986), but surprisingly, a significant proportion of asthmatics cannot achieve optimal
performance of their diaphragm when assessed with twitch interpolation (Allen et aI.,
1993b). Another example of fallacious measurement of maximal neural drive to muscles
comes from measurement of the efferent phrenic neurogram in animal experiments: even
when the chemical stimulus to ventilation is massive, this estimate of drive is well below
maximal, as judged by transdiaphragmatic pressures (Sieck & Fournier, 1989). Hence
un tethered estimates of force, EMG or even the neurogram are flawed for measurement of
maximal voluntary output.
Mosso (1904), Merton (1954), and Bigland and Lippold (1954) correctly perceived
that maximal voluntary force could only be judged by reference to what can be achieved by
non-voluntary electrical stimulation of motor output. The first approach was to compare the
force in tetanic stimulation with that in voluntary contraction. These forces are similar for
thumb adduction achieved by voluntary effort and by tetanic stimulation of the ulnar nerve
(Merton, 1954; see also Bigland & Lippold, 1954) although the biomechanics of thumb
movements are such that it is not difficult to exceed the tetanic force voluntarily (Gandevia,
unpublished observations). Stimulus rates above 100 Hz are said to be needed to produce
the largest tetanic force for this muscle group (Marsden et aI., 1983).
Interpolation of an electrical stimulus to the motor nerve during a maximal effort
should produce an increment in force at the appropriate latency either if the stimulated axons
are not all recruited voluntarily or if they are discharging at sub-tetanic rates. This technique
has now been applied not only to intrinsic muscles of the hand (Merton, 1954; Gandevia &
McKenzie, 1988) but also to proximal muscles such as quadriceps and elbow flexors (e.g.,
Bigland-Ritchie et aI., 1978; Bigland-Ritchie et aI., 1983; McKenzie & Gandevia, 1991;
284 S. C. Gandevia et al.

A B

100 ''':j.!.I:''
I •• : I I I I Control twitch
I: !'
90
Voluntary
acti vation 80
%
70

60 I.....--.-'
Control subjects lOOms

Figure 1. A, percentage of voluntary activation of biceps brachii (brachialis) during maximal elbow flexion
measured in 10 normal subjects using twitch interpolation. Each vertical column of dots represents 10 scores
for a single subject (i.e., one dot per score). B, raw traces for subject 10 (with asterisk) from A. One control
twitch (large response) evoked by supramaximal stimulus delivered over relaxed elbow flexor muscles and
10 responses evoked by stimuli delivered at the peak force during attempted maximal elbow flexion.
Amplitudes of the evoked response are expressed as a fraction of the control twitch amplitude and then
subtracted from I and expressed as a percentage to give a voluntary activation score as shown in A. This data
is from a subject with a very high level of voluntary activation.

Lloyd et aI., 1991), tibialis anterior and/or soleus (Gandevia & McKenzie, 1988; Belanger
& McComas, 1981) and even respiratory muscles (Bellemare & Bigland-Ritchie, 1984;
Gandevia & McKenzie, 1985; Allen et aI., 1993b). The usual context for its usage has been
to show that the MVCs are rather close to optimal and any deficit, if detected (see below) is
ignored.
The clinical utility of the twitch interpolation method for detection of functional
components to muscular weakness has long been noted (e.g., McComas et aI., 1983; Lloyd
et aI., 1991; Jacobsen et aI., 1991). However, rarely has the sensitivity of the method been
assessed. Unless separate amplification of DC-offset twitches and their superimposition on
control twitches is used to show the electromechanical delay (e.g., Hales & Gandevia, 1988;
their Fig. I), the technique will not be sensitive enough to reveal whether more than 95% of
the optimal force (from the stimulated axons) has been achieved voluntarily. This is clear
from our measurements of Merton's original figure (Fig. 2). However, with appropriate
techniques, superimposed twitches of <1 % of the size of control potentiated twitches can be
detected, and it is doubtful that further technical improvement would have much physiologi-
cal significance. Pitfalls with the twitch interpolation technique abound (Gandevia, 1992).
For example, stimulation must be constant and activate no axons producing antagonist
torque. Therefore, common peroneal nerve stimulation is technically inadequate for twitch
interpolation for ankle dorsiflexors, and ulnar stimulation is not accurate for abduction/ad-
duction of the fingers. Sensitivity of the technique will be higher when a large twitch or brief
tetanus is interpolated. In addition, the compliance of the myograph used to measure force
must be minimal (Loring & Hershenson, 1992).
For twitch interpolation, the stimulus or stimuli may be applied to the nerve or
intramuscular nerve fibers of the relevant muscle. Transcranial magnetic or electrical
stimulation can also be used to recruit relevant motor axons (Marsden et aI., 1980) or
alternatively, to show that not all the potential synergists are driven maximally in complex
muscle actions, depending on the task (Gandevia et aI., 1990b). A new approach has been to
examine the transcranially evoked compound muscle action potentials before and after
exercise (Brasil-Neto et aI., 1993; see also Gandevia et aI., 1994).
Conventional twitch interpolation has revealed that although most subjects have the
ability to activate motoneuron pools to a high level (voluntary activation scores::::: 95%), this
Central Fatigue 285

Adductor pollicis

sensitivity - 95%

sensitivity> 99%
'--------'
lOOms

Figure 2. Example of the differing sensitivities of the twitch interpolation technique. On the left are shown
the original traces from Merton (1954), on the right our more sensitive version of the technique. One large
control twitch evoked by stimuli delivered to a relaxed muscle is shown with superimposed twitches (smaller
responses) evoked at the peak force during maximal voluntary contractions. High resolution of superimposed
twitches may reveal failure in voluntary drive previously undetected. As calculated from the original figure,
Merton's system allows detection of a failure of 5% from optimal drive, with the more sensitive technique
able to resolve failure of voluntary drive ofless than 1%.

level is not consistently 100%. (This is not surprising given that the fluctuations in voluntary
force during a MVC are absent in a fused tetanic contraction.) The elbow flexors are a muscle
group for which high levels of maximal drive can be achieved with minimal training and
hence they are appropriate for exercise testing (e.g., Lloyd et aI., 1991; Allen et aI., 1993a).
Even so, on repeated testing under ideal conditions some normal subjects activate the muscle
fully in only 2-5% of contractions. If 10 briefMVCs are performed with a sufficient interval
to prevent peripheral fatigue, 95% of the population will have a median level of voluntary
activation for bicepslbrachialis at or above 92%. Data from 10 typical normal subjects are
shown in Fig. I. Furthermore, this level will not improve significantly with repeated testing
(Allen et aI., 1993a). It is as if this level is set at a high but characteristic level in each subject.
Not surprisingly, the level of maximal voluntary drive varies between muscles. In direct
comparisons it was lower for the diaphragm than for the elbow flexors (Allen et aI., 1993b),
and lower for the plantar-flexors than dorsiflexors of the ankle (Belanger & McComas,
1981).
Transcranial stimulation ofthe motor cortex has recently revealed that motor cortical
output is close to optimal for the elbow flexors at the onset of maximal contractions when
they are not fatigued, but that it is suboptimal as central fatigue develops (Gandevia et aI.,
1995). Ultimately, it will be necessary to focus on the factors effectively upstream of the
motor cortex because they can clearly limit performance.

VARIATION IN MAXIMAL VOLUNTARY DRIVE

Voluntary activation will be suboptimal when the muscle, or nearby skin provides a
nociceptive input either due to disease, recent surgery (Rutherford, et aI., 1986) or deliber-
ately induced painful restraint ofthe limb (Gandevia & McKenzie, 1988). Interestingly, when
joint effusions were produced artificially in normal subjects, maximal voluntary knee torques
were reduced (e.g., Wood et aI., 1988). Many reflex inputs can reduce motoneuronal output,
and while the pathways for this are not established, oligo synaptic spinal pathways are likely
contributors.
286 S. C. Gandevia et al.

Limb immobilization produces a disproportionate reduction in maximal voluntary

forces compared with twitch and tetanic forces (e.g., Duchateau & Hainaut, 1987). Although
twitch interpolation was not used, this result, if confirmed, implies that voluntary usage is
necessary to retain the ability to produce high levels of drive to the motoneuron pool.
Levels of voluntary activation have not been measured rigorously under environ-
mental extremes. However, at simulated altitude (P02, 282 torr) some subjects had poorer
neural drive to ankle dorsiflexors, although the effect was not well quantified (Gamer et aI.,
1990). In extreme heat, muscle performance diminishes, perhaps in part because twitch
contraction times shorten requiring higher motoneuronal discharge frequencies, although
other central factors probably contribute (cf. Bigland-Ritchie et aI., I 992b).
Finally, if the results obtained for isometric contractions are to be generalized, we
need to know the extent to which high levels of voluntary drive can be achieved (in the
absence of fatigue) during tasks involving many muscle groups and during non-isometric
tasks. When untrained subjects contract multiple muscle groups isometrically, lower levels
offorce and EMG may occur than for contractions of the individual muscles (e.g., Howard
& Enoka, 1991) and this has been confirmed with twitch interpolation (Herbert & Gandevia,
unpublished observations). While it is technically more demanding to apply twitch interpo-
lation during concentric contractions, some results suggest that voluntary drive can be high
(see Newham et aI., 1991).

EXERCISE-INDUCED REDUCTIONS IN VOLUNTARY DRIVE:

CENTRAL FATIGUE

As indicated above, during attempted maximal contractions ofthe unfatigued elbow

flexors, the level of voluntary activation measured by twitch interpolation is about 95% for
brief efforts of 2-3 s duration. How does this change when the muscle contracts for longer
and develops peripheral fatigue?
Thomas and colleagues (1989) used prolonged isometric contractions of FDI and
tibialis anterior. They found that while voluntary drive showed some diminution, it could be
restored by an additional extra effort. This paralleled earlier findings (Bigland-Ritchie et aI.,
1978) in which central fatigue was observed with sustained quadriceps contractions in 4 of
9 SUbjects. Inspection of their records suggests that central fatigue probably occurred in them
all. Maximal contractions of elbow flexors sustained for 3 min are accompanied by a
progressive reduction in force. While voluntary drive is initially high (>95%), it usually
declines and becomes more variable with time (Gandevia & McKenzie, 1993; text deleted
Fig. 3). In an alternative approach subjects performed intermittent MVCs of 10 s duration
with duty cycles less than 50%. These were also accompanied by concomitant peripheral

Superimposed :
twitch
1_ _ _ _ _ _ _ _
(%MVC) ~ Figure 3. Diagrammatic representation of the
force decline during a sustained maximal el-
100 bow flexion (3 min duration) which occurs
80 with a simultaneous increase in the superim-
posed response evoked by supramaximal stim-
MVC 6

:f\-
(% initial) 0 uli over biceps brachii from highly-trained
subjects. An example of decline in peripheral
twitch amplitude across the 3 min is shown at
bottom and the accompanying increase in re-
o0 3 min sponses superimposed on the maximal effort.
Central Fatigue 287

Voluntary activation Voluntary activation Voluntary activation

100% 87% 85%

lOOms
contraction #3 contraction #18

Figure 4. Data from one subject during a fatigue protocol (50% duty cycle). The larger response is that evoked
by supramaximal paired (10 ms interval) stimuli in the relaxed muscle (control), the smaller (shown at
increased gain X2) is that evoked by the stimuli delivered during a MVC. Stimuli are delivered at the arrow.
Note the decline in control twitch amplitude which occurs as the response evoked by paired stimuli during a
MVC to the elbow flexors increases. Voluntary activation, initially at 100%, declines as the contraction number
increases. This illustrates the simultaneous development of both peripheral and central fatigue (data redrawn
from McKenzie & Gandevia, 1991).

and central fatigue (McKenzie & Gandevia, 1991). Data from a typical experiment are given
in Fig. 4: voluntary activation was 100% in the trial contraction but declined thereafter.
Submaximal exercise at 30% MVC (60% duty cycle) produces a small but measurable degree
of central fatigue (e.g., Lloyd et aI., 1991). However, when this level of drive is plotted
together with that for the maximal voluntary force, it is evident that perhaps half of the
voluntary force decline after 45 min of exercise may be attributed to central fatigue.
Isokinetic exercise is also accompanied by the development of central fatigue (e.g., Newham
et aI., 1991).
The development of central fatigue is accompanied by changes in the excitability of
the human motor cortex (Gandevia et aI., 1994). Evidence for simultaneous reductions in
the thresholds for excitation and inhibition has been obtained. However, once central fatigue
has developed, motor cortical output is less than optimal and, recent studies have shown that
the changes in cortical excitability can be dissociated from the presence of central fatigue
(Gandevia et al., 1995). Hence, it is unlikely that the changes in cortical excitability per se
are necessarily the direct cause of central fatigue.

CONTRIBUTION OF MOTOR UNIT FIRING RATES

It is obvious that to obtain optimal output from a muscle all motor units must be
driven at or above a rate sufficient to produce a fused tetanus. Indeed, motor unit firing rates
are often below 20 Hz and rarely beyond 50 Hz in maximal isometric efforts, not rates that
would necessarily imply full tetanic fusion of all motor units (even if they were all recruited)
(Bellemare et aI., 1983; Gandevia et aI., 1990a; Thomas et aI., 1991; Vander Linden et aI.,
1991; Big1and-Ritchie et aI., 1992a). However to exceed the rate which can just achieve
maximal force output risks failure of neuromuscular transmission.
One concept that has gained support is that motor unit firing rates are appropriately
matched to the contractile machinery's twitch properties. This arose initially from the
realization that red and white muscles subserved different tasks, and had different contractile
speeds. There is also heterogeneity of contractile speed between units within muscle and
between muscles. Moreover, contractile properties of muscle vary with temperature, con-
traction history and fatigue.
288 S. C. Gandevia et al.

/voluntary
100
80 'stimulated
Force 60
(% of maximum) 40

20
quadriceps
O+----.---.----r---~--~
o 20 40 60 80 100
Frequency (Hz)

Figure S. Force-frequency curves for stimulated (open circles) and voluntary contractions (closed circles). Up
to 50% MVC, the stimulated and voluntary curves are similar, but above this the voluntary contractions
generated higher forces at a particular frequency of excitation. Note the almost linear increase in force and
motor unit discharge frequency for voluntary but not stimulated contractions (data redrawn from Bigland-
Ritchie & Rice, 1994).

The properties of the force-frequency relationship require that motor unit firing
frequencies remain in the ascending portion of the curve to allow frequency modulation of
force (see Fig. 5). There is a natural tendency for a decline in firing rate of motor units during
maximal voluntary contractions (e.g., Marsden et aI., 1969; see also Grimby et aI., 1981;
Fig. 6). Evidence exists for the mechanisms that allow this alteration. First, motoneurons
subjected to intracellular current injection discharge with progressively lower frequencies,
a property linked appropriately to their type (Kernell & Monster, 1982). However, it is
unlikely that this mechanism occurs during natural contractions (cf. Gandevia et aI., 1990a;
Macefield et aI., 1993). Second, Bigland-Ritchie and colleagues found that motor unit firing
rates diminish during single maximal isometric contractions which produce contractile
slowing, and that this diminution remains so long as the muscle is held ischemic with a cuff
inflated above systolic pressure (Woods et aI., 1987). This suggests that while only minimal
central fatigue developed, the reflex reduction in motor unit firing rate would match the

30

Motor unit 20
discharge rate
(Hz)
- spindle
10 firing rate

~a1
o --~.--~.r---T'--~
o 40
r.

10 20 30
Time(s)
--------MVC-------
Figure 6. Motor unit firing rate declines in three human muscles with fatigue induced by a sustained maximal
contraction. For adductor pollicis, mean firing rate declined by 42% over a 40 s contraction (Bigland-Ritchie
et aI., 1983), in quadriceps mean firing rate declined by 38% over 40 s (Woods et aI., 1987) and for first dorsal
interosseous, the mean motor unit firing rate declined by 39% over a 30 s contraction (Gandevia, et aI., 1990).
Muscle spindle firing rate (- - - - ) also declines in isometric contractions at 20-60% MVC (Mace field et aI.,
1991). Decline in amplitude of response to stimulation over the relaxed muscle is shown diagrammatically at
bottom of figure.
Central Fatigue 289

slowing of contractile speed. This would preserve flexibility in the operating point on the
force-frequency relationship such that units did not continue to discharge at supratetanic
rates.
In contrast, during intermittent isometric contractions at 30% MVC (60% duty
cycle, Vellestad et aI., 1988; Lloyd et aI., 1991) peripheral fatigue develops but motor
unit firing rates are relatively well preserved. This occurs despite a tendency for contractile
speed to increase and for some central fatigue to develop (Lloyd et aI., 1991). Clearly,
the relationship between muscle speed and motor unit firing rate during exercise will be
complex in the face of potentiation, temperature changes, peripheral and central fatigue,
and the variable behavior of individual motor units. It will be further complicated during
dynamic exercise.
A new approach has involved recording the discharge of motor cortical cells in the
monkey during phasic isotonic/isometric elbow flexion: reductions in discharge rates were
detected for some cells (Maton, 1991). An alternative approach has been to study the
discharge of motor units during sustained contractions in the absence of any peripheral
feedback. This has been achieved for motor units supplying intrinsic hand muscles (Gandevia
et aI., 1990a) and tibialis anterior (Macefield et aI., 1993) by recording with a microelectrode
from a fascicle proximal to a complete nerve block produced by injection oflocal anesthetic.
In these circumstances, subjects can maintain a constant output to the paralyzed muscle
provided that ancillary auditory or visual feedback of the neurogram is provided. However,
most importantly, the motor unit firing rates (across a range of efforts, including maximal
ones) are lower during the blockade of peripheral afferents, implying that fusimotor-medi-
ated spindle facilitation is necessary to achieve peak rates. Secondly, the motor unit discharge
rates in sustained efforts do not show the progressive decline noted for intact axons. This
implies that alterations in reflex input normally contribute to this decline (see also Gooch et
aI., 1990).
Both a decline in spindle facilitation, as directly observed in nerve recordings,
and an increase in the inhibition mediated by group III or IV afferents could contribute
to the reduction in motor unit firing rates with isometric contractions (Hayward et aI.,
1988). Ample evidence exists for activation of small-diameter group III and IV muscle
afferents by muscle contraction, with group III afferents discharging more in response
to static than dynamic contractions (Kaufman et aI., 1984b) and group IV afferents
discharging more under ischemic conditions (Paintal, 1960; Kaufman et aI., 1984a). These
inputs may provide late facilitation to fusimotor neurons when contractile force declines
(Ljubisavljevic et aI., 1992), but it is probably insufficient to affect spindle sensitivity
such that the tendon jerk is restored to control levels following fatigue (Balestra et aI.,
1992; see also Kukulka et aI., 1986). The observation that muscle vibration of a type
likely to activate muscle spindles can increase force during a sustained MVC suggests
that additional peripheral facilitation can replace the reduced spindle input (Bongiovanni
& Hagbarth, 1990). However, this would be expected whether motoneuronal output had
diminished due to a loss of reflex or descending excitation.
The neural substrate exists for potent modulation of motoneuronal output during
sustained contractions. Future studies will no doubt tease out the relative reflex contributions
of different afferent classes, the irradiation of relevant reflexes, and the temporal gating of
excitation reaching the motoneuronal pool. It is all very well to know how the discharge of
the various muscle afferents alters, but the effectiveness of these inputs at the motoneuron
will need to be assessed under a range of conditions. For example, studies in the cat indicate
that Ib homonymous inhibition may be gated out by presynaptic inhibition during electrically
induced contractions (Lafleur et aI., 1992) and studies in humans show that Ia excitation
may be reduced through presynaptic inhibition (Meurnier & Pierrot-Deseilligny, 1989).
290 S. C. Gandevia et al.

CENTRAL FATIGUE IN A BROADER CONTEXT

Our studies began with the use of modified twitch interpolation to show that the
phrenic motoneuron pool could be driven by volition to produce the near optimal force for
the diaphragm (Gandevia & McKenzie 1985). These studies complemented those of Bigland-
Ritchie which provided similar conclusions for combined inspiratory and expulsive maneu-
vers (Bellemare & Bigland-Ritchie, 1984). Initially, we interpolated brieftetani to enhance
the size of the force increments during maximal efforts and subsequently devised a special
amplifier to capture the small superimposed twitches evoked by single stimuli (Hales &
Gandevia, 1988). The more carefully we looked, the more clear it was that voluntary
activation was less than optimal. In limited training sessions, this capacity remains un-
changed (Allen et aI., 1993a), although others have exploited the potential capacity to
improve activation with merely imagined effort (Yue & Cole, 1992). As variable voluntary
activation occurs in virtually every normal subject over a range of ages and athletic abilities,
it is not an oddity but biological reality. If it was not recognized previously, this reflects in
part the available methods.
Just because a biological phenomenon can be measured does not establish its
importance. One implication from the studies on suboptimal drive to the motoneuron pool
with an unexercised muscle is that some motoneurons during an MVC are not on the plateau
oftheir force- frequency relationship. This assumes that all motoneurons have been recruited,
a possibility which seems likely at least for distal muscles given that they tend to use
predominantly frequency modulation to increase force (De Luca et aI., 1982). Indeed, from
25 - 100% MVC, motor unit firing frequency increases linearly without evidence for a
plateau (Bigland-Ritchie et aI., 1992a; Bigland-Ritchie & Rice, 1994; their Fig. 5). Thus,
motor unit frequencies and hence force must remain sensitive to any change in descending,
interneuronal or reflex drive impinging on the motoneurons. Apart from preserving the
flexibility of the force-frequency relationship, is there anything else that may explain or
provide a functional correlate with the observation of suboptimal activation and central
fatigue? When might the phenomenon have pathophysiological or evolutionary signifi-
cance?
First, a neural limitation on the upper bounds of muscle output could reside at a motor
cortical level or upstream of it (Gandevia et aI., 1995). This would act to limit recruitment
of many muscle groups simultaneously and limit the progress of peripheral fatigue. It may
not be coincidental that athletes approach but rarely achieve their theoretical maximal
breathing capacity during severe exercise. A second argument is suggested by observations
in asthmatic subjects: some fail to achieve optimal output to their diaphragm during
self-selected maximal efforts despite greater use of this muscle due to increased airway
resistance (Allen et aI., 1993b). Were this to occur during severe airway narrowing, those
with the lowest levels of voluntary drive could be predisposed to acute ventilatory failure
and death. Hence the capacity to drive some muscles voluntarily at a high level may have
survival value. By analogy, central drive to major limb muscles may have survival necessity.
The concept that central fatigue may develop for the diaphragm in a dramatic way during
loading has been demonstrated during expulsive contractions when the diaphragm and
abdominal muscles are coactivated to elevate abdominal pressure (Bellemare & Bigland-
Ritchie, 1987; McKenzie et aI., 1992). However, the diaphragm shows much less central
fatigue when purely inspiratory contractions are required. Therefore, the development of
central fatigue is dependent on the task (McKenzie et aI., 1992). This is not especially
surprising: the selection of motor cortical neurons and the setting of spinal and long-latency
reflexes behave in a task-dependent way (Porter & Lemon, 1993).
Central Fatigue 291

Thirdly, discussion so far has focused on signs consistent with central fatigue during
different forms of exercise. Why do we stop exercise: is it because we reach theoretical limits
to cardiac, pulmonary or muscle performance? In a detailed statistical analysis of results
from subjects undergoing cycle ergometry, Killian (1992) concluded that normal subjects in
attempted maximal exercise, cycle at heart rates, levels of ventilation and even levels of
symptom intensity (breathlessness and leg effort) which are submaximal. He likened this to
giving up when a certain threshold for tolerance had been exceeded. To express the data in
another way: central fatigue had terminated the exercise.
Finally, these arguments focus the spotlight on what is known about other supraspinal
processes specific to muscle fatigue. It is not simple to establish whether a change in any
aspect of CNS function reflects the contribution to the central fatigue or is caused by it.
Nonetheless, we are at a stage where these questions need to be addressed. Already, changes
in premovement potentials have been reported with isometric fatigue (Fruede & Ullsperger,
1987) although what caused them to change is unclear. However, it is tempting to ascribe
them to the greater subjective effort and recruitment of synergists which occurs during
repeated efforts at 80% MVC. Motivation alters with exercise and may directly or indirectly
change exercise performance (Heyes et aI., 1985; Chaouloff, 1991).
During exercise, changes occur at all steps in force production from upstream of the
motor cortex to the motoneuron, and peripherally from the muscle mitochondria to the
myofibril. It would be surprising and biologically unsound if force production always failed
because of an impairment at one step in this chain. Evolutionary design dictates that multiple
steps will tend to fail together; but ultimately, it is the central nervous system which may
decide when enough is enough.

ACKNOWLEDGMENTS

The authors wish to thank Ms. Jane Butler and Dr. Janet Taylor for comments on the
manuscript. Our laboratory has been supported by National Health & Medical Research
Council of Australia, and the Asthma Foundation of New South Wales, Australia. Attendance
of the authors at the 1994 Bigland-Ritchie conference was supported, in part, by the Prince
of Wales Medical Research Institute, Australia ([Link].), the Muscular Dystrophy Associa-
tion (USA; S. C. G. and [Link].), and the University of Miami ([Link]. and S. C. G.).

REFERENCES
Allen OM, Oandevia SC & McKenzie DK (1993a). Accurate measurements of maximal strength and maximal
drive with twitch interpolation. Electroencephalography and Clinical Neurophysiology 86, 52P.
Allen OM, McKenzie DK, Oandevia SC & Bass S (1993b). Reduced voluntary drive to breathe in asthmatic
subjects. Respiration Physiology 93, 29-40.
Balestra C, Duchateau J & Hainaut K (1992). Effects of fatigue on the stretch reflex in a human muscle.
Electroencephalography and Clinical Neurophysiology 85, 46-52.
Belanger AY & McComas AJ (1981). Extent of motor unit activation during effort. Journal of Applied
Physiology: Respiratory, Environmental & Exercise Physiology 51,1131-1135.
Bellemare F & Bigland-Ritchie B (1984). Assessment of human diaphragm strength and activation using
phrenic nerve stimulation. Respiration Physiology 58, 263-277.
Bellemare F & Bigland-Ritchie B (1987). Central components of diaphragmatic fatigue assessed by phrenic
nerve stimulation. Journal ofApplied Physiology 62,1307-1316.
Bellemare F, Woods II, Johansson R & Bigland-Ritchie B (1983). Motor-unit discharge rates in maximal
voluntary contractions of three human muscles. Journal ofNeurophysiology 50, 1380-1392.
Bigland B & Lippold OCl (1954). Motor unit activity in the voluntary contraction of human muscles. Journal
ofPhysiology (London) 125,322-335.
292 S. C. Gandevia et al.

Bigland-Ritchie B, Dawson NJ, Johansson RS & Lippold OCJ (1986). Reflex origin for the slowing of
motoneurone firing rates in fatigue of human voluntary contractions. Journal ofPhysiology (London)
379,451-459.
Bigland-Ritchie B, Furbush FH, Gandevia SC & Thomas CK (1992a). Voluntary discharge frequencies of
human motoneurons at different muscle lengths. Muscle & Nerve 15,130-137.
Bigland-Ritchie B, Johansson R, Lippold OCJ & Woods JJ (1983). Contractile speed and EMG changes during
fatigue of sustained maximal voluntary contractions. Journal ofNeurophysiology 50,313-324.
Bigland-Ritchie B, Jones DA, Hosking GP & Edwards RHT (1978). Central and peripheral fatigue in sustained
maximum voluntary contractions of human quadriceps muscle. Clinical Science and Molecular
Medicine 54,609-614.
Bigland-Ritchie B & Rice CL (1994). Comparison of force-frequency relations in human voluntary and
stimulated contractions. Proceedings of the Physiological Society, C103.
Bigland-Ritchie B, Thomas CK, Rice CL, Howarth JV & Woods JJ (1992b). Muscle temperature, contractile
speed, and motoneuron firing rates during human voluntary contractions. Journal ofApplied Physi-
ology 73, 2457-2461.
Bongiovanni LG & Hagbarth K-E (1990). Tonic vibration reflexes elicited during fatigue from maximal
voluntary contractions in man. Journal ofPhysiology (London) 423, 1-14.
Brasil-Neto JP, Pascual-Leone A, Valls-sole J, Cammarota A, Cohen LG & Hallet M (1993). Postexercise
depression of motor evoked potentials: a measure of central nervous system fatigue. Experimental
Brain Research 93,181-184.
Brouwer B, Ashby P & Midroni G (1989). Excitability of corticospinal neurons during tonic muscle contrac-
tions in man. Experimental Brain Research 74, 649-652.
ChaouloffF (1991). Cerebral monoamines and fatigue. In Atlan G, Beliveau L, Bouissou P (eds.), Muscle
Fatigue: Biochemical and Physiological Aspects, pp. 234-240. Paris: Masson.
De Luca CJ, Lefever RS, McCue MP & Xenakis AP (1982). Behaviour of human motor units in different
muscles during linearly varying contractions. Journal of Physiology (London) 329, 113-128.
Duchateau J & Hainaut K (1987). Electrical and mechanical changes in immobilized human muscle. Journal
ofApplied Physiology 62, 2168-2173.
Fruede F & Ullsperger P (1987). Changes in Bereitschaftspotential during fatiguing and non-fatiguing hand
movements. European Journal ofApplied Physiology and Occupational Physiology 56, IOS-108.
Gandevia SC (1992). Some central and peripheral factors affecting human motoneuronal output in neuromus-
cular fatigue. Sports Medicine 13, 93-98.
Gandevia SC, Allen GM, Butler JE & Taylor JL (199S). Supraspinal factors in human muscle fatigue: evidence
for suboptimal output from the motor cortex. Journal of Physiology (London) In press.
Gandevia Sc, Butler JE, Allen GM & Taylor JL (1994) Prolongation ofthe 'silent' period following transcranial
magnetic stimulation. Proceedings of the Physiological Society, C138.
Gandevia SC, Macefield G, Burke D & McKenzie DK (1990a). Voluntary activation of human motor axons
in the absence of muscle afferent feedback. The control of the deafferented hand. Brain 113,
1563-1581.
Gandevia SC & McKenzie DK (1985). Activation of the human diaphragm during maximal static efforts.
Journal of Physiology (London) 367, 4S-S6.
Gandevia SC & McKenzie DK (1988). Activation of human muscles at short muscle lengths during maximal
static efforts. Journal of Physiology (London) 407, 599-613.
Gandevia SC & McKenzie DK (1993). Central factors in human muscle performance. Proceedings of the
International Union of Physiological Sciences 122.6/0.
Gandevia SC, McKenzie DK & Plassman BL (l990b). Activation of human respiratory muscles during
different voluntary manoeuvres. Journal of Physiology (London) 428, 387-403.
Gamer SC, Sutton JR, Burse RL, McComas AJ, Cymerman A & Houston CS (1990). Operation Everest II:
neuromuscular performance under conditions of extreme simulated altitude. Journal of Applied
Physiology 68, 1667-1172.
Gooch JL, Newton BY & Petajan JH (1990). Motor unit spike counts before and after maximal voluntary
contraction. Muscle & Nerve 13, 1146-1151.
Grimby L, Hannerz J & Hedman B (1981). The fatigue and voluntary discharge properties of single motor
units in man. Journal of Physiology (London) 316, 545-S54.
Hales JP & Gandevia SC (1988). Assessment of maximal voluntary contraction with twitch interpolation: an
instrument to measure twitch responses. Journal of Neuroscience Methods 25, 97-102.
Hayward L, Breitbach D & Rymer Z (1988). Increased inhibitory effects on close synergists during muscle
fatigue in the decerebrate cat. Brain Research 440, 199-203.
Central Fatigue 293

Heyes MP, Garnett ES & Coates G (1985). Central dopaminergic activity influences rats ability to exercise.
Life Sciences 36,671-677.
Hill AV (1926). Muscular Activity. Baltimore: Williams & Wilkins.
Howard JD & Enoka RM (1991). Maximum bilateral contractions are modified by neurally mediated interlimb
effects. Journal ofApplied Physiology 70, 306-316.
Jacobsen S, Wildschiodtz G & Danneskiold-Samsoe B (1991). Isokinetic and isometric muscle strength
combined with transcutaneous electrical muscle stimulation in primary fibromyalgia syndrome.
Journal ofRheumatology 18, 1390-1393.
Kaufman MP, Rybicki KJ, Waldrop TG & Ordway GA (1984a). Effect of ischemia on responses of group III
and IV afferents to contraction. Journal ofApplied Physiology 57,644-650.
Kaufman MP, Waldrop TG, Rybicki KJ, Ordway GA & Mitchell JH (1984b). Effects of static and rhythmic
twitch contractions on the discharge of group III and IV muscle afferents. Cardiovascular Research
18, 663-668.
Kernell D & Monster AW (1982). Motoneurone properties and motor fatigue. Experimental Brain Research
46, 197-204.
Killian K (1992). Symptoms limiting exercise. In: Jones NL, Killian KJ (eds.), Breathlessness, pp. 132-142.
Hamilton, Canada: Boehringer Ingelheim.
Kukulka CG, Moore MA & Russell AG (1986). Changes in human alpha-motoneuron excitability during
sustained maximum isometric contractions. Neuroscience Letters 68,327-333.
Lafleur J, Zytnicki D, Horcholle-Bossavit G & Jami L (1992). Depolarization ofIb afferent axons in the cat
spinal cord during homonymous muscle contraction. Journal of Physiology (London) 445, 345-354.
Ljubisavljevic M, Jovanovic K & Anastasijevic R (1992). Changes in discharge rate of fusimotor neurones
provoked by fatiguing contractions of cat triceps surae muscles. Journal ofPhysiology (London) 445,
499-513.
Lloyd AR, Gandevia SC & Hales JP (1991). Muscle performance, voluntary activation, twitch properties and
perceived effort in normal subjects and patients with the chronic fatigue syndrome. Brain 114, 85-98.
Loring SH & Hershenson MB (1992). Effects of series compliance on twitches superimposed on voluntary
contractions. Journal ofApplied Physiology 73,516-521.
Macefield G, Hagbarth K-E, Gorman R, Gandevia SC & Burke D (1991). Decline in spindle support to alpha
motoneurones during sustained voluntary contractions. Journal of Physiology (London) 440, 497-
512.
Macefield VG, Gandevia SC, Bigland-Ritchie B, Gorman RB & Burke D (1993). The firing rates of human
motoneurones voluntarily activated in the absence of muscle afferent feedback. Journal ofPhysiology
(London) 471, 429-443.
Marsden CD, Meadows JC & Merton PA (1969). Muscular wisdom. Journal ofPhysiology (London) 200, 15P.
Marsden CD, Meadows JC & Merton PA (1983). "Muscular wisdom" that minimizes fatigue during prolonged
effort in man: peak rates of motoneuron discharge and slowing of discharge during fatigue. Advances
in Neurology 39, 169-211.
Marsden CD, Merton PA & Morton HB (1980). Maximal twitches from stimulation of the motor cortex in
man. Journal of Physiology (London) 312, 5P.
Maton B (1991). Central nervous changes in fatigue induced by local work. In Atlan G, Beliveau L, Bouissou
P (eds.), Muscle Fatigue: Biochemical and Physiological Aspects, pp. 207-221. Paris: Masson.
McComas AJ, Kereshi S & Quinlan J (1983). A method for detecting functional weakness. Journal of
Neurology, Neurosurgery & Psychiatry 46, 280-282.
McKenzie DK, Bigland-Ritchie B, Gorman RB & Gandevia SC (1992). Central and peripheral fatigue of
human diaphragm and limb muscles assessed by twitch interpolation. Journal ofPhysiology (London)
454, 643-656.
McKenzie DK & Gandevia SC (1986). Strength and endurance of inspiratory, expiratory, and limb muscles
in asthma. American Review of Respiratory Disease 134, 999-1004.
McKenzie DK & Gandevia SC (1991). Recovery from fatigue of human diaphragm and limb muscles.
Respiration Physiology 84, 49-60.
Merton PA (1954). Voluntary strength and fatigue. Journal of Physiology (London) 123, 553-564.
Meumier S & Pierrot-Deseilligny E (1989). Gating of the afferent volley of the monosynaptic stretch reflex
during movement in man. Journal of Physiology (London) 419, 753-763.
Mosso A (1904). Fatigue. London: Sonnenschein & Co.
Newham DJ, McCarthy T & Turner J (1991). Voluntary activation of human quadriceps during and after
isokinetic exercise. Journal ofApplied Physiology 71, 2122-2126.
Paintal AS (1960). Functional analysis of group III afferent fibers of mammalian muscles. Journal of
Physiology (London) 152, 250-270.
294 S. C. Gandevia et al.

Phillips CG & Porter R (1977). Corticospinal Neurones, Their Role in Movement. London: Academic Press.
Porter R & Lemon R (1993). Corticospinal Function and Voluntary Movement. Oxford: Clarendon Press.
Rutherford OM, Jones DA & Newham DJ (1986). Clinical and experimental application of the percutaneous
twitch superimposition technique for the study of human muscle activation. Journal of Neurology,
Neurosurgery and Psychiatry 49, 1288-1294.
Sieck GC & Fournier M (1989). Diaphragm motor unit recruitment during ventilatory and nonventilatory
behaviours. Journal ofApplied Physiology 66, 2539-2545.
Thomas CK, Bigland-Ritchie B & Johansson RS (1991). Force-frequency relationships of human thenar motor
units. Journal ofNeurophysiology 65,1509-1516.
Thomas CK, Woods JJ & Bigland-Ritchie B (1989). Impulse propagation and muscle activation in long
maximal voluntary contractions. Journal ofApplied Physiology 67, 1835-1842.
Vander Linden DW, Kukulka CG & Soderberg GL (1991). The effect of muscle length on motor unit discharge
characteristics in human tibialis anterior muscle. Experimental Brain Research 84, 210-218.
V"llestad NK, Sejersted OM, Bahr R, Woods JJ & Bigland-Ritchie B (1988). Motor drive and metabolic
responses during repeated submaximal contractions in humans. Journal of Applied Physiology 64,
1421-1427.
Wood L, Ferrell WR & Baxendale RH (1988). Pressures in normal and acutely distended human knee joints
and effects on quadriceps maximal voluntary contractions. Quarterly Journal of Experimental
Physiology 73, 305-314.
Woods JJ, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneurone firing rates. Journal ofNeurophysiology 58, 125-137.
Yue G & Cole KJ (1992). Strength increases from the motor program: comparison of training with maximal
voluntary and imagined muscle contractions. Journal ofNeurophysiology 67, 1114-1123.
 21

SINGLE-TRIAL READINESS POTENTIALS

AND FATIGUE

D. Popivanov, A. Mineva, and J. Dushanova

Institute of Physiology
Bulgarian Academy of Sciences
Acad. G. Bonchev Str., BI. 23, 1113 Sofia, Bulgaria

ABSTRACT

The authors propose that the cognitive processes related to internal motivation and volition
(e.g., intention and preparation of a voluntary action), influenced by central fatigue, could
be identified and characterized by cerebral readiness potentials (RP) using methods of
chaotic dynamics. The boundaries of single-trial RP and its successive phases can be detected
by tracking the data dynamics, and are represented by chaotically behaved short EEG
transitions.

INTRODUCTION

The performance of a voluntary movement is preceded by a foreperiod during which

the programming process takes place. The segment of scalp recorded EEG occurring in the
1-2 s period before movement onset, is considered to reflect this process and is known as
the readiness potential (RP). Since the time of their discovery by Kornhuber and Deecke
(1964), the brain potentials preceding the performance of voluntary actions have been
extensively investigated to determine to what extent they reflect the organization of the
voluntary motor acts (e.g., Libet, 1985; Barrett et aI., 1986; Tarkka & Hallett, 1990).
Accordingly, the relations between parameters of the movement and the preceding RP have
been studied. The results of Hashimoto and colleagues (1980) revealed that the amplitude
and duration of the averaged cortical slow potential preceding voluntary movements in
monkeys were positively correlated with increasing loads. Similar results have been pub-
lished for humans by Kristeva and colleagues (1990), who showed a significant effect of the
load on the late amplitudes of the RP during unilateral finger flexion.
The participation of the CNS during muscular fatigue has been investigated at
different levels. There exists data supporting the role of monoaminergic systems in the
perception of fatigue during exercise (Chaouloff, 1991; see Newsholme & Blomstrand,
Chapter 23). Maton (1991) suggested that during the fatiguing task, central adaptive
processes are required to compensate for the fatigue in order to keep the performance as

295
296 D. Popivanov et al.

constant as possible. His results on monkeys indicated a positive correlation between the
force and firing rate of precentral cortical cells, but this correlation did not always increase
with muscular fatigue. Characterizing central fatigue by the increased perceived effort
required to maintain desired workload, Nethery (1991) suggested that there were attentional
influences on the effort sense during prolonged exercise. Thus the participation of cognitive
processes in the sensation of the exertion was proposed. There are also cognitive processes
related to internal motivation and volition (e.g., expectancy of an event, intention and
preparation of a voluntary action) which are accompanied by the specific trend-like pattern
ofRP. To the extent that these processes can be measured by brain electrical activity, it seems
reasonable to expect changes in the RP parameters during compensatory adjustment to
fatigue. The results of Freude and colleagues (1987) showing an earlier onset and increased
amplitude of averaged RP during fatigue, support this assumption. However, the changes of
RP during fatigue have not been widely investigated. This could be due to two main reasons:
I) because of the classical technology to obtain RP, based on averaging, the dynamics of
rapid changes are lost, which makes the method inappropriate for detecting the onset and
time course of fatigue and its possible phases; and 2) the perception of fatigue is a gradual
process and its onset should be detected by tracking the EEG dynamics. However, the
detection ofRP with single trials is difficult because of its low amplitude and low signal-to-
noise ratio. Nevertheless, such an approach has been developed for single-trial RP in order
to identify its onset and successive phases (Popivanov et aI., 1995). It would be expected
that using appropriate methods for detecting the onset of single-trial RP, the onset of fatigue
might also be detected provided the central fatigue changes the RP parameters.
The general feature of RP is its ramp-like shape which may consist of successive
phases. This slow-transient, supposedly governed by a simple mathematical law, could be
the signal related to the preparation of voluntary movement. It means that the estimation of
voluntary movement preparation in the sense of RP identification can be reduced to the
determination of a few parameters. The statistical hypotheses that could be tested to identify
RP are based on the assumption that EEG activity is a stochastic process. Using the stochastic
dynamics approach the following hypothesis could be tested: RP is a sum of a smooth
function and an autoregressive (AR) process having an order similar to the EEG background
activity. Although this hypothesis was not rejected, neither for averaged RP (Popivanov,
1992) nor for single-trial RP (Popivanov et aI., 1995) the dynamic behavior ofEEG preceding
voluntary movement is surely more complicated.
The promising results in the application of nonlinear dynamics to the EEG activity
offer a new approach to the identification of changes in EEG activity preceding voluntary
action. For the sake of simplicity we shall denote these EEG segments as the RP time series.
The description of the dynamic behavior of the RP time series could be done not only in the
time or frequency domain, but also in the phase (state) space (PS). In general, PS is identified
with a topological manifold. Using Takens theorem (Takens, 1981), it is possible to
reconstruct the values xiC t), Xi( t+ T), ... , Xi( t+(M-1)T ) as a vectors in M-dimensional PS, where
M is called an embedding dimension and T is a fixed time increment. These vectors form
the so-called trajectory matrix X. Every instantaneous state of a process is represented by a
set (x" ... ,XM) which defines a point in the [Link] sequence of such points (states) over the
time defines a curve in a PS called a trajectory. As time increases, the trajectories either
penetrate the entire PS or they converge to a lower-dimensional subset. In this case, the subset
is called an attractor. If the dimension of the attractor is a non-integer, i.e., fractal, the attractor
is strange and its properties characterize the so-called deterministic chaos. A comprehensive
review of the modem algorithms of chaotic dynamics, applied in physiology, was published
recently by Elbert and colleagues (1994).
In experimental practice, the data have unknown dynamics, they are noisy and usually
are presented as short time series. Strictly speaking, the dimension is defined as a measure of
Single-Trial Readiness Potentials and Fatigue 297

the number of independent variables needed to specify the state of the dynamical system at a
given instant. In order to distinguish deterministic (or multiply-periodic), chaotic and random
processes, two other dimensions have been involved, namely the correlation dimension D2 and
the statistical dimension S, satisfying the relations D2 ::;; S(M) ::;; M (Broomhead & King, 1986;
Vautard & Ghil, 1989) and M ~ 2 • D2 + 1 (Takens, 1981). Methods are proposed for estimating
D2, S and M, which can help to analyze the dynamical behavior ofRP data.

HYPOTHESES AND AIM

In the following sections we describe our experience in application of chaotic

dynamics (methods and procedures) for detecting the onset of a RP (and its phases) in each
single-trial record, assuming that each of these instances reflects the change in the state of
EEG during the preparation of the voluntary motor act. Further we suppose that these
instances considered as boundaries of more deterministic segments, could be changed by
central fatigue. Our hypothesis is that RP time series consists of successive steady state
segments and shorter transitions between them. While the former could be described by a
simple mathematical law (e.g., by appropriate curve fitting), the latter are chaotic processes
that pre-adjust the system to pass from one stable state to another.

METHODS AND PROCEDURES

The applied procedures comprise two main steps:
1. Checking whether chaotic dynamics exists within the RP time series without
detection of the chaotic segments in time. This includes computing the correlation
dimension D2, nonlinear prediction, and computing the statistical dimension S.
2. Tracking the dynamics of the RP time series in order to detect the instances when
chaotic transients appear. This includes singular spectrum analysis (SSA), non-
linear prediction, point D2 correlation dimension, and nonlinear filtering.
Some examples are shown in Fig. I using RP-data files from experiments in which
subjects performed abrupt flexion of the right hand fingers, voluntarily, at irregular intervals
(Popivanov et aI., 1995). The data were collected using Ag/AgCl electrodes placed at Cz,
Fz, C4, pz and C3 according to the 10/20 system with linked earlobes as a reference.
The amplifiers (Nihon Kohden EEG-43 14) were set to an upper frequency (-3 dB)
of 120 Hz and a time constant of 5 s. A 12 bit ADC with sampling rate of250 samples/s was
used (CED 1401) for EEG, EOG, EMG and mechanogram channels, the latter used to
synchronize the data record to the movement onset. Thus the typical file containing RP time
series consists of 1024 samples, 90% of which were prior to the movement onset.

DETECTION OF A STATIC CHAOTIC BEHAVIOR

Correlation Dimension
One of the measures of deterministic chaos is the so-called correlation dimension
D2. The idea is to compute how the number of pair of vectors Xi which are separated by
distances less then r, changes as a function of r. This is the so-called correlation integral
defined by Grassberger and Procaccia (1983) as C(r,n) - (l/n2). [number of pairs (i,j) for
which IXi-xjl ::;; r] for large number of pairs. D2 is estimated as a slope in the linear region of
298 D. Popivanov et aI.

30

Cz
-40

30

Fz
-40

30

c.4
-40

30

PZ
-40

30

C3
-40

30

EOG
-40

m.m: I
]. 1024

Figure 1. An example ofa single-trial RP records at Cz, Fz, C4, pz and C3, EOG and mechanogram (top to
bottom). Amplitudes in 1lV, 1024 samples, sampling interval 4 IDS (trace duration 4 s). The onset of the
mechanogram is accepted as the onset of a finger flexion.

log C(r,n) vs. log(r) for incrementing embedding M, i.e., D2 = d[logC(r,n)]/d[log(r)] is valid
for increasing embedding M. Thus D2 is estimated by a plot D2(M,r) vs. log(r). It is
concluded that a chaotic attractor is present ifD2(M,r) = f(log(r» is converging to a plateau
with increasing M and D2 is a non-integer. Thus D2 measures how much subspace of the
PS is really occupied by the attractor, which may be less than Mmax. IfD2 is an integer, the
process is multiply-periodic, which is indicated with a torus in PS. For a random process it
is believed that there would be no saturation ofD2 with increasing M, (although this is not
the case for some colored noises). Since the calculation ofD2 for very small r is dominated
by noise, particularly when dealing with experimental data, there is no convergence of the
slope for very small r. In practice, D2 is defined in a scaling region (r[, r2), where the curves
saturate to a flat plateau, which does not change with increasing M. It could be concluded
that D2 is enough to distinguish the chaotic from periodic or random sets of data. However,
the experimental data represent a complex process and its dynamics are unknown. Moreover,
the reliability ofD2 is less for short data sets (and/or mixed with colored noise) as it is in
our case. In some cases the curves D2 vs. log(r) converged asymptotically to a plateau for
RP time series data. However, the behavior ofD2 should be interpreted with caution because
of non stationarity of the data and insufficient data samples. For example, Fig. 2A,B illustrates
the behavior of C(r) vs. log (r) and D2(M,r) vs. log(r) for RP data at Cz, respectively. The
computations were performed for n = 1024 samples, time delay T =4 ms and embedding M
= 2,4, ... ,20. The curves in the second plot do not converge to a plateau, which means that
Single-Trial Readiness Potentials and Fatigue 299

-4 o

Figure 2. Plots of the correlation in-

tegral C(r) vs. log(r) (A) and correla-
tion dimension D2 vs. log(r) (B) for
'i:'
increasing values of the embedding ~-
dimension M (M = 2,4, ... ,20; bottom N
to top). The computations are per- o
formed using 1024 samples, time de-
lay T = 4 ms ofRP time series at Cz.
The slopes of C(r) are not parallel (A)
and the curves D2(M,r) do not con- -4 o
verge to a plateau (B). log(r)

D2 could not be determined exactly. Thus, for RP time series, D2 is not appropriate, moreover
its computation is time consuming. Hence, it should be considered as a qualitative measure.

Nonlinear Prediction
We recommend the method proposed by Sugihara and May (1990), which can detect
chaotic behavior in short data segments. The method is based on a library of past patterns jn a
time series, which is used to compute the prediction of the future pattern. If the time series is
chaotic, the correlation coefficient Ro between predicted and actual values falls down as a
function of a time step Tp; and the normalized error En as a function ofT p is an increasing curve
like a mirror ofRo =f(T p). If En =0, the prediction is perfect; En =1 indicates that the prediction
is not better than a data mean value. Because the above results are sensitive to the chosen
embedding M, the curve Ro = f(M) can show the optimal M, where Ro is max. This value can
be used for further analysis. Fig. 3A shows Ro =f(Tp) and En =f(Tp) for RP time series at Cz.
The same data values were taken as a library; that is, the library and forecasts span the same
time period. This is recommended if some systematic changes (i.e., trends) exist (nonstation-
arity). The steep decay of the upper curve is an indication of chaotic behavior of the data. In
Fig. 3B the two optimal values of M are indicated; that is, M = 4 and M = 7. These results
correspond to the attractors with D2 ::; 3, according to Takens formula M ~ 2 * D2+ 1.

Statistical Dimension
Statistical dimension S is defined as a number of eigenvalues of the matrix (XT. X),
which are above the noise level. Here X is the trajectory matrix for a given embedding M.
Thus the significant part could be distinguished from the noise part of a given data record.
300 D. Popivanov et al.

A B
1 0.972

0
0::

0.6
~ ~
0
0::

0.967
0 Tp 16 0 M 11

V
0.9 0.25

C C
w W

0.2 0.22
0 Tp 16 0 M 11

Figure 3. Top to bottom: A, plot of the correlation coefficient Ro between predicted and observed values as a
function of prediction time for M = 4 and T = i. Plot of the normalized error En as a function of prediction
time. The prediction accuracy decreases with increasing time, which characterizes the chaotic dynamics
generated by some low-dimensional attractor. RP- data are at Cz. The En was computed according to Farmer
and Sidorowich (1987, p. 846). B, plot of the Ro as a function ofM for Tp=\, T=i. Plot of the En as a function
ofM. The highest value ofRo and the lowest En give the optimal M (see text).

We have used the algorithms of Broomhead and King (1986) and Vautard and Ghil (1989),
known as Singular Spectrum Analysis (SSA). SSA gives quantitative and qualitative infor-
mation about the stochastic and dynamic components of the process even if the time series
is short and noisy. Fig. 4 shows normalized eigenvalues spectra log( crlEcrk) vs. k, where k
= 1,2" ... ,M for single RP time series at C3, computed for several embedding dimensions M.
The curves represent decreasing sequences of eigenvalues converging to the noise level. The
end of steep decay determines S for each M. To verify the computed D2 (e.g., the value of
D2 where the plateau is reached) the following relations should be satisfied: D2 ~ S(M) ~
M for 2 • D2 + 1 ~ M ~ Msat , where Msat is the value of M, when D2 reached its saturation
value (see above, Correlation Dimension). IfD2 > S, the time series could contain noise or
its length is too short. In the examples shown above, the corresponding to M values of S are
4,6, 7, 9 and 10. However, the left term of these relations could not be checked.

Figure 4. SSA eigenvalues spectra for

RP time series at C3 computed for M =
-4
\ 0, 15, 20, 25 and 30 (bottom to top). The
1 )0 end of the steep decay determines statis-
order k tical dimension S for each M (see text).
Single-Trial Readiness Potentials and Fatigue 301

DETECTION OF THE CHAOTIC TRANSIENTS IN TIME DOMAIN

Analysis of the Principal Components

The principal components (PC) are defined as the orthogonal projection coefficients
of the original series into their corresponding eigenvectors. Thus PC are the filtered versions
of the original time series. It is accepted that PC corresponding to the eigenvalues above the
noise level, reflect the intrinsic (significant) part of the process, while those corresponding
to the eigenvalues below the noise level have low amplitude and reflect the noise. The
disadvantage of PC is that their length is n - M + 1 (n = number of data samples); that is,
they are shorter than the original data record. In the analysis of RP time series, we used the
so-called reconstructed components (RC) instead of PC, described by Vautard and colleagues
(1992, formulae 2.17a,b,c). RC represent series oflength n corresponding to a given sets of
eigenvalues (selected by the researcher). They are obtained from the PC and its eigenvectors.
From the spectral point of view eigenvectors correspond to data-adaptive moving-average
filters. Using RC, parts of the time series could be extracted, which are characterized
according to some feature selected by the researcher (e.g., amplitude or frequency). Plots of
three significant RCs from RP time series at C4 are shown in Fig. 5. The remaining RC made
a small contribution and were considered as noise. However, this interpretation should be
accepted with caution. It is possible, that the so-called noise contains hidden significant
components, which may indicate chaotic transients between the bursts.

Point-D2 Correlation Dimension

The point-D2 (PD2) estimate of the correlation dimension has been proposed by
Skinner and associates (Molnar & Skinner, 1992, Elbert et aI., 1994) in order for the
correlation dimension to be less sensitive to nonstationarities. The PD2 does not accept every
data point as a reference point but only those that pass two criteria: linear scaling in the
10g(C(r,n,nref)) vs. log(r) plots and convergence of slope vs. M (see above, Correlation

20

RC1
-20

10

RC2
-10

Figure 5. Plots of three RCs from RP time series at C4. The first plot (RC I) represents the trend, which is
obtained from the first principal component. The second plot (RC2) shows the component reflecting the
frequency band around 8 Hz (alpha rhythm). The diminishing of alpha spindle could be seen around I s before
the movement onset, which could indicate the RP onset (PfurtscheIIer, 1992). The third plot (RC3) shows a
similar picture for the frequency band around 16 Hz (RC3). RC2 and RC3 could be interpreted as a detrended
significant parts.
302 D. Popivanov et aJ.

1 1020

Figure 6. Plots of RP time series at Cz (A), absolute error between predicted and actual data values (B) and
PD2 (C). The arrows mark the instances of highest error, which coincide with diminishing PD2. These instances
indicate the possible onset of chaotic transients.

Dimension). Thus the method seeks stationary subepochs of the same type as that in which
the reference vector is located, then tests and rejects those vectors for which linear scaling
and convergence cannot be found. PD2 can be used to track the changes in both time and
dimension, although there is, as yet, no mathematical proof that PD2 is a temporal charac-
terization of the D2. Using mean PD2 for different subepochs containing event-related
potentials (ERP), Molnar & Skinner demonstrated PD2 decreasing around ERP. We used
PD2 to detect the onset ofRP. Fig. 6C shows the plot ofPD2 computed for RP time series
atCz.

Nonlinear Prediction Error

Ifthe absolute error is computed between predicted and actual data values (see above,
Nonlinear Prediction), it can indicate the instants in time that the chaotic transients are
located. It can serve as an additional test for detecting the chaotic transients in time domain
(Smith, 1992). Fig. 6B shows the plot of absolute error for RP time series at Cz. The same
data values were taken as a library, i.e., the library and forecasts span the same time period.

Nonlinear Filtering
A nonlinear digital filtering approach is proposed for tracking the changes in chaotic
features of a nonstationary time series (Saltzberg et aI., 1988). By using the recurrence
relations for logistic, cubic, and tent maps and the least-squares estimation procedure, the
recurrence relations for the filter output were able to detect chaotic transients in short, noisy
time series (Popivanov et aI., 1993). Using this approach and the results for D2 and M
obtained by the methods mentioned above, the logistic map proved to be appropriate for
tracking the dynamics ofRP time series. The filter output recovered the bifurcation parameter
of the logistic map when short overlapping segments of data were fed into the filter. These
segments served as a moving window shifted in steps along the record. The filter output (i.e.,
the bifurcation parameter) was displayed in each step until the end of the record. The output
values over 3.57 indicated the chaotic transients. The details of the procedure are given
elsewhere (Popivanov et aI., 1993). For RP time series, M was selected as a length ofthe
moving window and T was selected as the time decay corresponding to lie of the autocor-
relation function initial value (e - 2.72).
Single-Trial Readiness Potentials and Fatigue 303

DISCUSSION

The application of several methods to test the dynamics of RP time series show that
our hypothesis with regard to single trial RP was not rejected. Chaotic transients were
detected within data records near the same places by different methods. These instances are
marked by arrows in Figs. 5 and 6. It is interesting that the instances of chaotic transients do
not coincide always with the beginning of the local trend components. This means that the
changes in the dynamical behavior of the process is not reflected in the trend components
only. Instead, the key to understanding the intrinsic dynamics could be hidden in the faster
components that are erroneously considered as it noise.
All the methods described above seem to be necessary for tracking the dynamics of
RP data. At this stage of the study, we feel that the preprocessing procedures, including
preliminary tests for existence of chaotic transients, are necessary before the application of
any algorithm for tracking the dynamics. Unfortunately, the computations are very time
consuming. The most important methods are those recommended for short, noisy time series.
They could be used on data segments even when these overlap. However, more research is
needed to select the proper segments and the time lag T (Elbert et ai., 1994).
The results shown above suggest the following: 1) there exists chaotic transients
somewhere within RP data records. However, because of the unknown dynamics of the data
and unknown effect of central fatigue on the single RP, time consuming preprocessing is
needed to obtain evidence of chaotic behavior; and 2) several tests are required to confirm
chaotic behavior of RP data. This is because data is acquired in various experimental tasks,
and chaotic behavior may be task specific.

ACKNOWLEDGMENTS

The authors wish to thank Prof. Douglas Stuart for his encouragement to contribute
this work, and for his and Dr. Robert Lansing's helpful comments. The authors' laboratory
has been supported, in part, by the National Fund for Scientific Research of Bulgaria (grant
B-18/1994).

REFERENCES
Barrett G, Shibasaki H & Neshige R (1986). Cortical potentials preceding voluntary movement: evidence for
three periods of preparation in man. Electroencephalography and Clinical Neurophysiology 83, 327-
339.
Broomhead DS & King GP (1986). Extracting qualitative dynamics from experimental data. Physica 20D,
217-236.
ChaouloffF (1991). Cerebral monoamines and fatigue. In: Atlan G, Beliveau L, Bouissou P, (eds.), Muscle
Fatigue. Biochemical and Physiological Aspects, pp. 234-240. Paris: Masson.
Elbert T, Ray WJ, Kowalik ZJ, Skinner JE, Graf KE & Birbaumer N (1994). Chaos and physiology:
deterministic chaos in excitable cell assemblies. Physiological Reviews, 74, 1-47.
Farmer JD & Sidorowich JJ (1987). Predicting chaotic time series. Physical Review Letters 59, 845-848.
Freude G, Ullsperger P & Pietschmann M (1987). Are self-paced repetitive fatiguing hand contractions
accompanied by changes in movement-related brain potentials? In: Gantchev GN, Dimitrov B, Gatev
P (eds.), Motor Control, pp. 99-103. New York: Plenum.
Grassberger P & Procaccia I (1983). Measuring the strangeness of strange attractors. Physica 9D, 189-208.
Hashimoto S, Gemba H & Sasaki K (1980). Premovement slow cortical potentials and required muscle force
in self-paced hand movements in the monkey. Brain Research 117,415-423.
Komhuber HH & Deecke L (1964). Hirnpotentialanderungen beim Menschen vor und nach Willkurbewegun-
gen, dargestellt mit Magnetbundspeicherung und Ruckwartsanalyse. Pjlugers Archiv 281, 52.
304 D. Popivanov et al.

Kristeva R, Cheyne D, Lang W, Lindinger G & Deecke L (1990). Movement-related potentials accompanying
unilateral and bilateral fInger movements with different inertial loads. Electroencephalography and
Clinical Neurophysiology 75, 410-418.
Libet B (1985). Unconscious cerebral initiative and the role of conscious will in voluntary action. Behavioral
and Brain Sciences 8, 529-566.
Maton B (1991). Central nervous changes in fatigue induced by local work. In: Atlan G, Beliveau L, Bouissou
P (eds.), Muscle Fatigue. Biochemical and Physiological Aspects, pp. 207-221. Paris: Masson.
Molnar M & Skinner JJ (1992). Low-dimensional chaos in event-related brain potentials. International Journal
ofNeuroscience 66,263-276.
Nethery VM (1991). Central fatigue: the contribution ofattentional focus. In: Atlan G, Beliveau L, Bouissou
P (eds.), Muscle Fatigue. Biochemical and Physiological Aspects, pp. 243. Paris: Masson.
Pfurtscheller G (1992). Event-related synchronization (ERS): an electrophysiological correlate of cortical
areas at rest. Electroencephalography and Clinical Neurophysiology 83, 62-69.
Popivanov D. (1992). Time series analysis of brain potentials preceding voluntary movements. Medical &
Biological Engineering & Computing 30,9-14.
Popivanov D, Dushanova J, Mineva A & Krekule I (1993). IdentifIcation of slow EEG-transients by using
chaotic models. In: Rosenfalck A (ed.), Proceedings of the IMIA-IFMBE Working Conference on
"Biosignal Interpretation", pp. 198-201. Aalborg, Denmark.
Popivanov D, Dushanova J, Mineva A & Krekule I (1995). Single-trial readiness potentials: stochastic and
chaotic dynamical behavior. In: Stuart DG, Gantchev GN, GurfInkel VS, Wiesendanger M (eds.),
Motor Control VII, pp. 00-00. Tucson, AZ: Motor Control Press.
Saltzberg B, Burton WD & Skinner IE (1988). Filtering for [Link]: Harris G, Walker C (eds.), Proceedings
of the 10th Annual International Conference of the IEEE in Medicine & Biology Society, pp.
1072-1073. New Orleans, LA
Smith LA (1992). IdentifIcation and prediction of low dimensional dynamics. Physica D58, 50-76.
Sugihara G & May RM (1990). Nonlinear forecasting as a way of distinguishing chaos from measurement
error in time series. Nature 344,734-741.
Takens F (1981). Detecting strange attractors in turbulence. In: Rand DA, Young LS (eds.), Lecture Notes in
Mathematics, vol. 898, pp. 366. Berlin: Springer.
Tarkka 1M & Hallett M (1990). Cortical topography of premo tor potentials preceding self-paced, voluntary
movement of dominant and non-dominant hands. Electroencephalography and Clinical Neurophysi-
ology 75, 36-43.
Vautard R & Ghil M (1989). Singular spectrum analysis in nonlinear dynamics with applications to paleocli-
matic time series. Physica D35, 395-424.
Vautard R, Yiou P & Ghil M (1992). Singular spectrum analysis: A toolkit for short, noisy chaotic signals.
Physica D58, 95-126.
 22

THE SENSES OF EFFORT AND FORCE

DURING FATIGUING CONTRACTIONS

L. A. Jones

Department of Mechanical Engineering, Room 3-148

Massachusetts Institute of Technology
77 Massachusetts Ave, Cambridge, Massachusetts 02139

ABSTRACT

The sensory basis upon which judgments of force are made during fatiguing isometric
contractions is reviewed. At issue is whether the perception of force is based on centrally
generated sensations arising from the motor command, known as the sense of effort, or from
peripheral sensations originating in the muscle, and termed a sense offorce. The results from
a number of studies indicate during sustained constant-force contractions the perceived
magnitude of the force increases, which is consistent with the idea that judgments of force
are based on centrally generated signals. It is, however, possible for some subjects to
dissociate effort and force and to make accurate judgments of the magnitude of forces during
fatigue.

INTRODUCTION

The basis on which judgments of force and weight are made has been subj ect to much
debate and experimentation over the last century (for reviews see McCloskey, 1981; Jones,
1986). At issue, is whether the perception of force is primarily derived from centrally
generated sensations arising from the innervation of the efferent pathways (i.e., sensations
of innervation) as Helmholtz (1866/1925) proposed, or from peripheral sensations originat-
ing in the muscles, skin and joints as Bell (1826) hypothesized. Centrally mediated sensations
are often termed a sense of effort (McCloskey et al., 1974), and are assumed to arise from
internal neural correlates (corollary discharges) of the descending motor command. They
presumably reflect the magnitude of the voluntary motor command generated (McCloskey,
1981). The second source of sensory information, described as a sense offorce or tension,
is derived from peripheral receptors in muscles, tendons and the skin and is presumed to
signal intramuscular tension and, therefore, provides a measure of the actual force exerted
by the muscle (Roland & Ladegaard-Pedersen, 1977). Changes in muscle force are signaled
by the Golgi tendon organs (Crago et al., 1982) and these are the most likely source of a
sense of force (Jami, 1992). The relation between the force generated by a muscle and the
discharge rate of tendon organs does, however, depend on a number of factors (Proske &

305
 306 L. A. Jones

Gregory, 1980; Gregory, 1990; Jami, 1992}, but under natural conditions of muscle activation
it has been reported to be linear (Crago et aI., 1982). Although the senses of effort and force
are often presented as dichotomous systems, they should be considered as complementary
rather than competitive, in that both appear to be involved in mediating the perception of
force.
In most situations, the sense of effort and force provide congruent information; that
is, as the motor command increases, perhaps due to a change in the mass of an object being
supported, so, too, would the discharge rate of muscle receptors signaling the force of
contraction (Crago et aI., 1982; Jami, 1992). Under some conditions, however, it is possible
to dissociate the senses of force and effort by decoupling the relation between the motor
input to the muscle and peripheral feedback, arid in these situations one can evaluate whether
subjects' judgments of force appear to be based on centrally or peripherally arising sensa-
tions. For example, during sustained submaximal constant-force contractions, there is an
increase in the amplitude of the electromyogram (EMG) of the fatiguing muscle, reflecting
the recruitment of additional motor units as contractile failure occurs in those units already
active (Bigland-Ritchie, 1981). Under these conditions one can investigate whether judg-
ments of the force generated by the fatiguing muscle covary with the excitatory input to the
muscle (EMG), which is increasing, or remain essentially constant, consistent with the actual
force being produced.
The smoothed rectified EMG has been used as an indirect measure of the excitatory
input or neural drive sent to a muscle, as it reflects the number of active motor units and their
discharge rates (Cafarelli, 1988). An indeterminate amount of this signal is, however, due to
the activation of alpha motoneurons via the gamma loop and muscle spindle afferent fibers.
The contribution of these segmental reflex pathways to the generation of muscle force is
small, but not insignificant (Stein et aI., 1995), and in the absence of this feedback there is
a substantial decline (about 30%) in the firing rates ofmotoneurons (Gandevia et aI., 1990).
Support for using the surface EMG as an estimate of the descending motor command comes
from a study by Cafarelli and Bigland-Ritchie (1979) in which they varied the force-gener-
ating capacity of a muscle by changing its length and asked subjects to match the forces in
two corresponding muscles held at different lengths. Although the slope of the resulting
force-matching function was proportional to the ratio ofthe maximal voluntary contractions
(MVCs) of the two muscle groups, the relation between the EMGs recorded from the two
muscles was not influenced by muscle length, which suggests that when the excitatory inputs
to the muscles were the same, the sense of effort associated with each contraction was similar.

MEASUREMENT

A variety of methods have been used to study the perception of force ranging from
ratio-scaling procedures (Gescheider, 1985) which require that subjects assign numbers to
forces in proportion to the perceived magnitude of the force (Eisler, 1965), to the contralateral
limb-matching method (McCloskey et aI., 1974) in which the forces exerted by a muscle
group in one limb (the reference limb) are matched in subjective magnitude by contractions
ofthe corresponding muscle group on the contralateral side (the matching limb). The subj ect
usually has feedback of the force produced by the reference limb and attempts to match the
force sensations in both limbs during the matching contraction (Cafarelli, 1982). In addition
to these methods which measure how subjects scale force, sensory thresholds have been
estimated and these provide an index of the sensitivity of subjects to changes in force (Ross
& Reschke, 1982; Pang et aI., 1991). The relation between force discrimination and
judgments of the magnitude offorces is not known, and there is no reason a priori to assume
that the ability to estimate force accurately is related to the capacity to discriminate changes
Sensing Effort and Force during Fatiguing Contractions 307

in force (Holmes, 1917; Jones, 1989). Ross and colleagues (1984) have shown that the
relation is not monotonic in that force discrimination can deteriorate (i.e., thresholds
increase) under conditions in which forces are perceived to be smaller than normal (e.g., in
a O-G environment). The results from a number of studies indicate that the differential
threshold or just noticeable difference for force is approximately 6% (Ross & Reschke, 1982;
Jones, 1989; Pang et aI., 1991).
When subjects make numerical estimates of the perceived magnitude of forces, the
perceived force grows as the 1.7 power of the force exerted. This means that when the force
exerted increases by 50%, perceptually its magnitude doubles (Eisler, 1965; Cain & Stevens,
1971). The exponent of the power function does change with variations in experimental
procedure (range: 0.8-2.0), which affect the way in which observers give judgments, but
when the same procedure is used, consistent results are obtained even when the muscle group
varies (Eisler, 1965; Stevens & Cain, 1970; Jones, 1986). In contrast to this nonlinear relation
between force amplitude and perceived intensity, a linear relation between the reference and
matching forces is obtained with the contralateral limb matching procedure. This result is
consistent with those based on ratio-scaling methods in that they predict that the exponents
of the force functions for the two limbs are identical, and hence the relation linear. The
contralateral limb-matching method avoids some of the response biases associated with
numerical estimation of stimulus magnitude (discussed in Poulton, 1979), and has the
additional advantage that subjects are not able to remember previous responses with the same
degree of accuracy with which numbers can be recalled. This is important in experiments in
which the effects of varying the instructions given to subjects on judgments of sensory
magnitude are studied (e.g., estimate force or effort during a constant force contraction).

FORCE PERCEPTION DURING MUSCLE FATIGUE

Changes in the perceived magnitude of forces occur as a function of the duration of
the muscular contraction. The perceptual changes associated with maintaining a constant
force over time have been studied by asking subjects to estimate numerically the perceived
magnitude of a force sustained for varying periods of time. Stevens and Cain (1970) found
that the perceived magnitude of a handgrip contraction grew as a power function of both the
isometric force exerted (exponent of 1.7) and the duration of the contraction (exponent of
0.7). This means that when subjects are asked to judge the magnitude of forces exerted by
fatigued muscles, there is a uniform ratio increase in the perceived amplitude of the forces,
and so the exponent of the power function relating perceived to physical force does not
change with fatigue (Cain & Stevens, 1971). This finding also indicates that from the point
of view of endurance, it is better to bear a light load for a long period of time than a heavy
load for a short time (Stevens & Cain, 1972).
Using a different psychophysical procedure, known as cross-modal matching, in
which the loudness of tones was matched to the force of handgrip contractions, Teghtsoonian
and colleagues (1977) reported that muscle fatigue was associated with a substantial change
in the exponent of the force power function. In their experiment, fatigue was induced by
requiring subjects to produce five consecutive MVCs prior to the matching trials. After the
fatiguing contractions, tones of higher sound pressure levels were matched with forces that
were smaller than those produced prior to fatigue, and softer tones were now matched with
larger forces. This latter unexpected finding suggests that the absolute threshold for detecting
force may increase with muscle fatigue (Teghtsoonian et aI., 1977). Production of a brief,
presumably non-fatiguing, MVC (i.e., 5 s) does influence the perception offorce, but in this
situation small amplitude forces (between 2 and 4% MVC) are over and not under estimated
in magnitude (Hutton et aI., 1987; Thompson et aI., 1990). This error in estimating force has
308 L. [Link]

l00~---------------- ______ ~

~ 80
~
~ 60

!
.~
40 '--'-.--

~
~
20
Figure 1. The matching forces exerted by
the left arm as one of four constant forces
01-~--~~--~-r--r-~~~~ (30%[.], 45%[.], 60%[&], and 75%[.]
o 30 60 90 120 150 180 210 240 270 MVC) was maintained by the right arm until
Endurance Time (s) maximal endurance (data from one subject).

been attributed to post-contraction potentiation of stretch reflex pathways (Hutton et aI.,

1987) and tendon organ desensitization (Thompson et aI., 1990).
The effect of fatigue on either absolute or differential thresholds for force does not
appear to have been measured. Knowledge about changes in these sensory thresholds would
contribute to our understanding of the performance decrements associated with fatigue. In
the auditory modality, it has been known since the early 1940s that there is an increase in
the detection threshold after prolonged exposure to a (fatiguing) sound (Botte & Scharf,
1980). Similar findings have been reported for visual (Geisler, 1979) and olfactory stimuli
(Cain & Engen 1969). If fatigue does result in an elevation in the detection threshold for
force, then this would affect the control offorce in those tasks highly dependent on feedback
(e.g., fine manual tasks such as threading a small nut on a bolt).
When subjects are asked to judge the magnitude of a sustained constant force by
producing a brief matching contraction on the contralateral side, there is a progressive and
linear increase in the perceived amplitude of the force being maintained, with the rate of
increase dependent on the amplitude of the constant force being exerted, as shown in Fig. I
(McCloskey et aI., 1974; Gandevia & McCloskey, 1978; Jones & Hunter, 1983a, 1983b;
Cafarelli & Layton-Wood, 1986). The overestimation in force as the muscle fatigues is
considerable, with errors at the limit of endurance being in the order of 70% (McCloskey et
aI., 1974) to 130% (Jones & Hunter, 1983a).
In these studies, subjects were simply instructed to make both arms the same or make
the forces produced by the two arms the same, and were unaware during the experiment of
the extent to which they were overestimating the sustained forces (Jones & Hunter, 1983b).
These results suggest that subjects base their judgments of force on a signal related to the
motor command sent to the muscle, or some variable proportional to it, rather than on the
force actually generated by the fatiguing muscle. The overestimation of force results from
efferent signals of comparable magnitude being transmitted to fatigued and unfatigued
muscles, the latter having greater force-generating capacity. This interpretation is supported
by the relation observed by Jones and Hunter (1983b) between the EMG of the fatiguing
muscle and the matching force produced on the contralateral side. As the EMG increased,
due to either a larger force or progressive fatigue in the muscle, there was a corresponding
increase in the magnitude of the matching force as shown in Fig. 2.
It seems unlikely that the overestimation of force during fatigue can be attributed to
the activity of receptors in the muscle. Macefield and colleagues (1991) have shown that
when subjects maintain a submaximal voluntary contraction, there is a decline in the
discharge rate of most muscle spindle afferents and this occurs in the presence of a
Sensing Effort and Force during Fatiguing Contractions 309

100. .- - - - - -_ _ _ _ _ _ _........,

U
;> 80 •• ••
~
•
~
60 • • ••
...t
Q
• ••• •
'"= 40
OIl

:E
Figure 2. Relation between surface EMG re- ~ 20
~
corded from the biceps muscle of the right
arm during constant force contractions main-
tained at 35%[.&], 50%[_] and 65%[e] 0
MVC and the matching forces produced by 500 750 1000 1250 1500 1750 2000
the left arm (data from one subject). Reference Arm EMG (arbitrary units)

progressive increase in the EMG. The data on the responses ofGolgi tendon organ receptors
under these conditions are more limited, but they do suggest that there is a decay in the
discharge rates of some receptors with time (Macefield et ai., 1991). It also seems probable
that their response rates saturate at high forces and that there is decrease in tendon organ
sensitivity during fatigue (Hutton & Nelson, 1986; Gregory, 1990). The hypothesis that
peripheral receptors are not the primary source of force information during fatigue is
supported by Cafarelli and Layton-Wood's (1986) finding that vibrating a fatigued muscle
has no effect on force sensation. They reported that whereas vibration increased the sensation
of force in fresh (unfatigued) muscles, it had no effect on the perception of forces generated
by fatigued muscles, which continued to be overestimated during vibratory stimulation.

Dissociation between Effort and Force during Fatigue

The results from the studies reviewed above suggest that subjects do not perceive the
actual forces exerted by fatiguing muscles, and that their judgments of force are always based
on the effort associated with generating the contraction. Under some conditions, however,
it has been shown that subjects can distinguish between a sense of effort and a sense offorce.
Roland and Ladegaard-Pedersen (1977) demonstrated that during partial curarization, sub-
jects could dissociate a sense of effort from force if they were instructed to disregard the
increased effort that was required to produce the muscle contraction. Although the subjects'
judgments of the forces generated by the paretic muscles were more variable than under
control conditions, there was not a consistent tendency to overestimate the force as would
be expected if subjects were basing their judgments on the descending motor command. This
result was obtained during isometric force matching and when subjects were matching the
force of compression of two hand-held springs. Despite disruption of the normal relation
between motor signals and muscle output that resulted from the neuromuscular blockade,
these subjects were able to produce equivalent forces in two limbs, a task that must have
involved peripheral afferent feedback.
When the same instructions were given to subjects during sustained constant-force
contractions, Jones (1983) did not find any consistent evidence of a dissociation between
force and effort. On some trials (26%), however, subjects were able to match the force
accurately during the fatiguing contraction, suggesting that it may be possible to dissociate
effort from force during fatigue. Subjects did not perceive that their performance on these
trials was any different from that on trials in which forces were overestimated, and were not
able to describe the cues they used to match forces accurately. The results from one subject
 310 L. A. Jones

l00·~----------------------~

80
...0-~
..
60
~
t
~ 40
101)
.S Figure 3. The matching forces exerted by
.c
~
OJ
20 the left arm as one of three constant forces
~ (35%[.], 50%[.] and 65%[e] MVC) was
maintained by the right arm until maximal
0
endurance. This subject was the most accu-
0 30 60 90 120 150 180 210 rate in judging the amplitude of the sustained
Endurance Time (s) force.

who was able to estimate the forces accurately are shown in Fig. 3. The typical performance
of subjects in this study, as in the other experiments on fatigue, was for forces to be
overestimated. There are data, however, that do suggest that during fatigue of the inspiratory
muscles, most subjects (6/8) can judge accurately the inspiratory pressure developed (i.e.,
muscle tension) as well as the effort associated with producing the contraction (Gandevia et
aI.,1981).
It appears that the sense of effort dominates the perception of muscle force during
fatigue and that it is extremely difficult for subjects to estimate the actual forces being
produced by the muscle, although for some this is possible (Gandevia et aI., 1981; Jones,
1983). With training and feedback of performance, it may be possible to facilitate this
dissociation between effort and force. The preponderance of a sense of effort during fatiguing
contractions may permit a subject to gage the endurance capacity of the muscle and so may
function as a protective mechanism in that the subject can anticipate when muscle failure is
imminent. .

MUSCLE WEAKNESS AND PERCEIVED EFFORT AND FORCE

The changes in the perception of force during fatigue are also evident during other
states or conditions that affect the force-generating capacity of the muscle. Among the
clinical symptoms first described by Holmes (1917) in his studies of patients with unilateral
cerebellar lesions who were without sensory loss were disturbances in the appreciation of
weight (and force). Even though there was no difference between the patients' hands in their
ability to discriminate weights, the majority of these people overestimated the heaviness of
weights lifted on the affected side. The overestimation of weight was attributed to the
asthenia of the limb and a loss of muscle tone. This presumably resulted from a reduction in
the excitability of the motoneuronal pool which then required an increased neural input in
order to achieve any specified level of force. Similar results have been reported in patients
with paresis due to upper motoneuron disease who are without any clinical evidence of
sensory loss. Gandevia and McCloskey (1977a) found that these subjects consistently
overestimated the magnitude of forces generated on the hemiparetic side, and hypothesized
that this was due to the increased neural output of unaffected corticofugal pathways required
to compensate for the pathways no longer functional.
Paresis can be experimentally induced by partially blocking transmission at the
myoneural junction using a neuromuscular blocking agent. This procedure has been used
Sensing Effort and Force during Fatiguing Contractions 311

to study the perception of force and effort during paresis and to examine the time course
of the perceptual changes. When a muscle is weakened by local infusion of a blocking
agent, the forces generated by the muscle are perceived to be greater than those produced
under control conditions, as reflected in the magnitude of the forces generated on the
contralateral (unaffected) side (Gandevia & McCloskey, 1977b; Roland & Ladegaard-
Pedersen, 1977). This overestimation of force during partial curarization is consistent
with the idea that judgments of force are based on the magnitude of the centrally generated
motor command producing the contraction, rather than on the sensory signals arising
from the muscle. With paresis, there is an increase in the efferent signal required to
produce a given level of muscle force (Tang & Rymer, 1981), but the peripheral afferent
discharges from muscle spindle and tendon organ receptors continue to signal the actual
force of contraction.

SUMMARY

The results from numerous experiments indicate that whenever there is a change in
the voluntarily generated motor command producing a muscle contraction, there is a parallel
change in the perceived amplitude of the force produced by the muscle. Although these
changes occur in parallel, there is no directly proportional relation between the two. For
example, during partial curarization, the maximal strength of the muscle may be reduced to
10% of its normal value, but the perceived heaviness of a weight is overestimated by only
40% (Gandevia & McCloskey, I 977b). Similarly, at the point of endurance of a sustained
submaximal contraction, the force produced by the contralateral unfatigued muscle is not
the MVC, but a submaximal force (Jones & Hunter, 1983a). An exception to these findings
is the results of Roland (1978) who found that when the subjective effort of isometric
contractions exerted by a non-paretic and paretic hand were matched, the matching forces
produced could be predicted quite accurately from the maximal strength of the paretic hand.
The absence of a proportional relation between judgments of force and the de-
scending efferent command is not unexpected given the numerous systems involved in
generating and sensing muscle force. A peripheral signal from the muscle must be used
to indicate the success of a muscle contraction (Gandevia & McCloskey, 1978), that is
that the force generated is adequate for the task at hand. A signal of intramuscular force
could also be used to scale a corollary discharge by modulating the sensory effects of
these signals as McCloskey and colleagues (1983) have proposed. Although the evidence
presented here strongly implicates a centrally mediated perception of force, the contri-
bution of intramuscular receptors to force perception cannot be discounted. Unfortunately,
it is not possible to devise an experiment in which tendon organs would be the sole source
of force information. For many ofthe experiments described in this article, assumptions
were made, on the basis of rather limited data, about the behavior of tendon organ receptors
during these perturbations. In some of this work, it was also assumed that they were the
only available signals of peripheral force. There is clearly a need for more detailed studies
of muscle afferent discharge patterns during normal conditions of muscle activation and
as the muscle fatigues.

ACKNOWLEDGMENTS

The author wishes to thank Dr. Simon Gandevia for his helpful comments on the
manuscript. Her laboratory has been supported by the Medical Research Council of Canada
312 L. A. Jones

and the Office of Naval Research (USA). Her attendance at the 1994 Bigland-Ritchie
conference was supported, in part, by the University of Miami.

REFERENCES
Bell C (1826). On the nervous circle which connects the voluntary muscles with the brain. Philosophical
Transactions of the Royal Society of London 116, 163-173.
Bigland-Ritchie B (1981). EMG/force relations and fatigue of human voluntary contractions. Exercise and
Sport Sciences Reviews 9,75-117.
Botte M-C & ScharfB (1980). La sonie - Effets simultanes de fatigue et de masque. Acustica 46,100-106.
Cafarelli E (1982). Peripheral contributions to the perception of effort. Medicine and Science in Sports and
Exercise 14, 382-389.
Cafarelli E (1988). Force sensation in fresh and fatigued human skeletal muscle. Exercise and Sports Sciences
Reviews 16, 139-168.
Cafarelli E & Bigland-Ritchie B (1979). Sensation of muscle force in muscles of different length. Experimental
Neurology 65,511-525.
Cafarelli E & Layton-Wood J (1986). Effect of vibration on force sensation in fatigued muscle. Medicine and
Science in Sports and Exercise 18, 516-521.
Cain WS & Engen T (1969). Olfactory adaptation and the scaling of odor intensity. In: Pfaffman C (ed.),
Olfaction and Taste Ill, pp. 127-141. New York: Rockefeller University Press.
Cain WS & Stevens JC (1971). Effort in sustained and phasic handgrip contractions. American Journal of
Psychology 84, 52-65.
Crago PE, Houk JC & Rymer WZ (1982). Sampling of total muscle force by tendon organs. Journal of
Neurophysiology 47, 1069-1083.
Eisler H (1965). The ceiling of psychophysical power functions. American Journal ofPsychology 78, 506-509.
Enoka RM & Stuart DG (1992). Neurobiology offatigue. Journal ofApplied Physiology 72,1631-1648.
Gandevia SC, Killian KJ & Campbell EJM (1981). The effect of respiratory muscle fatigue on respiratory
sensations. Clinical Science 60, 463-466.
Gandevia SC, Macefield G, Burke D & McKenzie DK (1990). Voluntary activation of human motor axons in
the absence of muscle afferent feedback. Brain 113, 1563-1581.
Gandevia SC & McCloskey DI (1977a). Sensations of heaviness. Brain 100, 345-354.
Gandevia SC & McCloskey DI (l977b). Changes in motor commands as shown by changes in perceived
heaviness, during partial curarization and peripheral anaesthesia in man. Journal of PhYSiology
(London) 272, 673-689.
Gandevia SC & McCloskey DI (1978). Interpretation of perceived motor commands by reference to afferent
signals. Journal of Physiology (London) 283, 493-499.
Geisler WS (1979). Evidence forthe equivalent-background hypothesis in cones. Vision Research 19,799-805.
Gescheider GA (1985). Psychophysics: method, theory, and application (2nd Ed.). Hillsdale, NJ: Erlbaum.
Gregory IE (1990). Relations between identified tendon organs and motor units in the medial gastrocnemius
muscle of the cat. Experimental Brain Research 81, 602-608.
Helmholtz H von (1925). Treatise on Physiological Optics vol. 3, Southall JPC (ed., translator). Menasha WI:
Optical Society of America (original work published 1866).
Holmes G (1917). The symptoms of acute cerebellar injuries due to gunshot injuries. Brain 40, 461-535.
Hutton RS, Kaiya K, Suzuki S & Watanabe S (1987). Post-contraction errors in human force production are
reduced by muscle stretch. Journal ofPhysiology (London) 393, 247-259.
Hutton RS & Nelson DL (1986). Stretch sensitivity of Golgi tendon organs in fatigued gastrocnemius muscle.
Medicine and Science in Sports and Exercise 18, 69-74.
Jami L (1992). Golgi tendon organs in mammalian skeletal muscle: Functional properties and central actions.
Physiological Reviews 72, 623-666.
Jones LA (1983). Role of central and peripheral signals in force sensation during fatigue. Experimental
Neurology 81, 497-503.
Jones LA (1986). Perception of force and weight: Theory and research. Psychological Bulletin 100,29-42.
Jones LA (1989). Matching forces: Constant errors and differential thresholds. Perception 18, 681-687.
Jones LA & Hunter IW (1983a). Perceived force in fatiguing isometric contractions. Perception & Psycho-
physics 33, 369-374.
Jones LA & Hunter IW (1983b). Effect offatigue on force sensation. Experimental Neurology 81,640-650.
Sensing Effort and Force during Fatiguing Contractions 313

Macefield G, Hagbarth K-E, Gorman R, Gandevia SC & Burke D (1991). Decline in spindle support to
!l-motoneurones during sustained voluntary contractions. Journal of Physiology (London) 440,
497-512.
McCloskey DI (1981). Corollary discharges: Motor commands and perception. In: Brookhart JM, Mountcastle
VB (sec. eds.), Brooks VB (vol. ed.), Handbook ofPhysiology. sec. 1. vol. II. pt 2. The Nervous System:
Motor Control., pp. 1415-1448. Bethesda, MD: American Physiological Society.
McCloskey 01, Ebeling P & Goodwin GM (1974). Estimation of weights and tensions and apparent involve-
ment of a sense of effort. Experimental Neurology 42, 220-232.
McCloskey 01, Gandevia SC, Potter EK & Colebatch JG (1983). Motor commands and judgements about
muscular contractions. In: Desmedt IE (ed.), Motor Control in Man. pp. 151-167. New York: Raven
Press.
Pang XD, Tan HZ & Durlach NI (1991). Manual discrimination offorce using active finger motion. Perception
& Psychophysics 49.531-540.
Poulton EC (1979). Models for biases in judging sensory magnitude. Psychological Bulletin 86,777-803.
Proske U & Gregory IE (1980). The discharge rate: tension relation of Golgi tendon organs. Neuroscience
Letters 16, 287-290.
Roland PE (1978). Sensory feedback to the cerebral cortex during voluntary movement in man. Behavioral
and Brain Sciences 1, 129-171.
Roland PE & Ladegaard-Pedersen H (1977). A quantitative analysis of sensations of tensions and of
kinaesthesia in man. Brain 100, 671-692.
Ross HE, Brodie E & Benson A (1984). Mass discrimination during prolonged weightlessness. Science 225,
219-221.
Ross HE & Reschke MF (1982). Mass estimation and discrimination during brief periods of zero gravity.
Perception & Psychophysics 31,429-436.
Stein RB, Hunter IW, Lafontaine SR & Jones LA (1995). Analysis of short latency reflexes in human elbow
flexor muscles. Journal ofNeurophysiology In press
Stevens JC & Cain WS (1970). Effort in isometric muscular contractions related to force level and duration.
Perception & Psychophysics 8, 240-244.
Stevens JC & Cain WS (1972). Effort in isometric contractions: Buildup and recovery. 17th 1nternational
Congress. International Association ofApplied Psychology 1, 399-407.
Tang A & Rymer WZ (1981). Abnormal force-EMG relations in paretic limbs ofhemiparetic human subjects.
Journal ofNeurology. Neurosurgery and Psychiatry 44,690-698.
Teghtsoonian R, Teghtsoonian M & Karlsson JG (1977). The effects of fatigue on the perception of muscular
effort. In: Borg G (ed.), Physical Work and Effort. pp. 157-180. Oxford: Pergamon Press.
Thompson S, Gregory IE & Proske U (1990). Errors in force estimation can be explained by tendon organ
desensitization. Experimental Brain Research 79, 365-372.
 23

TRYPTOPHAN,5-HYDROXYTRYPTAMINE
AND A POSSIBLE EXPLANATION FOR
CENTRAL FATIGUE

E. A. Newsholme 1 and E. Blomstrand2

1 Cellular
Nutrition Research Group, Department of Biochemistry
University of Oxford
South Parks Road, Oxford OXI 3QU, United Kingdom
2 Pripps Research Laboratory
Pripps Bryggeri
Bromma, Sweden

ABSTRACT

In prolonged exercise the plasma level of branched-chain amino acids (BCAA) may fall and
that of fatty acid increases: the latter increases the free tryptophan level, so that the plasma
concentration ratio, free tryptophanIBCAA may increase leading to higher levels of trypto-
phan and therefore of5-hydroxytryptamine (5-HT) in brain. The latter increases the activity
of some 5-HT neurons in the brain which can cause sleep and which could, therefore, increase
the mental effort necessary to maintain athletic activity. Drinks containing branched-chain
amino acids should restore vigor to athletes whose performance is depressed by an excess
of cerebral 5-HT. Recent work suggests that intake of branched-chain amino acids may
improve performance in slower runners in the marathon and decrease perceived physical and
mental exertion in laboratory experiments. This suggestion is supported by pharmacological
manipulations that result in either increased or decreased physical performance.

INTRODUCTION

The mechanisms that underlie fatigue during physical exercise have attracted the
attention of physiologists and biochemists for many years. Many studies have been published
concerning muscle (peripheral) fatigue and several hypotheses have been put forward
including accumulation of protons, depletion of muscle glycogen or failure of neuromuscular
transmission. However, little is known about the biochemical mechanisms of central fatigue
(i.e., fatigue resulting from changes within the central nervous system). Mechanisms that
have been suggested to cause central fatigue, include changes in brain monoamine concen-
trations (Newsholme, 1986) or accumulation of ammonia in the brain during exercise
(Okamura et aI., 1987). Changes in plasma amino acid concentrations could playa role in

315
316 E. A. Newsholme and E. Blomstrand

central fatigue by increasing the rate of synthesis and hence the level of the neurotransmitter
5-HT in parts of the brain (Newsholme, 1986). Thus, it has been considered that 5-HT is
involved in sleep, aggression and mood, so that it was suggested that this neurotransmitter
might also be involved in the mental fatigue that can occur during and after vigorous and
sustained physical exercise.

PLASMA FREE TRYPTOPHAN CONCENTRATION AND FATIGUE

The following is a summary of several important cellular nutritional facts that form
the basis for the hypothesis.
1. Tryptophan is converted in the brain to the neurotransmitter 5-HT.
2. Branched-chain amino acids (leucine, isoleucine and valine) are not taken up by
liver but by muscle and their rate of uptake from the blood by muscle is increased
during exercise.
3. Both branched-chain amino acids and tryptophan enter the brain upon the same
amino acid carrier, so that competition between the two types of amino acids for
entry into brain can occur (Fernstrom, 1990)
4. None of the enzymes involved in the conversion of tryptophan to 5-HT appears
to approach saturation with substrate (i.e., there is no flux-generating step in this
series of reactions) (Fernstrom, 1990; see Newsholme & Leech, 1983 for a
discussion of flux-generating steps).
5. An increased level of tryptophan will, therefore, be expected to increase the rate
of formation of 5-HT and hence increase the level of this neurotransmitter in the
brain. This could result in increased firing of some 5-HT neurons. It is known that
5-HT is involved in sleep so that it may also result in tiredness and possibly
fatigue; that is, it could increase the mental effort necessary to maintain the pace
of running (Newsholme & Leech, 1983; Newsholme, 1986).
6. Tryptophan is unique amongst the amino acids in that it is bound to plasma
albumin, so that it exists as a bound and a free form which are in equilibrium; this
equilibrium is changed in favor offree tryptophan when the plasma fatty acid level
is increased (Blomstrand et ai., 1988).

AN HYPOTHESIS FOR CENTRAL FATIGUE

An increase in the plasma fatty acid level and/or a decrease in that of branched-chain
amino acids could markedly influence the plasma concentration ratio of free tryptophan to
branched-chain amino acids to favor entry of tryptophan into the brain and consequently
increase the level of 5-HT in at least some parts of the brain.
In exercise, either intermittent or continuous, there is elevation in the blood catecho-
lamine levels and a decrease in that of insulin which will result in fatty acid mobilization
from adipose tissue. This increases the plasma fatty acid level. If there is precise control
between the mobilization of fatty acids, the extent of vasodilation in muscle, and the
stimulation of fatty acid oxidation in muscle, the increased rate of fatty acid oxidation by
muscle may occur with little increase in the plasma level of fatty acids. Hence the free
tryptophan level may not change much. However, if this coordination is poor, due to lack of
training or due to hypoglycemia, the blood fatty acid concentration could be increased to
sufficiently high levels to increase the plasma concentration offree tryptophan. Furthermore,
in intermittent exercise in which there is usually a greater dependence upon anaerobic
Tryptophan, S-Hydroxytryptamine and Central Fatigue 317

exercise and, therefore, less opportunity to oxidize these fatty acids, the plasma level of fatty
acid could rise and hence increase the concentration of free tryptophan.
In prolonged exercise it is likely that muscle will use branched-chain amino acids for
energy when the muscle glycogen level is depleted. Similarly, when the liver glycogen store
is depleted, fatty acid mobilization may increase raising the blood level of fatty acids, so that
the plasma level of free tryptophan will increase. Thus, an interesting possibility is fatigue
that is caused by the depletion of glycogen in muscle could be due not to a direct effect of
glycogen depletion on the muscle but to an increase in the level of 5-HT in a specific area
of the brain. Failure of the motor centers in the brain to stimulate muscle to contract would
mean that the power output would fall, i.e. fatigue due to a change in the balance of
concentration of key amino acids in the blood but initiated by depletion ofliver and muscle
glycogen stores.

Testing of the Hypothesis

Recent experiments have provided evidence in support of the hypothesis. A summary
of the experimental findings supporting the hypothesis is as follows: 1) the plasma concen-
tration ratio, free tryptophanlbranched-chain amino acids is increased in man and in the rat
after prolonged and exhaustive exercise (Blomstrand et ai., 1988; 1989); 2) in the rats, the
levels of tryptophan were increased in all areas of the brain studied: in contrast the level of
5-HT was increased in only two areas, the brain stem and the hypothalamus: the level of
5-hydroxyindole acetic acid was also increased in these areas (Blomstrand et ai., 1989); 3)
such changes in neurotransmitter levels even in behaviorally significant areas of the brain
such as the brain stem and hypothalamus, provides only indirect evidence in support. How
is it possible to provide more direct evidence? Two approaches have been tried. First, athletes
during various events have taken a drink containing branched-chain amino acids sufficient
to maintain the resting free tryptophanlbranched-chain amino acids concentration ratio. The
effects on physical and mental fatigue have then been studied. Secondly, the effect of
pharmacological manipulation of the 5-HT neurotransmitter system on physical performance
have been studied in both rats and man.
A summary of the results of key experiments is as follows:
1. In a Stockholm marathon, 193 volunteers were randomly divided into two groups
and the experiment was performed double-blind. In the experimental group, a
drink containing a mixture of branched-chain amino acids was given four times
during the race and the other group was given a placebo drink. Considering the
whole group of subj ects, the difference in performance between the experimental
and placebo groups was small and did not reach significant levels. However, when
the subjects were divided into two subgroups based on their marathon time, the
performance was statistically significantly better for the slower runners (3.05-3.30
h to complete this marathon) in the experimental group compared with the placebo
group. The difference in time at this pace would mean an improvement in
performance of 5-6 min (Blomstrand et ai. 1991 a). The reason for the lack of effect
on performance in the faster runners «3.05 h to complete this marathon) is not
known. It might be that the more well-trained (faster) runners are more resistant
to fatigue, both central and peripheral fatigue, and, therefore, less sensitive to a
supply of branched-chain amino acids. Alternatively, the slower runners may have
depleted their glycogen stores at an earlier stage of the race so that they would
increase the plasma level of fatty acids and decrease that of branched-chain amino
acids at an earlier time in the race compared with the better runners. In this case,
they would change their free tryptophanlbranched-chain amino acid ratio earlier
318 E. A. Newsholme and E. Blomstrand

in the race and this would cause fatigue according to the hypothesis. For this
reason, the effect of the amino acid supplementation would be easier to detect in
these slower runners.
2. A preliminary experiment was carried out in eight runners who completed two 24
km cross-country runs separated in time by one week. After each race the runners
were asked to recall their perceived physical and mental effort. On the first
occasion four subjects drank the placebo drink before the run and after 13 km,
and four subjects drank the branched-chain amino acids at the same time periods.
The placebo and the branched-chain amino acid were indistinguishable from one
another and the experiment was performed double-blind. Drinks were then
switched and the procedure repeated one week later. There was no difference in
running time between the two occasions. However, both the perceived physical
and mental efforts, as measured on a modified Borg scale, were lower when
branched-chain amino acids were taken, but the difference was only statistically
significant for the perceived mental effort (Newsholme et aI., 1991).
3. Twenty-four participants were questioned after a marathon on their mental attitude
towards the race. All of the 12 participants who drank the placebo drink experi-
enced aversion to running the race during the last 10 km; in contrast, of the 12
participants who drank the branched-chain amino acids, only seven experienced
aversion as indicated from a questionnaire. The difference was statistically sig-
nificant (Newsholme et aI., 1991).
4. The Stroop color and word test (CWT) was given to 16 subjects who participated
in a 30 km cross-country race (Blomstrand et aI., 1991 b). Research on the CWT
has established that the test provides a useful tool in the study ofneuropsychologi-
cal and cognitive processes. In the group who took the branched-chain amino acids
during the race, performance in this test improved after the race in comparison
with before the race, while no statistically significant difference was found for the
subjects who took the placebo drink.
It is possible to change the effectiveness of neurotransmitters in a synapse in the brain
by pharmacological manipulations: agents that can prevent the binding of the neurotrans-
mitter to the post-synaptic receptor and elicit no response from the receptor are known as
antagonists: those that bind and do elicit a response are known as agonists. In addition, some
agents can prevent the uptake of the neurotransmitter from the synaptic cleft so reducing its
effect. All these approaches have been used in testing the 5-HT and central fatigue hypothe-
sis: 1) administration of a 5-HT agonist to rats impairs running performance in a dose-related
manner (Bailey et aI., 1992); 2) in contrast, administration of a 5-HT antagonist to the rats
improved running performance (Bailey et aI., 1993); 3) administration of a 5-HT re-uptake
blocker to human subjects lowered physical performance; exercise time to exhaustion during
standardized exercise was decreased in comparison to a control condition (Wilson &
Maughan, 1992).
Although no signs of circulatory effects of the drug were observed, it cannot be ruled
out that these drugs, especially at high does, might have effects on the peripheral metabolism
or central circulation and such changes could influence performance.

EFFECT OF INGESTION OF BRANCHED-CHAIN AMINO ACIDS

AND/OR CARBOHYDRATE ON FATIGUE

It has been suggested that an intake of BCAA might be detrimental to physical

performance due to an increased production of ammonia originating from the metabolism
Tryptophan, 5-Hydroxytryptamine and Central Fatigue 319

of BCAA during exercise. However, when small amounts of BCAA are supplied but
sufficient to maintain the plasma concentration ratio of free tryptophanIBCAA (6-8 g of
BCAA during exercise) at resting levels, no increase in the plasma concentration of ammonia
has been found as a consequence of BCAA supplementation, that is, exercise caused the
same increase in plasma ammonia concentration when BCAA or water were supplied
(Blomstrand & Newsholme unpublished observations).
When BCAA were supplied together with carbohydrates during exercise, no
difference in the perceived mental and overall fatigue could be detected during 80 min
of exercise at 75% of the maximal oxygen uptake when compared with provision of
carbohydrate alone. Provision of carbohydrate before or during exercise may prevent or
decrease the increase in plasma fatty acid concentration that can occur during exercise.
This may be caused by a stimulation of insulin secretion, or, more interestingly, by
increasing the rate of esterification of fatty acids in the liver. Whatever the mechanism,
a decrease in the plasma concentration of fatty acids would cause a decrease in the free
tryptophan level in the plasma. Hence, carbohydrate intake may be effective in delaying
fatigue not only by preventing a fall in the blood glucose level but also by decreasing
or preventing the increase in the free tryptophan concentrations in the bloodstream. In a
recent study (Davis et aI., 1992) it has been reported that provision of carbohydrate
attenuates, in a concentration-related manner, the increase in the plasma free tryptophan
concentration during sustained exercise to fatigue (2-4 h duration). However, when the
exercise continues for longer than 2-3 h there may be an increase in the plasma concen-
tration of fatty acids and hence that of free tryptophan even when carbohydrate is
consumed (Wright et aI., 1991; Davis et aI., 1992). This might explain why it was possible
to detect an effect of BCAA supplementation on the marathon performance even when
carbohydrate was provided during the race (Blomstrand et aI., 1991a), but not in shorter
events. A competitive marathon is extremely heavy exercise and, in the experiment
reported above, almost 200 volunteers were involved in the study which made it possible
to detect a small difference in performance (3%).

Training and the 5-HT-System

Recently some evidence has been presented that endurance-training alters the
sensitivity of the 5-HT system. The sensitivity to a neurotransmitter can be modified
by changing the number of postsynapic receptors. Since the release of prolactin from
the pituitary is controlled, in part, by the activity of the 5-HT-system in the hypo-
thalamus, prolactin release can be used as test of 5-HT sensitivity (Cowen et aI.,
1990). This sensitivity is lower in endurance-trained athletes compared with untrained
individuals (Jakeman et aI., 1994). This provides further support for the theory that
the 5-HT system is involved in central fatigue during endurance exercise and indicates,
furthermore, that central fatigue, like peripheral fatigue, is affected by physical training.
Thus, effects of BCAA supplementation on physical and mental fatigue may depend
not only on the changes in the plasma free tryptophan concentrations but also on the
sensitivity of the 5-HT system to changes in the level of 5-HT in neurons in specific
parts of the brain.

ACKNOWLEDGMENTS

Support from Pripps Bryggeria (Stockholm, Sweden) is gratefully acknowledged.

320 E. A. Newsholme and E. Blomstrand

REFERENCES
Bailey SP, Davis IM & Ahlborn EA (1992). Effect of increased brain serotonergic activity on endurance
performance in the rat. Acta Physiologica Scandinavica 145, 75-76.
Bailey SP, Davis IM & Ahlborn EA (1993). Neuroendocrine and substrate responses to altered brain 5-HT
activity during prolonged exercise to fatigue. Journal ofApplied Physiology 74, 3006-3012.
Blomstrand E, Celsing F & Newsholme EA (1988). Changes in plasma concentrations of aromatic and
branched-chain amino acids during sustained exercise in man and their possible role in fatigue. Acta
Physiologica Scandinavica 133, 115-121.
Blomstrand E, Hassmen P, Ekblom B & Newsholme EA (1991a). Administration of branched-chain amino
acids during sustained exercise--effects on performance and on plasma concentration of some amino
acids. European Journal ofApplied Physiology 63, 83-88.
Blomstrand E, Hassmen P & Newsholme EA (1991b). Effect of branched-chain amino acid supplementation
on mental performance. Acta Physiologica Scandinavica 143, 225-226.
Blomstrand E, Perrett D, Parry-Billings M & Newsholme EA (1989). Effect of sustained exercise on plasma
amino acid concentrations and on 5-hydroxytryptamine metabolism in six different brain regions of
the rat. Acta Physiologica Scandinavica 136, 473-481.
Cowen PJ, Anderson 1M & Grahame-Smith DG (1990). Neuroendocrine effects of azapirones. Journal of
Clinical Psychopharmacology 10, 215-255.
Davis IM, Bailey SP, Woods JA, Galiano FJ, Hamilton MT & Bartoli WP (1992). Effects of carbohydrate
feedings on plasma free tryptophan and branched-chain amino acids during prolonged cycling.
European Journal ofApplied Physiology 65,513-519
Fernstrom JD (1990). Aromatic amino acids and monamine synthesis in the CNS: influence of diet. Journal
ofNutritional Biochemistry 10, 508-517.
Jakeman PM, Hawthorne JE, Maxwell SRI, Kendall MJ & Holder G (1994). Evidence for down regulation of
hypothalamic 5-hydroxytryptamine receptor function in endurance-trained athletes. Experimental
Physiology 79, 461-464.
Newsholme EA (1986). Application of knowledge of metabolic integration to the problem of metabolic
limitations in middle distance and marathon running. Acta Physiologica Scandinavica 128 (suppl.
556),93-97.
Newsholme EA, Blornstrand E, Hassmen P & Ekblom B (1991). Physical and mental fatigue: do changes in
plasma amino acids playa role? Biochemical Society Transactions 19, 358-362
Newsholme EA, & Leech AR (1983). Biochemistry for the Medical Sciences. Chichester, England: John Wiley
& Sons.
Okamura K, Matsubara F, Yoshioka Y, Kikuchi N, Kikuchi Y & Kohri H (1987). Exercise-induced changes in
branched chain amino acid/aromatic amino acid ratio in the rat brain and plasma. Japanese Journal
ofPharmacology 45, 243-248.
Wilson WM & Maughan RJ (1992). Evidence for a possible role of 5-hydroxytryptamine in the genesis of
fatigue in man: administration of paroxetine, a 5-HT re-uptake inhibitor, reduces the capacity to
perform prolonged exercise. Experimental Physiology 77, 921-924.
Wright DA, Sherman WM & Dernbach AR (1991). Carbohydrate feedings before, during or in combination
improve cycling endurance performance. Journal ofApplied Physiology 71, 1082-1088.
SECTION VII
Task Dependency of Fatigue Mechanisms

This section specifically addresses how fatigue varies during different tasks. Histori-
cally, the realization of the different tasks performed by red and white muscles was one of
the forerunners of this discussion. The term task dependency reminds us that the pattern and
recruitment of motor units, factors influencing blood flow, and variations in the state of the
segmental and supraspinal apparatus will ultimately influence muscle force and the duration
for which it can be sustained. Consequently, the duration that a task can be sustained will
vary as the details of the task vary.
In Chapter 24, Sargeant and Jones review the relationship between fatigability and
the speed of contraction. They assess how the known properties of human single muscle
fibers would affect maximal power output. Analysis of human cycling is used to explore the
factors that alter the relationship between velocity (pedaling rate) and maximal peak power.
These factors include the proportions of the different fiber types, the extent to which they
are recruited, and muscle temperature. Their modeling relies on results from experiments on
single muscle fibers but, most valuably, exposes gaps in our knowledge about the neural
control of muscle during concentric and eccentric exercise.
When a muscle contracts the pressure within it increases to levels 2-3 fold greater
than the arterial blood pressure. Elevated intramuscular pressures have three important
implications for muscle fatigue. First, the increased pressure will impede the blood flow and
thus impose relatively ischemic conditions; second, it may interfere mechanically with the
transmission of force into the tendon. Third, the elevated pressure will alter the forces that
govern the transfer of fluids across the capillary wall. Sejersted and Hargens (Chapter 25)
analyze the precise mechanism for the increase in intramuscular pressure and its implications
for muscle fatigue under different circumstances, ranging from isometric contractions to
walking. Interestingly, under some conditions the intramuscular pressure is highly correlated
with intramuscular force and thus the effective neural drive to the muscle. Their analysis
emphasizes the need to consider the biomechanical effects of intramuscular architecture on
pressure and force generation: the so-called clinical compartment syndromes reinforce this
View.
In Chapter 26, Botterman begins, like many others, with the assumption that the
neuromuscular system has evolved to minimize fatigue. After reviewing the differences in
fatigability among type-identified motor units in the cat, he then describes how the discharge
rate of a motor unit must change if the desired outcome is to hold constant (or clamp) the
force produced by a single motor unit. Such a paradigm is one that occurs regularly under
natural conditions, of course. Not surprisingly, type S units can produce a constant submaxi-
mal force with a constant discharge frequency for extended periods of time. The concurrent
322 Task Dependency of Fatigue Mechanisms

expression of potentiation and fatigue (including contractile slowing), however, adds a

complication to how the eNS might control all motor units to produce a constant force. This
chapter also brings out the value of using a variety offatigue-inducing stimulation paradigms
to bring out different features of motor unit fatigue and recovery.
Bigland-Ritchie and colleagues attempt a global view of fatigue in Chapter 27.
Fatigue is defined as "any reduction in a person's ability to exert force or power in response
to voluntary effort, regardless of whether or not the task itself can be performed success-
fully". Rather than accepting that fatigue will occur because of a failure at one site in a chain
from the motor cortex to the actomyosin crossbridge, they hypothesize that fatigue will
depend on the amount ofstress placed at each site. A corollary they suggest is that impairment
at one site is balanced by an improvement at another. They indicate that there is no simple
relation between muscle metabolism and muscle fatigue, and that impairment of excitation-
contraction coupling is a major reason why the relation is not simple. Overall, because the
neuromuscular system has evolved to operate in humans under a wide variety of conditions,
the authors argue that fatigue, at least conceptually, is a unitary process.
The chapters in this section reinforce the view that the mechanisms responsible for
fatigue depend on the exact task performed by the neuromuscular system.
 24

THE SIGNIFICANCE OF MOTOR UNIT

VARIABILITY IN SUSTAINING
MECHANICAL OUTPUT OF MUSCLE

A. J. Sargeant' and D. A. Jones 2

'Department of Muscle and Exercise Physiology, Faculty of Human

Movement Sciences
Vrije University
Van der Boechorstraat 9, 1081 BT Amsterdam, The Netherlands
2 School of Sport and Exercise Sciences
The University of Birmingham
Edgbaston, Birmingham Bl5 2TT, UK

ABSTRACT

Neuromuscular function and fatigue have been studied using a wide variety of preparations.
These range from sections of single fibers from which the cell membrane has been removed
to whole muscles or groups of muscles acting about a joint in the intact animal. Each type
of preparation has its merits and limitations. There is no ideal preparation; rather the question
to be answered will determine the most appropriate model in each case and sometimes a
combination of approaches will be needed. In particular, it is important to understand how
the mechanical output of whole muscle can be sustained to meet the demands of a task and
to take into account the organized variability of the constituent motor units.

INTRODUCTION

In a mature muscle one motoneuron will, through its axonal branches, supply a
number of muscle fibers throughout the muscle. In a healthy muscle the innervation is such
that adjacent fibers will most probably be supplied by branches from different motoneurons.
A motoneuron and all of the muscle fibers that it innervates forms a motor unit. As a
consequence of a common pattern of activation, muscle fibers within a single motor unit will
be relatively homogenous in terms of contractile and metabolic properties. There may,
however, be considerable variability between motor units within a single muscle, although
the degree of variability will depend on the range of mechanical output required from a
muscle in a particular species. Not only will muscle fiber properties be influenced by the
pattern of activation but also by the environment in which they operate. Thus some limited
variability can be expected even within a single motor unit as a consequence of the

323
324 A. J. Sargeant and D. A. Jones

geographical scattering of the component fibers. This scattering may, for example, lead to
marked differences in operating temperature and intra-muscular pressure depending on
whether a fiber is located superficially or deep within the muscle (Sargeant, 1994a). In
addition, changes in interstitial concentrations of potassium. for example, consequent upon
activation of some fibers, may influence the excitability of other previously quiescent fibers.
These changes may be important in influencing mechanical output during fatiguing submaxi-
mal contractions (for discussion, see McComas et al., 1993; see Sj0gaard & McComas,
Chapter 4).

VARIABILITY OF MUSCLE FIBERS

As early as the seventeenth century, a number of authors had commented on the fact
that muscles differ in their appearance. However, it was not until 1873 that Ranvier
recognized that skeletal muscles not only differ in color, but that they also had different
contractile properties. For example, in most mammals the soleus muscle is red in appearance
and slow to contract while the extensor digitorum longus is white and contracts and relaxes
rapidly. Subsequently, various combinations of histochemical stains have been used to
identify different muscle fiber types. The most commonly used stains are: l) myosin ATPase
with preincubation at either alkaline or acid pH; 2) NADH tetrazolium reductase or succinic
dehydrogenase, a marker of mitochondrial activity; and 3) myophosphorylase, a marker of
glycolytic activity. Using this histochemical approach, three main fiber types can be
distinguished. However, the histochemical techniques are, to a greater or lesser extent,
manipulated to create apparently discrete fiber types, whereas in reality there is a continuum
in most if not all contractile and metabolic properties (Figs. lA and 2).
Although the histochemically based classifications may be adequate for many
purposes, it is now possible to show that there are a complex set of genes controlling the
expression of the contractile proteins which determine the intrinsic contractile properties. In
human muscle fibers, for example, it has been shown that many type II muscle fibers
co-express variable proportions of types A and B rather than expressing a single isoform of
the myosin heavy chain. The continuum in the degree of co-expression is also reflected by:
1) the staining intensity for myosin ATPase with pre-incubation at pH 4.6 (San!' Ana et al.,
1995); 2) by the maximal rate of myofibrillar ATPase (Sant' Ana Pereira & van der Laarse,
personal communication); and 3) the maximal velocity of shortening (Fig. 2; Larsson &
Moss, 1993).

~
)(
Q)
100
.t. itt.t.
-HII-/tf/+At>t.t.
+ ___ 0+
tl
I.
I.'
;;100
~
)(
Q)
r~s
FR

'C + t. A
'C
.: B
s:::
50 .... : 50 ,FOnt}
+ +,+

e
Q)
+ Q)
::J ::J
CI
~~+~ CI

10 ~+: 10 FF
u.. o 10 ' !
!

30
!

50
u.. 0
10 20 :II £0
Twitch CT (ms) Twitch CT (ms)

Figure 1. Co-variation between different contractile properties for muscle units of cat's peroneus longus
muscle. Fatigue index (%) plotted vs. twitch time-to-peak (ms) for 80 units (A) and for average values of the
same units after categorization into types S, FR, F(int) and FF (B). In B, the diameter of each symbol is
proportional to the maximal tetanic force of the respective unit category (mean force for FF units: 27.9 g)
(reproduced from Kernell et aI., 1983 (A), by permission of Springer-Verlag; and Kernell, 1986 (B».
Motor Unit Variability and Sustaining Mechanical Output of Muscle 325

20
Type I
15
10
5
0
20
15 Type IIA

10
5
~
.a 0
E
;:, 20
z Type liB
15
10
5
0
20
Figure 2. Distributions of Vmax values
from human quadriceps and soleus muscle 15 Type IINB
fibers classified according to myosin
heavy chain type, i.e., type I, IIA, IIB or 10

--
IIA/B MHCs. Shortening velocities from 5

- t
two fibers co-expressing types I and IIA
MHCs (0.3 and 1.0 MLls) are not shown 01 • II
in the figure (reproduced with permission 0 2 3 4
from Larsson & Moss, 1993).

The Relationship between Contractile Properties and Histochemistry

Experiments in which individual motor units are stimulated enables characterization
of their contractile properties. Subsequent repetitive stimulation to deplete the constituent
muscle fibers of glycogen then allows, in serial sections of the muscle, identification and
histochemical characterization of the muscle fibers of that same motor unit (Burke et ai.,
1973; Kugelberg, 1973). On the basis of size, speed, and fatigability, motor units fall between
two extremes: large, fast and fatigable or small, slow and fatigue resistant. Histochemically,
the large units tend to be made up of type lIB fibers while the small units are predominantly
composed of type I fibers. Type IIA motor units span a range of size and fatigue resistance
which is reflected in the broad spectrum of their mitochondrial enzyme activities. While
mammalian species show similar varieties of fibers when classified by histochemical
techniques, there may be considerable differences in the contractile characteristics of an
identified fiber type between species, between muscles in the same species, and possibly
even within the same muscle. It should also be noted that compared with other species, there
is a much less striking covariation of contractile properties and a smaller endurance range
in at least the human thenar muscles (Thomas et aI., 1991; see Thomas, Chapter 10).
However, it is unclear whether this finding can be extrapolated to larger human limb muscles,
to other exercise protocols or even during voluntary contractions. Furthermore, while three
326 A. J. Sargeant and D. A. Jones

fast myosin heavy chain isoforms have been identified in rat and rabbit muscle (IIA, IIX,
and lIB), only two major fast isoforms (designated IIA and lIB) have been found in human
muscle. Using gene cloning techniques, it has recently been shown that human type lIB is
highly homologous to the rat IIX rather than rat lIB (Ennion et aI., 1995). This finding has
functional significance since the type lIB muscle fibers in rodents have been shown to have
the highest velocities of shortening (Bottinelli et aI., 1991). It is possible therefore that the
type lIB myosin heavy chain isoform is not expressed in humans because it is associated
with too high an intrinsic velocity for larger animals (see the discussion of Hill, 1950, with
respect to the constraints of animal size and the number of sarcomeres in series, on the
intrinsic speed of shortening). Thus, although general principles may be similar across
species there is a need for some caution in extrapolating from animal models to humans since
there may not be an exact correspondence or covariance of properties.

The Control of Fiber Type Expression - Usage Dependent Plasticity

Contractile activity imposed on the muscle unit by the motoneuron is an important
factor determining the contractile and metabolic properties of the constituent muscle fibers.
This was first demonstrated in the classic cross-innervation studies of Buller and colleagues
(1960) and in subsequent chronic stimulation studies (e.g., Salmons and Vrbova, 1969).
These observations have led to the widely held belief that activity is the primary determinant
of motor unit properties. Recent observations, on the effects of inactivity and re-innervation
of adult muscle, indicate, however, that factors independent of activity may also play a
significant role (e.g., Pierotti et aI., 1991; Unguez et aI., 1993).
The major changes seen after chronic stimulation include an increase in capillary
density, proliferation of mitochondria, decrease in sarcoplasmic reticulum and the expression
of different troponin and myosin isoforms. These changes also occur with different time
courses (Pette & Vrbova, 1992). Increased capillary density and mitochondrial content,
which are amongst the first changes to be seen in response to prolonged activity, will
predominantly affect the fatigability of the muscle. Changes in the sarcoplasmic reticulum
and contractile proteins, which require more activity, influence the speed of the muscle.
Although there has been a great deal of work on stimulated animal muscle there are relatively
few investigations of the effects of chronic stimulation on human muscle. The studies that
have been undertaken show that changes in fatigability can be produced but probably without
altering the myosin isoform expression (Rutherford & Jones, 1988). Similarly while training
programs based on voluntary activation increase fatigue resistance of human muscle, it is
difficult to demonstrate changes in myosin expression. Perhaps in both cases there was not
a significant enough departure from basic postural and locomotory activities to reach the
threshold needed to change the myosin isoform expression. Furthermore, the normal hierar-
chical pattern of recruitment used in training exercise will mitigate against changing the
myosin of high threshold motor units. Nevertheless, if training is extremely intense, it may
be possible to effect changes in the contractile properties (e.g., Vmax' as shown by Fitts et
aI., 1989, using human skinned single fibers from the deltoid muscle of competitive
swimmers).

FORCE MODULATION BY MOTOR UNIT RECRUITMENT AND

RATE CODING

In order to meet the demand for steady and finely gradable force output, motor units
are organized in a hierarchy according to size and contractile properties. There is in addition
Motor Unit Variability and Sustaining Mechanical Output of Muscle 327

100 100

E
~ 75 75 E
e
.j(
."
E
;g
011
~ 50
:l
E
.~
~ 25
Figure 3. Proportion of maximal force utilized,
and muscle fibers active, in relation to exercise
intensity. Values are given for the total and com-
ponent fiber type populations; type I, 0; type IIA
~; IIAB & lIB combined 0 (reproduced with 25 50 75 100
permission from Greig et ai., 1985). Exercise intensity (% Va,max)

to this a task related recruitment strategy. Thus at low levels of required force, small (low
force) slow motor units with thin axons and small, easily excitable motoneurons are
recruited. As the requirement for force increases, there is systematic recruitment of larger
(high force) motor units so that the force steps caused by successive recruitment remains
rather constant relative to the prevailing force output. (Kernell, 1992, for review; see Kernell,
Chapter 9). Not surprisingly, and as a consequence of the patterns of activation implicit in
such a recruitment hierarchy there is, in general, an association with increasing size and
increasing fatigue sensitivity and power output (Fig. 1B).
As well as varying the number of motor units recruited, muscle force can also be
modulated by changing the frequency of stimulation, so called rate-coding. Modulation by
frequency changes in the steep part of the force frequency curve provide an additional very
sensitive force control for the neuromuscular system. The extent to which the demand for
force in human muscle contractions is met by the recruitment of more motor units or by
rate-coding is not clear and may vary somewhat between, for example, the small hand
muscles compared to large limb muscles (Kukulka & Clamann, 1981). In the latter context,
there is evidence that in human whole body dynamic exercise such as cycling, there is, in
general, an orderly intensity dependent hierarchical recruitment of motor units such that type
I fibers are recruited first followed by type IIA then type lIB (Gollnick et aI., 1974; Vellestad
et aI., 1984; Vellestad & Blom, 1985). Nevertheless, there is also strong evidence for a large
concurrent element of rate-coding as evidenced by metabolic activity in type II human fibers
at low exercise intensities (Ivy et aI., 1987). Similarly in cycling exercise at intensities close
to V0 2max , which require only -50% of the maximal muscle force (measured at the same
velocity), there is evidence that virtually the whole muscle fiber popUlation has been, to some
degree, metabolically active (Fig. 3; Greig et aI., 1985). This observation implies a significant
element of rate-coding with presumably the more powerful, but fatigable, motor units at the
top of the hierarchy being relatively less activated than the less powerful fatigue resistant
motor units which are lower in the recruitment order. Such an approximate matching of
fatigue resistance and degree of activation could be a useful strategy for sustaining power
output at the highest possible level during prolonged activity. Nevertheless, it does seem that
fatigue of whole muscle power output during high intensity submaximal exercise can be the
result of a mismatch of activation level and fatigue sensitivity leading to a selective fatigue
of fast fatigue sensitive motor units even though they may be activated at relatively low
firing frequencies (Beelen & Sargeant, 1991; Beelen & Sargeant, 1993; Beelen et aI., 1993).
328 A. J. Sargeant and D. A. Jones

TYPES OF CONTRACTION

The mechanical output required from skeletal muscle in-vivo may require one, or a
combination of, three types of contraction. These include: I) isometric contraction, 2)
eccentric contraction; and/or 3) concentric contractions as described below.

Isometric Contraction
This term implies that muscle is activated but is held at a constant length. In fact in
whole body human and other animal studies, it is usually defined operationally as an
activation of muscle(s) when the joint(s) spanned by the muscle(s) are held in a fixed
position. The important point here is that there will always be an element of shortening of
the muscle due, amongst other things, to stretching of the tendinous connections and
compression of joints. For example, in metabolic experiments involving short duration
repetitive isometric contractions (that is, with a fixed joint position) the whole activation
period may be taken up with shortening of the muscle fibers. In addition to this situation,
there may be rather complex and simultaneous lengthening and shortening of sarcomeres in
series along the length of the individual muscle fibers, especially when fatigue is occurring
due to heterogeneity of metabolic and mechanical properties between sarcomeres. Notwith-
standing these qualifications, the generation of isometric force is for largely technical
reasons, the most commonly studied form of contraction. It is also of considerable impor-
tance, for stabilizing joint complexes during locomotion and maintaining posture.

Eccentric Contractions
In these contractions, the muscle is activated but instead of shortening it is lengthened
due to an external force which exceeds that generated by the degree of activation. Eccentric
contractions are a normal and common element of everyday human muscle function. The
role of different fiber types in the generation of eccentric force is not well understood
although a rather small number of fibers may be activated to generate eccentric as compared
to concentric force (Bigland-Ritchie & Woods, 1976) and some investigators have suggested
that there may be a selective recruitment of type II fibers during such contractions (Nardone
et aI., 1989). In relation to activity involving eccentric contractions, it is worth noting that
single fiber studies indicate that levels of fatigue that reduce the maximal velocity of
shortening and modify the (concentric) force velocity relationship may not affect eccentric
force generation to the same extent (Curtin, 1990). This maybe of significance in movement
patterns involving a pre-stretch. Here the preservation of high force generation in the initial
eccentric phase will maintain the amount of elastic energy stored. During the subsequent
power generating concentric phase, the contribution of the recovered elastic energy could
offset the effect of fatigue when compared to a pure concentric contraction (de Haan et aI.,
1991 ).

Concentric Contractions
These occur when the muscle is activated and shortens, power is generated and work
is done by the muscle. Human muscles have been studied much less often under dynamic
than isometric conditions, and often on the implicit assumption that isometric function will
systematically reflect the functional status ofthe muscle during concentric contractions. This
is not always the case as shown, for example, by the independent effects of fatigue on
maximal isometric force and maximal velocity of shortening (de Haan et aI., 1989; Beelen
Motor Unit Variability and Sustaining Mechanical Output of Muscle 329

-;
f
-'
300

...!
~

a:
w 200
o==
0.

~ 100
:E
)(
c(
:E

50 100 150 200

VELOCITY (pedalling rate • rev/min)

Figure 4. Relationship of maximal peak power to pedaling rate. Power is standardized for upper leg muscle
volume. Data for 5 subjects (different symbols), with 50% type II fibers (-). For comparison, data are given
for an ultra-marathon runner with only 4% type II fibers (- - -) (group data from Sargeant et aI., 1981) (reprinted
with permission from Sargeant, 1994b).

& Sargeant, 1991). Clearly, in human locomotion the capability of human muscles to
generate power in concentric contractions is of fundamental importance. As pointed out
previously, the measurement oftrue maximal human power is dependent upon ensuring that
the muscles will be contracting at optimal velocity for maximal power production as defined
by the force velocity relationships. The fact that in whole body exercise this velocity can
only be determined globally for all the contributing muscles does not detract from the
importance of this requirement (Sargeant & Beelen, 1993). In our own studies, we have
addressed the problem by controlling the pedaling frequency in maximal cycling exercise
using an isokinetic system (see e.g., Sargeant et aI., 1981; Beelen et aI., 1994). These studies
have indicated that maximal power in this cyclic leg extension exercise is achieved at an
optimal pedaling rate of about 120 crank rev/min (Fig. 4).

Stretch-Shortening Cycle
As well as these pure forms of contraction, there are many situations where a
combination of contraction types occurs. The most important of these is the stretch-shorten-
ing cycle where an eccentric muscle contraction immediately precedes a concentric power
generating phase. For example in running, the leg extensor muscle-tendon complex can act
not simply as a brake or shock absorber on landing but also as an energy store. On landing,
activation of the leg extensor muscles results in their series-elastic tendinous structures being
stretched. Some of this stored elastic energy may then be recovered and used to add to the
power generated in the subsequent push-off (see e.g., Cavagna et aI., 1968; Alexander, 1984;
Komi, 1984). The re-use of stored elastic energy in this way is common throughout the
animal kingdom, but perhaps the most well known development of the principle in mammals
330 A. J. Sargeant and D. A. Jones

is seen in the kangaroo. Less expected is the contribution that re-use of stored elastic energy
makes to the achievement of high sustained running speeds in the camel (Saltin & Rose,
1994).

WHAT IS THE POTENTIAL CONTRIBUTION OF DIFFERENT

HUMAN MUSCLE FIBER TYPES TO MAXIMAL POWER OUTPUT
A T DIFFERENT CONTRACTION VELOCITIES?

This is an important but unanswered question in relation to dynamic muscle function

in humans. Faulkner and colleagues (1980) measured the force-velocity characteristics of
bundles of human muscle fibers obtained at surgery. Taking account of the proportion of
type I and II fibers present, they suggested a ratio for the V max of around 1:4. A similar ratio
is indicated for the normal control subjects studied by Fitts and colleagues (1989) and this
ratio seems theoretically reasonable on dimensional grounds relative to other species (see
e.g., Hill, 1950, 1956; Rome et al., 1990). However, the recent investigation of Larsson and
Moss (1993) reports a wider range, although the mean value for type II fibers appears to be
of the same order. What is important to note from this last study is the wide range of V max
within the type II population and the large number of hybrid fibers co-expressing variable
proportions of myosin heavy chain isoforms (see also Sant' Ana Pereira et al., 1995).
The implications for power output from human muscles that have fiber populations
with different force-velocity characteristics can be shown by reference to a model. For
illustrative purposes the following assumptions and simplifications are made:
1. There are two discrete populations of muscle fibers (type I and II) with a 1:4 ratio
for their maximal velocities of shortening.
2. The specific force of these two fiber popUlations is such that each generates 50%
of the whole muscle maximal isometric force (after account is taken of their
respective cross-sectional areas).
3. The length-tension relationships, relative to the whole muscle, are the same for
both fiber populations.
4. The alP 0 constant which defines the curvature of the force-velocity relationship
is the same for both populations.
Of these 1) is a simplification based on mean values to make the model manageable
and comprehensible. In fact, there is wide range of V max values, especially for type II fibers
(Fig. 2; Larsson & Moss, 1993). Even among type II fibers expressing a single myosin heavy
chain isoform, there may be variation in V max related to the myosin light chain composition
(Lowey et al., 1993; Bottinelli et al., 1994) (see also the apparent plasticity of contractile
properties of human muscle fibers as indicated by the training study of Fitts et al., 1989; also
training induced changes in myosin isoforms, Baumann et al., 1987; Schantz & Dhoot,
1987); 2) is arbitrary but not unreasonable for human muscle of mixed composition; 3) is a
simplification, although on functional grounds one might expect differences related to the
range over which task related recruitment of different fiber populations occurs.(see e.g.,
Herzog & ter Keurs, 1988; de Ruiter et al., 1995; Kernell, 1994); and 4) is a simplification
made in the absence of systematic data (for a review of animal muscle see Woledge et al.,
1985; and for human data see Fitts et aI, 1989).
The relative force and power velocity relationships for types I and II fiber populations
in our hypothetical human muscle are shown in Fig. 5. Ifwe assume a summation of power
production from the two populations in the maximally activated whole muscle, the combined
power output will be as shown in Fig. 6. The great difficulty is to know how to relate the
Motor Unit Variability and Sustaining Mechanical Output of Muscle 331

/ __ " ----- Type I

\' I ' , - - T y p e II

\ I, '
\/
I~"
" " \
,'-\~:.
Figure 5. Force and power in relation to velocity for
..... ............ \
......- --. \
a slow (type I) and a fast (type II) population of fibers
that have the same specific isometric force but Vmax ".
in the ratio I : 4 (note the very low power generated
by the type I compared to type II fibers). VELOCITY

relative velocities of Figs. 5 and 6 to human locomotory movement. What we do know is

that in cycling exercise the optimal pedaling rate (Vopt) for maximal power output is -120
rev/min (Fig. 4). Ifit is assumed that this is the velocity at which the maximal power of the
two popUlations combined is generated, then a pedaling rate of 60 rev/min approximates to
the Vopt of the type I fibers. It also implies that the V max of type I fibers will not be exceeded
until a pedaling rate of- 165 rev/min is achieved and that the fastest fibers (type II) have an
optimal velocity well in excess of that seen in normal locomotion. In general these implica-
tions from the model seem not unreasonable. The power/velocity data for the ultra-marathon
runner shown in comparison to normal control subjects in Fig. 4 also seems to fit the general
case.
One important point to note from the model is the much greater power available from
the type II fibers within the locomotory range (see also the experimental human data in Fig.
4). As discussed by Faulkner and Brooks (1993), this implies that fast type II fibers could
with relatively low levels of activation sustain a power output level equivalent to maximally
activated slow type I fibers. This may be of particular significance for muscles containing a
large proportion of type IIA fibers which have a high oxidative potential.

SUSTAINED SUBMAXIMAL POWER OUTPUT

In locomotory function, the ability to sustain submaximallevels of power is impor-
tant. In Fig. 7, the x-axis from Fig. 6 is expanded and submaximal power output levels
approximating 25 and 80% of the maximal power are indicated by dashed lines. At the 80%
level of sustained power there is, by definition, a 20% reserve of the combined power
generating capability when pedaling at 120 rev/min while at 60 rev/min there is no reserve.
Similarly at the 25% level ,there is a reserve of 75% at 120 rev/min but only 55% at 60
rev/min. At first, the greater reserves at the fast pedaling rate might suggest that this rate is

Combined
-..... ",
a:
,<TYpe II
w
~ \
II. \
Figure 6. The component and combined power-velocity \
relationship for a whole muscle (modeled as if composed \
of two discrete populations of fiber types present in equal \
proportions, see text). The superimposed pedaling rates are
derived by reference to the group data in Fig. 4. 60 120 rev/min VELOCITY
332 A. J. Sargeant and D. A. Jones

100%

E 80% ~

...
....'"
:>
E
'•" ~

...
'" .....
E toO >
~ VI

...,......
~
>
"
W
~ OJ

:;r:o--.... =-
----...
0
'" 25% ~:::::::: Figure 7. Submaximallevels of power
(25% and 80% of maximum) in relation
to the first part of the combined power
velocity relationship expanded from Fig.
6. The maximal potential contribution
60 120
from the type I fiber population is also
VELOCITY (pedelllng rate • rev/min) shown (---).

preferable at both levels of power for resisting fatigue and hence sustaining power output.
For the lower level of power output, however, this ignores the potential contribution to power
output from the fatigue resistant type I fibers. In principle, a power output of 25% of
maximum could be achieved at 60 rev/min by recruiting only the type I fibers. At a pedaling
rate of 120 rev/min, however, the same power output can only be attained if there is a minimal
contribution of -15% power from the faster type II fiber population, which ultimately
includes fatigue sensitive fibers. Calculating the minimum possible contribution from the
type II fibers at different fractions of maximal power yields the data shown in Fig. 8. This
suggests that at a low level of sustained power output (25% of maximum), the optimal
pedaling rate to minimize the contribution required from type II fibers is 60 rev/min but at
high levels of sustained power (80% of maximum) the optimum increases to 120 rev/min.
However, as has already been pointed out, there is evidence for a significant element of rate
coding superimposed on the hierarchical pattern of motor unit recruitment in this form of
exercise. It should be emphasized therefore that this analysis is only describing the minimum
contribution necessary from the power generating capability of the type II fibers - assuming
an initial and full recruitment of the type I fiber population.

100

~
"'~
Oc 75 80% Max. power.
ZW
o!:
-;:, 70%
"til
1-0'

2" so 60% Figure 8. Relationship between velocity and

the minimum proportion of the type II power
"til
0"
"'~
50% required at different fractions of maximal power.
The dashed line indicates the progressive in-

--
~o 40%
;:''''
::.:- 25
30%
crease in the velocity at which the minimum
ZtIl contribution from type II power occurs. This is
-",
~>- 25% derived from Fig. 7 and based on the assumption
I- that the type I population is contributing maxi-
0 mally in all cases. As discussed in the text, this
60 120 may not be true leading to underestimation, es-
VELOCITY (pedalling rate . rev/min) pecially at the lower levels .
Motor Unit Variability and Sustaining Mechanical Output of Muscle 333

Interestingly, and despite all the assumptions and simplification, the model does seem
to reflect muscle function. We recently had the opportunity to study a group of experienced
competitive cyclists (Sargeant, 1994b). They were asked to cycle at different constant speeds
on their own bicycles and they were allowed to choose freely the gear at each speed studied.
As the exercise intensity increased so did their freely chosen pedal frequency; from -60
rev/min at the lowest levels to -1l0 at the highest (speed range, 20 - 47 km/hr; V0 2 range,
0.8 - 5.1 Llmin). In addition, it is worth considering the data derived from the world cycling
record for distance achieved in one hour. Knowing the gear ratios, the mean pedaling rate
can be calculated. Over all the years analyzed, these world class athletes consistently chose
a high pedaling rate of around 106 rev/min in order to maintain a high sustained level of
power output (see Sargeant, 1994b; his Table 1).

MECHANICAL EFFICIENCY AND PEDALING RATE

As pointed out previously (Sargeant, 1988), the theoretical advantage described
above of choosing fast pedaling rates at high levels of sustained submaximal power would
be negated if there was a disproportionate increase in the energy cost for the same external
power delivered; that is, mechanical efficiency decreased. Clearly, however, at least in well
trained cyclists, this is either not the case, or the advantage, in terms of fatigue resistance,
outweighs the extra energy cost.
Unfortunately, very little is known about the mechanical efficiency/velocity relation-
ships of different human muscle fiber types. On the basis of animal muscle experiments, it
is usually proposed that maximal efficiency occurs at a velocity that is close to, but slightly
below, the optimum for maximal power (see e.g., Goldspink, 1978; Lodder et aI., 1991;
Rome, 1993). Thus it might be speculated that the efficiency/velocity relationships for type
I and II fibers might be of the general form shown in Fig. 11. By reference to the earlier
discussion regarding the V opt for maximal power in cycling, it can be seen that the maximal
efficiency for the type I fibers might occur at - 50 rev/min and for type II fibers at -150
rev/min. In this model, the mechanical efficiency of the type I and II fiber popUlations is
equal at about 90 rev/min. Also within the normal locomotory range there is a reciprocal
change in the efficiencies of the two types. As a consequence, calculations from the model
predict that mechanical efficiencies in high intensity cycling exercise should be largely
independent of pedaling rate over a wide range from -60 to II 0 rev/min, a prediction
confirmed by experimental data (Zoladz et aI., 1995). It should be remembered, however,
that this is again a greatly simplified two compartment model. The reality is that there should
be a continuum of efficiency/velocity relationships reflecting the continuum of expression
in the contractile proteins.

ACUTE PLASTICITY OF THE NEUROMUSCULAR SYSTEM

The contractile properties of muscle and potential for power output can change
markedly and acutely as a consequence of exercise itself. In some circumstances, exercise
can lead to potentiation of mechanical output (see e.g., Lodder et aI., 1991). More commonly,
the effect of exercise is studied in relation to the generation of fatigue and this is often seen
as a failure of the system, but this approach is misleading and can limit comprehension of
the phenomenon. Alternatively, fatigue might be seen as protective, a down-regulation of
output to prevent an energy crisis (de Haan & Koudijs, 1994). Thus the muscle is temporarily
transformed becoming slower and much less powerful as demonstrated in animal and human
experiments (Figs. 9 and 10; de Haan et aI., 1989; Beelen & Sargeant, 1991). Such a
 334 A. J. Sargeant and D. A. Jones

-
'iii'
'::sc
1000

750
FIU:SII.

...
>- ~
!!
:a... 500
~ Figure 9. Power (arbitrary units) in rela-
250 tion to velocity for fresh and fatigued rat
a:: medial gastrocnemius muscle. Note: in the
W
fatigued state isometric force was reduced
0== 0 to -50% but power to -25% of the value for
El.
0 25 50 75 100 125 150 fresh muscle due to the additive effect of the
decrease in Vmax (derived from de Haan et
VELOCITY (mm/s) a1., 1989).

-
'iii
Ii
.!
1800 FRESII .
~

/1....-1--1 'AT""""
r:
a:: 1500
w
==
0

1
a.

1- 1- i-
~
1200
ct
w
a.
....I 900
ct Figure 10. Human maximal peak
:E power cycling at 5 different pedal-
><
ct ing rates in fresh and fatigued states.
.#
:E j
Mean (± SE) data for 6 subjects
60 90 120 150
(adapted from Beelen & Sargeant,
PEDALLING RATE (rev/min) 1991).

PEDAL RATE
(rev/min)

60 120

>-
o
z
,
w
(3 \
"TYPE II
,
ii: Figure 11. Schematic of the possible relation-
u. \
ships between mechanical efficiency and velocity
w
\
,
\ TYPE I '\ for the modeled fiber types. In the absence of
systematic data, no relative difference between
\ the maximal efficiencies is given; each type is
I
normalized to the same maximum. The velocity
range equivalent to pedaling rates of 60 to 120
VELOCITY rev/min is derived from Fig. 6.
Motor Unit Variability and Sustaining Mechanical Output of Muscle 335

:;-

-•..
....I
~

-.
300
~ - fest

!.
a:
w 12°C
:= 200

0
a..
Figure 12. The relationship of maximal peak
power during isokinetic cycling and pedaling ~

rate. Power is expressed as watts per liter of upper

leg muscle volume. Data is given for normal
""wa..
resting conditions at room temperature and fol- ::E
lowing 45 min immersion in water baths at 44°C, ~

18°C and l2°C. Limits of the experimental data

::E
)(
are given by thicker sections of the lines. Arrows
indicate the optimal velocity which increases with "":::ii 50 100 150 200 250
temperature (adapted from Sargeant, 1987). PEDALLING RATE (rev/min)

transformation might under some circumstances have the additional benefit of improving
the economy of isometric contractions (Jones & Bigland-Ritchie, 1986) and possibly,
depending on contraction velocity and recruitment patterns, even the efficiency of repetitive
dynamic contractions (Sargeant, 1994b).
The contractile properties of muscle will also be acutely modified by temperature
changes, whether as a consequence of metabolic heat production by the muscle itself or due
to environmental conditions. In humans, as would be expected from isolated animal muscle
experiments, there is a velocity-dependent effect of manipUlating muscle temperature (Fig.
12, Sargeant, 1987). Thus, just as exercise-induced fatigue may transform muscle into a
slower type, the exercise-induced increase in muscle temperature will make the muscle faster
which may obscure the true magnitude of the fatigue (Beelen & Sargeant, 1991).

ACKNOWLEDQMENTS
Attendance of the authors at the 1994 Bigland-Ritchie conference was supported, in
part, by the Muscular Dystrophy Association (USA), and the University of Miami (A.J.S.).

REFERENCES
Alexander RMcN (1984). Elastic energy stores in running vertebrates. American Zoologist 24,85-94.
Baumann H, Jaggi M, Soland F, Howald H & Schaub Me (1987). Exercise training induces transitions of
myosin isoform subunits within histochemically typed human muscle fibers. Pflugers Archiv 409,
349-360.
Beelen A & Sargeant AJ (1991). Effect of fatigue on maximal power output at different contraction velocities
in humans. Journal ofApplied Physiology 71, 2332-2337.
Beelen A & Sargeant AJ (1993). Effect of prior exercise at different pedal frequencies on maximal power in
humans. European Journal ofApplied Physiology 66, 102-107.
Beelen A, Sargeant AJ, Lind A, de Haan A, Kernell D & van Mechelen W (1993). Effect of contraction velocity
on the pattern of glycogen depletion in human muscle fiber types. In: Sargeant AJ, Kernell D (eds.),
Neuromuscular Fatigue, pp. 93-95. Amsterdam: Royal Netherlands Academy of Arts and Sciences.
336 A. J. Sargeant and D. A. Jones

Beelen A, Sargeant AJ & Wijkhuizen F (1994). Measurement of directional force and power during human
submaximal and maximal isokinetic exercise. European Journal ojApplied Physiology 69, 1-5.
Bigland-Ritchie B & Woods 11 (1976). Integrated EMG and O2 uptake during positive and negative work.
Journal oj Physiology (London) 260, 267-277.
Bottinelli R, Betto R, Schiaffmo S & Reggiani C (1994). Unloaded shortening velocity and myosin heavy
chain and alkali light chain isoform composition in rat skeletal muscle fibers. Journal oj Physiology
(London) 478, 3421-349.
Bottinelli R, Schiaffino S & Reggiani C (1991). Force-velocity relations and myosin heavy chain isoform
compositions of skinned fibers from rat skeletal muscle. Journal oj Physiology (London) 437,
655-667.
•
Buller AJ, Eccles JC & Eccles RM (1960). Interactions between motoneurones and muscles in respect of the
characteristic speeds of their responses. Journal ojPhysiology (London) 150, 417-439.
Burke RE, Levine DN, Tsairis P & Zajac, FE (1973). Physiological types and histochemical profiles in motor
units of the cat gastrocnemius. Journal ojPhysiology (London) 234, 723-748
Cavagna GA, Dusman B & Margaria R (1968). Positive work done by a previously stretched muscle. Journal
ojApplied Physiology 24, 21-32.
Curtin NA (1990). Force during stretch and shortening of frog sartorius muscle: effects of intracellular
acidification due to increased carbon dioxide. Journal oj Muscle Research and Cell Motility 11,
251-257.
de Haan A, Jones DA & Sargeant AJ (1989). Changes in power output, velocity of shortening and relaxation
rate during fatigue of rat medial gastrocnemius muscle. Pflugers Archiv 413,422-428.
de Haan A & Koudijs JCM (1994). A linear relationship between ATP degradation and fatigue during
high-intensity dynamic exercise in rat skeletal muscle. Experimental Physiology 79, 865-868.
de Haan A, Lodder MAN & Sargeant AJ (1991). Influence of an active pre-stretch on fatigue of skeletal muscle.
European Journal ojApplied Physiology 62, 268-273.
de Ruiter CJ, de Haan A & Sargeant AJ (1995). Physiological characteristics of two extreme muscle
compartments in gastrocnemius medialis of the anaesthetized rat. Acta Physiologica Scandanavica
153, 313-324.
Ennion S, Sant' Ana Pereira JA, Sargeant AJ, Young A & Goldspink G (1995). Characterisation of human
skeletal muscle fibers according to the myosin heavy chains they express. Journal oJMuscle Research
and Cell Motility 16, 35-43
Faulkner JA & Brooks SV (1993). Fatigability of mouse muscles during constant length, shortening, and
lengthening contractions: interactions between fiber types and duty cycles. In: Sargeant AJ, Kernell
D (eds.) Neuromuscular Fatigue, pp. 116-123. Amsterdam: Royal Netherlands Academy of Arts and
Sciences.
Faulkner JA, Jones DA, Round JM & Edwards RHT (1980). Dynamics of energetic processes in human muscle.
In: Cerretelli P, Whipp BJ (eds.), Exercise Bioenergetics and Gas Exchange, pp. 81-90. Amsterdam:
ElsevierlNorth-Holland Biomedical Press.
Fitts RH, Costill DL & Gardetto PR (1989). Effect of swim training on human muscle fiber function. Journal
ojApplied Physiology 66, 465-475.
Goldspink G (1978). Energy turnover during contraction of different types of muscle. In: Asmussen E,
Jergensen K (eds.), Biomechanics VI-A, pp. 27-39. Baltimore: University Park Press.
Gollnick PD, Piehl K & Saltin B (1974). Selective glycogen depletion pattern in human muscle fibers after
exercise of varying intensity and at varying pedaling rates. Journal oj Physiology (London) 241,
45-57.
Greig CA, Sargeant AJ & Vellestad NK (1985). Muscle force and fiber recruitment during dynamic exercise
in man. Journal oj Physiology (London) 371, 176P.
Herzog W & ter Keurs HEDJ (1988). Force-length relation of in-vivo human rectus femoris muscles. Pflugers
Archiv. European Journal oj Physiology 411,642-647.
Hill AV (1950). The dimensions of animals and their muscular dynamics. Proceedings Royal Institute oJGreat
Britain 34, 450-473.
Hill AV (1956). The design of muscles. British Medical Bulletin 12, 165-166.
Ivy JL, Chi M-Y, Hintz CS, Sherman WM. Hellendall RP & Lowry OH (1987). Progressive metabolic changes
in individual human muscle fibers with increasing work rates. American Journal oj Physiology 252,
C630-C639.
Jones DA & Bigland-Ritchie B (1986). Electrical and contractile changes in muscle fatigue. In Saltin B (ed.),
Biochemistry oj Exercise VI, pp. 337-392. Champaign, Ill: Human Kinetics Publishers.
Kernell D (1983). Functional properties of spinal motoneurons and gradation of muscle force. In: Desmedt IE
(ed.), Motor Control Mechanisms in Health and Disease, pp. 213-226. New York: Raven Press.
Motor Unit Variability and Sustaining Mechanical Output of Musde 337

Kernell D (1986). Organization and properties of spinal motoneurones and motor units. Progress in Brain
Research 64, 21-30.
Kernell D (1992). Organized variability in the neuromuscular system: a survey of task-related adaptations.
Archives Italiennes de Biologie 130, 19-66.
Kernell D (1994). Motor tasks and functional organization of the neuromuscular system. Plenary lecture.
Proceedings for Joint Meeting of the Dutch Physiological Society and The Physiological Society, 2P.
Kernell D, Eerbeek 0 & Verhey BA (1983). Motor unit categorization on basis of contractile properties: an
experimental analysis of the composition of the cat's m. peroneus longus. Experimental Brain
Research 50, 211-219.
Komi PV (1984). Physiological and biomechanical correlates of muscle function. Effects of muscle structure
and stretch shortening cycle on force and speed. In: Terjung R (ed.), Exercise and Sports Science
Reviews, vol. 12, pp. 18 -121. Lexington: Collamore Press.
Kugelberg E (1973). Histochemical composition, contraction speed, and fatiguability of rat soleus motor units.
Journal of Neurological Sciences 20, 177-198.
Kukulka CG & Clamann HP (1981). Comparison of the recruitment and discharge properties of motor units
in human biceps brachii and adductor pollicis during isometric contractions. Brain Research 219,
45-55.
Larsson L & Moss RL (1993). Maximum velocity of shortening in relation to myosin isoform composition in
single fibers from human skeletal muscles. Journal ofPhysiology (London) 472,595-614.
Lodder MAN, de Haan A & Sargeant AJ (1991). Effect of shortening velocity on work output and energy cost
during repeated contractions of the rat EDL muscle. European Journal of Applied Physiology 62,
430-435.
Lowey S, Waller GS & Trybus KN (1993). Skeletal muscle myosin light chains are essential for physiological
speed of shortening. Nature 365, 454-456.
McComas AJ, Galea V, Einhorn RW, Hicks AL & Kuiack S (1993). The role of the Na+, K+-pump in delaying
muscle fatigue. In: Sargeant AJ, Kernell D (eds.), Neuromuscular Fatigue, pp. 35-43. Amsterdam:
Royal Netherlands Academy of Arts and Sciences.
Nardone A, Romnano C & Schieppati M (1989). Selective recruitment of high-threshold human motor units
during voluntary isotonic lengthening of active muscles. Journal of Physiology (London) 409,
451-471.
Pette D & Vrbova G (1992). Adaptation of mammalian skeletal muscle fibers to chronic electrical stimulation.
Reviews in Physiology, Biochemistry and Pharmacology 120, 115-202.
Pierotti DJ, Roy RR, Bodine-Fowler SC, Hodgson JA & Edgerton VR (1991). Mechanical and morphological
properties of chronically inactive cat tibialis anterior motor units. Journal of Physiology (London)
444, 175-192.
Rome LC (1992). The design of the muscular system. In: Sargeant AJ, Kernell D (eds.), Neuromuscular
Fatigue, pp. 129-136. Amsterdam: Royal Netherlands Academy of Arts and Sciences
Rome LC, Sosnicki AA & Goble DO (1990). Maximum velocity of shortening of three fiber types from horse
soleus muscle: implications for scaling with body size. Journal ofPhysiology (London) 431, 173-185.
Rutherford OM & Jones DA (1988). Contractile properties and fatiguability of the human adductor pollicis
and first dorsal interosseous: a comparison of the effects of two chronic stimulation patterns. Journal
ofNeurological Science 85,319-331.
Salmons S & Vrbova G (1969). The influence of activity on some contractile characteristics of mammalian
fast and slow muscles. Journal ofPhysiology (London) 201, 535-549.
Saltin B & Rose RJ (eds.) (1994). The racing camel (Camelus dromedarius). Physiology, metabolic functions
and adaptations. Acta Physiologica Scandinavica Supplement 617, 9-18.
Sant'Ana Pereira JA, Wessels A, Nijtmans L, Moorman AFM & Sargeant AJ (1995). New method for the
accurate characterisation of single human skeletal muscle fibers demonstrates a relation between
mATPase and MyHC expression in pure and hybrid fibers types. Journal of Muscle Research and
Cell Motility 16, 21-34.
Sargeant AJ (1987). Effect of muscle temperature on leg extension force and short-term power output in
humans. European Journal ofApplied Physiology 56, 693-698.
Sargeant AJ (1988). Optimum cycle frequencies in human movement. Journal of PhYSiology (London) 406,
49P.
Sargeant AJ (1994a). 'Acute' and 'chronic' plasticity of human muscle and power output. Motor Control
Symposium. Proceedings of the 2nd World Congress of Biomechanics. Stichting World
Biomechanics, Nijmegen, The Netherlands, Volume II, 119.
Sargeant AJ (1994b). Human power output and muscle fatigue. International Journal of Sports Medicine 15,
116-121.
338 A. J. Sargeant and D. A. Jones

Sargeant AJ & Beelen A (1993). Human muscle fatigue in dynamic exercise. In: Sargeant AJ, Kernell D (eds.),
Neuromuscular Fatigue, pp. 81-92. Amsterdam: Royal Netherlands Academy of Arts and Sciences.
Sargeant AJ, Hoinville E & Young A(1981). Maximum leg force and power output during short-term dynamic
exercise. Journal ofApplied Physiology 51, 1175-1182.
Schantz PG & Dhoot GK (1987). Coexistence of slow and fast isoforms of contractile and regulatory proteins
in human skeletal muscle fibers induced by endurance training. Acta Physiologica Scandinavica 131,
147-154.
Thomas CK, Johansson RS & Bigland-Ritchie B (1991). Attempts to physiologically classify human thenar
motor units. Journal ofNeurophysiology 65, 1501-1508.
Unquez GA, Bodine-Fowler SC, Roy RR, Pierotti OJ & Edgerton VR (1993). Evidence of incomplete neural
control of motor properties in cat anterior tibialis after self-reinnervation. Journal of Physiology
(London) 472, 103-125.
V 0llestad NK & Blom PCS (1985). Effect of varying exercise intensity on glycogen depletion in human muscle
fibers. Acta Physiologica Scandinavica 125, 395-405.
V0llestad NK, Vaage 0 & Hermansen L (1984). Muscle glycogen depletion patterns in type I and subgroups
of type II fibers during prolonged severe exercise in man. Acta Physiologica Scandinavica 122,
433-441.
Woledge RC, Curtin NA & Homsher E (1985). Energetic aspects of muscle contraction. Monographs of the
Physiological Society No. 41. London: Academic Press.
Zoladz J, Rademacker A & Sargeant AJ (1995). Oxygen uptake does not increase linearly with power output
at high intensities in humans. Journal ofPhysiology (London) In press.
 25

INTRAMUSCULAR PRESSURES FOR

MONITORING DIFFERENT TASKS AND
MUSCLE CONDITIONS

O. M. Sejersted 1 and A. R. Hargens2

1 Institutefor Experimental Medical Research

University of Oslo
Ullevaal Hospital, Oslo, Norway
2 Gravitational Research Branch
NASA
Ames Research Center, Moffett Field, California

ABSTRACT

Intramuscular fluid pressure (IMP) can easily be measured in man and animals. It follows
the law of Laplace which means that it is determined by the tension of the muscle fibers, the
recording depth and by fiber geometry (fiber curvature or pennation angle). Thick, bulging
muscles create high IMPs (up to 1000 mmHg) and force transmission to tendons becomes
inefficient. High resting or postexercise IMPs are indicative of a compartment syndrome due
to muscle swelling within a low-compliance osseofascial boundary. IMP increases linearly
with force (torque) independent of the mode or speed of contraction (isometric, eccentric,
concentric). IMP is also a much better predictor of muscle force than the EMG signal. During
prolonged low-force isometric contractions, cyclic variations in IMP are seen. Since IMP
influences muscle blood flow through the muscle pump, autoregulating vascular elements,
and compression of the intramuscular vasculature, alterations in IMP have important impli-
cations for muscle function.

INTRODUCTION

IMP, the hydrostatic fluid pressure inside muscles can be measured using various
techniques. It is primarily thought to reflect the hydrostatic pressure within the interstital
matrix gel and the osmotic pressure of the immobilized macromolecules, the sum of which
is the pressure that equilibrates with the free fluid in the interstitium (Wiig, 1990). In resting
muscles fiber tension is low, so pressure gradients inside the muscle are small. In most resting
muscles compliance of the interstitial fluid compartment is high and hydrostatic and colloid
osmotic pressures are of the same order of magnitude (Reed & Wiig, 1981). There is a small
hydrostatic pressure gradient out of the capillaries (Aukland & Reed, 1993). However, during

339
340 O. M. Sejersted and A. R. Hargens

contraction, the rigid muscle cells are able to withstand pressure just like the heart wall, and
quite large pressure gradients can develop.
The physiological and clinical interest in IMP is related to several aspects of muscle
function, of which three seem the most important. First, fluid balance inside muscles is highly
dependent on hydrostatic pressures since these pressures constitute part of the Starling forces
that govern fluid transfer across the capillary wall and therefore also the volume of the
interstitial space. High IMP will impair fluid filtration, but promotes lymph formation
(Levick, 1991). However, as IMP increases, blood flow through the muscle will decrease
either due to an increased resistance at the level of venous outlet from the muscle or because
the capillaries collapse. Reduced microcirculation will cause ischemia. Hence, the second
important factor relates to metabolic control ofthe muscle cells. High pressures will promote
anaerobic metabolism. The third aspect is purely biomechanicai. High IMP will reduce
transmission of force to muscle tendons as outlined below.
With this background, we will review the literature on IMP. First, we describe the
factors that determine IMP and how it can be measured in various muscles. Second, we will
discuss the task dependency of IMP and its relation to the EMG signal. Finally, the
importance of IMP as an independent determinant of muscle function will be discussed.

FACTORS THAT DETERMINE INTRAMUSCULAR FLUID

PRESSURES

Laplace's Law
In physical terms pressure within a compartment is determined by the properties and
the geometry of the wall. This was also pointed out by Hill (1948) who stated that muscle
fibers "that are curved ... must necessarily exert a pressure inwards when they contract,
depending on their tension and radius of curvature." Modified for skeletal muscle with one
fiber direction, the law of Laplace states that

s
P = Po + n!':.h-
r

where P is IMP at any point, Po is pressure beneath the fascia, n is the number of concentric
muscle layers, !':.h is thickness of each layer, s is tension or stress of the muscle cells and r
is radius of fiber curvature (Sejersted et aI., 1984). Since we know little of the actual geometry
of muscle layers and individual fibers in skeletal muscle, the above version of the law might
not predict pressures accurately in all muscles. However, the law of Laplace has been
successfully applied to the heart (Huisman et aI., 1980; Skalak, 1982). Two extreme
situations will illustrate the consequences of the important principles of this law. First,
imagine a muscle with straight, parallel fibers attached to a broad tendon sheet so that the
tendon tension vector has exactly the same direction as the muscle fibers. During an isometric
contraction there will be no curving of fibers and all of the force of contraction will be
transmitted to the tendon. Since there is no force vector in other directions, no pressure will
develop. This prediction has been corroborated by measurements in the diaphragm and
trapezius muscle in which low IMP was found during contraction (Supinski et aI., 1990;
Decramer et aI., 1990; larvholm et aI., 1991). Second, the other extreme would be a circular
muscle fiber. During contraction no force will be transmitted to any tendon, and again if the
fiber is prevented from shortening, all of the developed tension will be directed inwards
towards the center of the ring. In the heart this is how pressure develops in the ventricular
cavity. Most skeletal muscles will have fiber geometries in which the direction of muscle
Intramuscular Pressures for Monitoring Different Tasks 341

fiber force development does not align with the direction that force is transmitted through
the tendons. Hence, fibers will tend to curve and force vectors will be present perpendicular
to force in the tendon. These force vectors will elevate IMP. Thus, in spindle shaped muscles
that bulge during contraction, and in pennate structured muscles, one would predict high
contraction pressures (Sejersted & Hargens, 1986). In fact pressures close to 600 mmHg
have been found in the human vastus medialis and supraspinatus muscles (Sejersted et aI.,
1984; Jarvholm et aI., 1991).
In addition to muscle fiber tension (s) and radius of curvature (r), muscle thickness
(n·M) determines IMP according to the law ofLaplace. This has been extensively corroborated.
IMP often increases linearly as a function of recording depth in many muscles (Sejersted et aI.,
1984; Jarvholm et aI., 1988). Large bulging muscles will not necessarily mean that a lot offorce
is transmitted to tendons, but rather that increased muscle fiber tension creates high IMPs.

Starling Forces
An important term in the equation above is the pressure beneath the fascia (Po). If
the muscle is surrounded by a thick fascia or bone so that compliance of the whole muscle
compartment is small, Po will rise if the muscle swells. Provided there is no tension of the
muscle(s = 0, which is probably never quite true), .the pressure will be homogenously
elevated throughout the compartment. Clearly contraction pressures will also be elevated.
Thus, elevated resting IMP is a cardinal sign of the chronic compartment syndrome that can
develop in the tibialis anterior muscle, for example (Pedowitz et aI., 1990). Supraspinatus
muscle and various forearm muscles are also encapsulated in a compartment (Mubarak et
aI., 1978; Jensen, 1991). Most other muscles have elastic fasciae, and volume changes can
occur with very little change of Po.
Hence, muscle volume is important not only because it might affect fiber geometry, but
because in certain muscles it will determine Po. Muscle volume can vary by at least 10-15 %
under normal conditions. During exercise, fluid is taken up by the active muscle. The reason
for this fluid transfer is not quite clear (see McComas & Sjf1Jgaard, Chapter 4). Vasodilation
leads to higher capillary perfusion pressures, and probably also a larger capillary surface area
which will promote ultrafiltration. However, muscle swelling is better related to exercise
intensity than to muscle blood flow, since swelling is linearly related to power also at very high
intensities whereas perfusion seems to reach a maximum at lower powers, at least during
running or bicycling (Gullestad et aI., 1993). Intracellular breakdown of creatine phosphate and
accumulation oflactate probably creates an osmotic driving force for fluid uptake by the cells
and interstitial colloid osmotic pressure will rise. Interestingly, this combination of hydrostatic
and osmotic forces can cause the muscle to swell within seconds, whereas restoration of the
muscle volume is a slow process (Sejersted et aI., 1986). During recovery, the intracellular
osmoles are removed quite rapidly so that fluid will leave the cells and accumulate in the
interstitium. However, reabsorption of fluid into the capillaries and drainage into peripheral
lymphatic vessels is slow since it is driven primarily by the colloid osmotic pressure gradient
and small hydrostatic pressure gradients, respectively. Hence, elevated post-exercise IMP is
quite characteristic of the chronic compartment syndrome (Pedowitz et aI., 1990).

MEASUREMENTS OF INTRAMUSCULAR FLUID PRESSURES

Methods of Measurements
IMP has been measured using several techniques. Hill (1948) measured the tempera-
ture rise in an oil-filled syringe attached to a needle inserted into a frog muscle. Compression
342 o. M. Sejersted and A. R. Hargens
of oil will cause a rise in temperature. Subsequent investigators have used similar fluid filled
systems, but with different pressure sensors. For example, pressure inside implanted porous
or perforated capsules or cylinders have been used (Guyton, 1963; Wiig, 1990). The wick
technique was introduced by Scholander and colleagues (1968) and later modified by Fadnes
and colleagues (1977) to a wick-in-needle method. Pressure can also be measured by
micropipette techniques. None of these methods are well suited for measurements in the
contracting muscle. Wick catheters respond slowly to pressure changes, and micropipettes
are difficult to keep in place and do break. The simplest technique and one which is similar
to the one used by Hill (1948) is to insert a fluid-filled small-bore catheter into the muscle
and record pressure with an ordinary pressure transducer (Sejersted et aI., 1984). The tip of
the catheter has been modified in various ways to ensure that it will remain open, and other
investigators have infused saline at slow or intermittent rates (Styf & Korner, 1986a).
Alternatively transducer-tipped catheters have been inserted into muscle (Sejersted et aI.,
1984; Crenshaw et aI., 1992; Ballard et aI., 1994b). These transducers are now available in
very small sizes and have been used to record pressures in the small thyroarythenoid muscle
of dogs and in human muscles (Cooper et aI., 1993; Crenshaw et aI., 1992). The virtue of
the transducer-tipped catheters is of course that one avoids the signal dampening effect and
oscillations of fluid filled systems. The frequency response is also very high. Potential
problems comprise mechanical interference with the sensor area and lack of ways to
zero-adjust during recording. There seems to be little difference between steady state
pressures measured by these methods (Wiig, 1985; Wiig, 1990), but the open catheters or
transducer-tipped catheters are the only means by which to measure rapid changes (Sejersted
et aI., 1984).

Intramuscular Fluid Pressures in Different Muscles

IMP has been measured in a variety of muscles in many species. Resting pressures
range from -2 to +14 mmHg in normal muscle and are higher than those reported in
subcutaneous tissue (Styf & Korner, 1987; Wiig, 1990; Pedowitz et aI., 1990). This range of
variation is primarily due to the position of the muscle relative to the rest of the body and
the tightness of osseofascial boundaries. It is also difficult to compare IMP obtained from
intact muscles with data from experiments in which the muscle surface has been exposed
and compression forces due to skin and fasciae have been removed. In an upright or sitting
position, IMP of resting thigh muscles in humans normally ranges between 0 to 10 mmHg
(Sejersted et aI., 1984).
During contraction, IMP will rise in most muscles. During maximal voluntary
contraction IMP can vary from below 40 mmHg in the masseter muscle to above 1000 mmHg
(Sylvest & Hvid, 1959; Sadamoto et aI., 1983) in the vastus medialis muscle. Later studies
have confirmed this variation range, but it is interesting that even in quite small muscles like
the frog gastrocnemius, the dog arythenoid and parasternal intercostal muscles, maximum
pressures can be 100 mmHg or more (Hill, 1948; Leenaerts & Decramer, 1990; Cooper et
aI., 1993). The variability is explained by Laplace's law and underscores the importance of
muscle morphology.
Similarly, repeated or multiple simultaneous measurements within the same muscle
will show large differences due to variable positioning of the pressure sensor. The depth of
recording is especially important (Sejersted et aI., 1984; Jensen, 1991; Nakhostine et aI.,
1993). Fig. 1 shows a typical example of force and pressure recordings in the vastus medialis
muscle of one human subject. IMP was recorded by three different techniques and varied by
almost three-fold between recording sites. In this experiment two open-tipped catheters and
a transducer-tipped catheter were used. In another study IMPs in four shoulder muscles were
compared. Maximum contraction pressures of 400-600 mmHg were observed in the supra-
Intramuscular Pressures for Monitoring Different Tasks 343

Figure 1. Force and three IMPs in the right human vastus lateralis muscle during isometric knee extension in
the sitting position. Force was recorded at the ankle. First and last contractions are MVCs, and 3 to 5 min rest
was allowed between contractions at 70, 50, 30 and 15% MVC. Pressure was recorded by a Millar Mikrotip
pressure transducer (A), Bentley Trantec transducer (B), and AE840 transducer (C) (reproduced with permis-
sion from Sejersted et ai., 1984)

and infraspinatus muscles, whereas pressure in the deltoid and trapezius muscles were just
above 100 mmHg (Jarvholm et aI., 1991). Muscle fiber geometry and muscle thickness
probably explain this variability.

TASK DEPENDENCY OF INTRAMUSCULAR FLUID PRESSURES

Relationship to Muscle Length and Mode of Contraction

Since fiber geometry and tension are important determinants ofIMP, muscle length might
affect both resting and contraction pressures. During passive stretch the radius of curvature of the
fiber will increase (pennation angle will decrease). The actual outcome will depend on which
factor dominates. For example, in dog parasternal intercostal muscles, stretch causes quite high
pressures (Leenaerts & Decramer, 1990), whereas stretching in the tibialis anterior muscle of
humans causes a pressure rise up to 35 mmHg (Baumann et aI., 1979; Gershuni et aI., 1984).
During contraction it has been shown first by Mazella (1954), and later by many
investigators that IMP is linearly related to force or torque (Sadamoto et ai., 1983; Sejersted
et aI., 1984; Sj0gaard et ai., 1986; Jarvholm et aI., 1989; Leenaerts & Decramer, 1990;
Jarvholm et aI., 1991; Aratow et aI., 1993). The slope of this relationship, however, varies
both within and between muscles (Sejersted et aI., 1984; Jensen, 1991; Jarvholm et aI., 1991).
As seen from Fig. 2, the linear relationship between IMP and torque is independent of
contraction velocity, intensity and mode of contraction (isometric or isokinetic). Compari-
sons between eccentric and concentric exercise also indicate that IMP does not differ greatly
344 O. M. Sejersted and A. R. Hargens

in those contractions except at extreme muscle lengths (Ballard et aI., 1994a). As expected,
IMP was greater per unit torque at short muscle lengths.

Cyclic Variations of IMP at Low Force, Prolonged Contractions

There are no studies that have examined contraction pressure during intermittent
contractions or dynamic exercise lasting more than a few minutes. It seems, however, that with
continuous recording in one place, pressure measurements are reproducible between contrac-
tions, except when muscle compartment compliance decreases due to swelling (as discussed
below), and the IMP force relationship is constant for a given recording. However, repeated
measurements may give different slopes due to variability in pressure gauge position.
The tight linear relationship between overall force and IMP seems to be broken during
low force prolonged isometric contractions. Two reports show that IMP at a given location may
change considerably with contractions maintained for more than a couple of minutes. Sejersted
and colleagues (1984) found oscillations ofIMP amounting to 60% of the highest recorded
pressure in the vastus medialis muscle when force was maintained at 15% of MVC. The
oscillation frequency was about I min-I. These authors also reported an experiment during
which the subject maintained one pressure constant for one catheter at about 180 mmHg. Over
a two minute period force declined by 30 % and the pressure in a nearby catheter fell initially
and thereafter remained fairly constant. Similarly Sj0gaard and colleagues (1986) reported that
during a maintained 5 % MVC contraction IMP in the rectus femoris muscle was quite variable.
In one location it was maintained stable for one hour. In other locations cyclic variations were
observed with a pressure variation range and periodicity similar to that seen by Sejersted and
colleagues (1984). On one occasion a period of low pressure was closely associated with
changes of the EMG signal. The importance of these observations has never been examined
properly. The pressure oscillations must reflect variation in local muscle fiber tension. The
absolute muscle volume that the pressure catheter can see is unknown. It is probably in the range
of several cm3 and is highly dependent on the depth of recording and interstitial compliance.
Hence, it seems that quite large portions of the muscle can be recruited and turned off in a
synchronous and cyclic manner, possibly to prevent fatigue.
In conclusion, task dependency of IMP is related primarily to the required force
output, and under normal circumstances other task requirements will not affect pressure. The
only known exception is prolonged isometric contractions during which cyclic changes in
local IMP occur. These changes probably reflect the motor unit recruitment pattern and are
observations that need further study.

400 y = 2.47x + 3.17

r = 0.97

300 .6
Cl
I ~~
.sE 200 M'
Figure 2. Soleus IMP and ankle joint torque during
isometric and isokinetic (concentric) contractions in one

~
a..

.,.,
subject. Contractions of various intensities were per-
~

100 - ... isometric formed at velocities of 0 (isometric), 60, 120 and 240

.... 0
0
/::;.
60 0/sec
1200/sec
2400/sec
degrees/so Isometric contractions were performed at 5
different joint angles. IMP was measured using a 3F Milar
transducer-tipped catheter. Knee and hip joints were
o Ii i·' flexed at 90°. The relationship between IMP and torque
0 50 100 150 is independent of velocity, intensity, and mode of contrac-
Torque (Nm) tion.
Intramuscular Pressures for Monitoring Different Tasks 345

THE RELATIONSHIP BETWEEN INTRAMUSCULAR FLUID

PRESSURE AND THE ELECTROMYOGRAPHIC SIGNAL (EMG)
The linear relationship between force and IMP has prompted several investigators to
suggest that IMP can be used as a force indicator during various tasks where force cannot
be directly measured (Parker et aI., 1984; Jiirvholm, 1990; Aratow et aI., 1993). Especially
with complex movements involving many muscles, the force contribution from one single
muscle might be assessed from pressure measurements. The alternative and widely used
non-invasive method is to measure EMG. However, there are various disadvantages with
this technique, for example: non-linear calibration curves, low reproducibility, and fatigue
related changes (Westgaard, 1988).Surface EMG can also pick up signals from several
muscles. During brief isometric contractions there seems to be a fairly good relationship
between the integrated or root mean square EMG signal and IMP (Jiirvholm et aI., 1989;
Jiirvholm et aI., 1991; Aratow et aI., 1993). Due to fatigue, however, IMP was superior as
an estimator of muscle force (Parker et aI., 1984). Interestingly, when movement is intro-
duced, the relationship between EMG and force seems to change. In contrast, the relationship
between IMP and force is preserved as outlined above (Aratow et aI., 1993). Hence, during
concentric contractions of either the soleus or tibialis anterior muscles in humans, the
coefficients of determination (R2) for the torque vs. IMP are significantly higher than those
between torques and EMG. The discrepancy becomes much more pronounced during
eccentric work. Fig. 3 shows that IMP is almost linearly related to torque, whereas EMG

300
A
mean R2 = 0.91 ± 0.02

Ci 200
:c
E
.s
a.
~ 100

Figure 3. Soleus IMP (A; n-9) and EMG (B; n

= 9) during eccentric exercise. Each line rep-
resents one MVC from a subject up to a maxi-
....I....-'jF----r--r----.--~---,-- mal torque. Linear regression analysis of each
80 100 line yielded an R2 value. Mean ± SE of all R2
values is shown (reproduced with permission
from Aratow et ai., 1993).
346 O. M. Sejersted and A. R. Hargens

I- GRFz - Soleus IMP -TAIMPI Figure 4. Soleus and tibialis anterior

200 150 E
Cl
IMP during treadmill walking. The
'0; graph represents average data from nine
~ 150 ~
100 ~
subjects, four gait cycles per subject.
.s 100
E N=9 0
.c
IMPs were measured using 3 F Millar
transducer-tipped catheters. GRFz is the
0.. 50 '0 vertical ground reaction force measured
~ 50 ~
~
u. N by an insole force sensor. In all subjects,
0 0 a: soleus IMP showed a single peak near
C!)
0 25 50 75 100 the push-off phase of gait while tibialis
Gait cycle (%) anterior was biphasic.

shows a highly curved relationship. The reason is that during eccentric exercise there is a
force peak late during the contraction that is not reflected in the EMG signal, only in IMP.
Thus for normal movements comprising a mixture of eccentric, concentric as well as
isometric contractions, IMP seems by far the best estimator of local muscle force.
IMP measurements have, therefore, been applied to analyze muscle function during
walking. In an early study by Baumann and colleagues (1979), wick catheters were used,
but responded too slowly. In a more recent study, transducer-tipped catheters were used
(Ballard et aI., 1994b). Fig. 4 shows a recording during one gait cycle in 9 subjects. At heel
strike, there is a pressure peak in the tibialis anterior muscle. It then relaxes while IMP builds
to a peak in soleus at push-off. There is a simultaneous second IMP peak in the tibialis anterior
at this point. Because IMP correlates linearly with force, IMP magnitude provides informa-
tion about the magnitude of contraction force for specific muscles during movement.

THE IMPORTANCE OF IMP AS A DETERMINANT OF MUSCLE

FUNCTION

Muscle Pump
An increase in IMP will of course squeeze blood out of the muscle. This is the
mechanism by which blood is propelled back to the heart, and the efficiency of this system
depends on functional venous valves (Hargens et aI., 1987).
During rhythmic exercise muscle blood flow increases rapidly, probably due to
vasodilation Folkow and colleagues (1970) and more recently Sheriff and coworkers (1993)
proposed that since the small venules and veins inside the muscle will be emptied during
contraction, the elastic recoil of these vessels on relaxation creates a low intraluminal
pressure. Hence, the pressure gradient across the capillary bed will be greatly enhanced. This
would be sufficient to increase flow substantially. This is in keeping with the finding that
the initial flow increment and its timing is independent of cardiac output, arterial blood
pressure, power, or autonomic function. Hence, the muscle pump is not only a push-forward
mechanism, but also exerts a sort of pull backwards.

Autoregulation of Arteriolar Resistance (Bayliss Reflex)

Another mechanism by which IMP might affect muscle blood flow is through
autoregulation of the resistance vessels. Autoregulation is very strong in the vascular beds
of organs like the kidneys, the brain and also the coronary circulation in the sense that blood
flow is maintained constant over a wide perfusion pressure range. In skeletal muscle
autoregulation is also present, but has a moderate gain (Mellander et aI., 1987; Braakman et
Intramuscular Pressures for Monitoring Different Tasks 347

aI., 1990). One important trigger of autoregulatory dilation during reduction in perfusion
pressure is the Bayliss reflex. An decreased pressure gradient across the arteriolar wall will
sometimes elicit dilation, and it can be created by raised IMP. However, it is well known that
increased interstitial pressure in the kidneys will elicit preglomerular vasodilation. The role
of autoregulatory vasodilation when IMP is elevated is moderate (Mellander & Albert, 1994).

Compression of Vessels of the Intramuscular Circulation

The third relation between IMP and nutritional flow relates to how IMP affects
nutritional flow. Clearly, during contractions arterial inflow to the muscle is abruptly
reduced, and may be transiently reversed due to expulsion of blood by compression of vessels
(V 011estad et aI., 1990). Compression of capillaries and small vessels may occur at pressures
above 30-50 mmHg (Hargens et aI., 1989). During intermittent contractions (dynamic or
isometric) perfusion may proceed during relaxation periods and the periods with high IMP
limit average muscle perfusion very little (Styf et aI., 1987; V0llestad et aI., 1990). Normally,
resting pressures are rapidly regained when contraction ceases (Sejersted et aI., 1984; our
Fig. 1).
Previously, it has been believed that the intramuscular vascular bed behaves like a
collapsible tube, a Starling resistor, which means that with increasing IMP a waterfall will
appear on the venous side. Intravascular pressures will build up proximal to this waterfall
since flow is reduced, and to some extent also because of vasodilation on the arteriolar side.
Thus, total occlusion of the vascular bed only occurs when pressures exceed the arterial
pressure head. It is also important that with rising IMP, but before occlusion ofthe capillaries,
the driving force for flow through the muscle is no longer the arteriovenous pressure
difference but the difference between arterial pressure and IMP. The concept of a Starling
resistor is to a large extent correct, but Mellander and Albert (1994) have recently argued
that the final short segment of the vein is compressed before it leaves the muscle creating a
venous outflow orifice resistance. This differs from the behavior ofa collapsible tube. Thus,
during sustained isometric contractions or in instances where resting pressures are elevated,
nutritional blood flow may be insufficient. Due to the large variability in the IMP-force
relationship in these situations, IMP is a much better predictor of intramuscular blood flow
than force. Again it is important to recognize that IMP during contraction is highly dependent
on depth, whereas with compartment syndromes, the high pressure is usually more uniform
throughout the compartment. Thus, with maintained isometric contractions only regional
ischemia may occur, whereas with compartment syndromes the ischemia is usually more
general.
The case of continuous isometric contraction is especially interesting, because fatigue
may occur subsequent to ischemia. With repeated, prolonged episodes oflow level isometric
contractions, it has been claimed that the microcirculation is disturbed which can cause
chronic myalgia (Henriksson & Bengtsson 1991). There is no evidence for this. In the study
ofSj0gaard and colleagues (1988), flow was well maintained at contraction forces up to 10%
of maximal voluntary force in the quadriceps muscles. Fatigue could not be related to lactate
output either, because the levels were quite low. Furthermore after isometric contractions at
this relative force, there is usually no relaxation hyperemia. Only at higher contraction forces
and IMPs does ischemia occur.

Compartment Syndromes
Ischemia due to compression of the microvasculature is accompanied by lactate
accumulation, which promotes further muscle swelling and elevation ofpressure. As outlined
above, muscle swelling during exercise may raise resting IMP and reduce compliance of the
348 o. M. Sejersted and A. R. Hargens
compartment further. Therefore, in chronic exertional compartment syndromes, elevated
post-exercise or relaxation pressures are primary diagnostic features. Interestingly, delayed
muscle soreness of the tibialis anterior muscle after eccentric exercise was also associated
with elevation of both resting and contraction pressures probably secondary to increased
muscle volume and strain-induced trauma (Friden et aI., 1986). In one study of patients with
trapezius myalgia, elevation of both resting and contraction pressures was associated with
extensive fibrosis (Hagert & Christenson, 1990). The authors suggest that this kind of
myalgia is a functional compartment syndrome. Similarly, it seems that the supraspinatus
muscle is surrounded by a tight compartment, and it has been suggested that shoulder pain
could emanate from this muscle (Jensen, 1991).
Clearly, muscle swelling, elevated resting IMP and ischemia as well as ischemic pain
seem to be related causally. Successful treatment therefore depends on increasing compliance
of the muscle compartment by fasciotomy or some other means (Qvarfordt et aI., 1983; Styf
& Komer, 1986b; Styf et aI., 1987).

CONCLUSIONS

IMP is a sensitive and direct measure of muscle force. It is easily monitored both in
animals and humans. The large variability within and between muscles is explained by the
law of Laplace, but at anyone location, IMP is highly reproducible. IMP is related linearly
to force independent of the mode or speed of contraction (isometric, concentric, eccentric).
One exception is the cyclic variations in IMP during low-force prolonged isometric contrac-
tions. Furthermore, IMP is an important determinant of muscle blood flow and as such, has
important influences on muscle function.

ACKNOWLEDGMENTS
This research was supported by the Anders Jahre's Fund for the Promotion of Science
and by National Aeronautics and Space Administration grants 199-14-12-04 and 199-26-12-
38. Attendance of [Link]. atthe 1994 Bigland-Ritchie conference was supported, in part, by
the Muscular Dystrophy Association (USA).

REFERENCES
Aratow M, Ballard RE, Crenshaw AG, StyfJ, Watenpaugh DE, Kahan NJ & Hargens AR (1993). Intramuscular
pressure and electromyography as indexes of force during isokinetic exercise. Journal of Applied
Physiology 74, 2634-2640.
Aukland K & Reed RK (1993). Interstitial-lymphatic mechanisms in the control of extracellular fluid volume.
Physiological Reviews 73, 1-7S.
Ballard RE, Styf J, Watenpaugh DE, Aratow M, Crenshaw AG & Hargens AR (1994a). Intramuscular pressure
per unit torque varies with joint angle. Transactions Orthopedic Research Society 19, 6S6.
Ballard RE, Watenpaugh DE, Breit GA, Murthy G, Whalen RT & Hargens AR (1994b). Intramuscular pressure
measurement for assessing muscle function during locomotion. Medicine and Science in Sports and
Exercise 26, 5141
Baumann JU, Sutherland DH & Hiinggi A (1979). Intramuscular pressure during walking: An experimental
study using the wick catheter technique. Clinical Orthopaedics and Related Research 145,292-299.
Braakman R, Sipkema P & Westerhof N (1990). Two zero-flow pressure intercepts exist in autoregulating
isolated skeletal muscle. American Journal of Physiology 258, HIS06-HISI4.
Cooper DS, Pinczower E & Rice DH (1993). Thyroarytenoid intramuscular pressures. Annals of Otology
Rhinology and & Laryngology 102,167-175.
Intramuscular Pressures for Monitoring Different Tasks 349

Crenshaw AG, Styf JR & Hargens AR (1992). Intramuscular pressures during exercise - an evaluation of a
fiber optic transducer-tipped catheter system. European Journal ofApplied Physiology and Occupa-
tional Physiology 65, 178-182.
Decramer M, Jiang TX & Reid MB (1990). Respiratory changes in diaphragmatic intramuscular pressure.
Journal ofApplied Physiology 68, 35-43.
Fadnes HO, Reed R & Aukland K (1977). Interstitial fluid pressure in rats measured with a modified wick
technique. Microvascular Research 14,27-36.
Folkow B, Gaskell P & Waaler BA (1970). Blood flow through limb muscles during heavy rhythmic exercise.
Acta Physiologica Scandinavica 80, 61-72.
Friden J, Sfakianos PN & Hargens AR (1986). Muscle soreness and intramuscular fluid pressure - comparison
between eccentric and concentric load. Journal ofApplied Physiology 61, 2175-2179.
Gershuni DH, Yaru NC, Hargens AR, Lieber RL, O'Hara RC & Akeson WH (1984). Ankle and knee position
as a factor modifying intracompartmental pressure in the human leg. Journal of Bone and Joint
Surgery 66A, 1415-1420.
Gullestad L, Hallen J & Sejersted OM (1993). Variable effects of f3-adrenoceptor blockade on muscle blood
flow during exercise. Acta Physiologica Scandinavica 149, 257-271.
Guyton AC (1963). A concept of negative interstitial pressure based on pressures in implanted perforated
capsules. Circulation Research 12,399-414.
Hagert CG & Christenson JT (1990). Hyperpressure in the trapezius muscle associated with fibrosis. Acta
Orthopaedica Scandinavica 61, 263-265.
Hargens AR, Akeson WH, Mubarak SJ, Owen CA, Gershuni DH, Garfin SR, Lieber RL, Danzig LA, Botte
MJ & Gelberman RH (1989). Tissue fluid pressures: from basic research tools to clinical applications.
Journal of Orthopaedic Research 7, 902-909.
Hargens AR, Millard RW, Pettersson K & Johansen K (1987). Gravitational haemodynamics and oedema
prevention in the giraffe. Nature 329, 59-60.
Henriksson KG & Bengtsson A (1991). Fibromyalgia - a clinical entity? Canadian Journal of Physiology &
Pharmacology 69, 672-677.
Hill AV (1948). The pressure developed in muscle during contraction. Journal of Physiology (London) 107,
518-526.
Huisman RM, Sipkema P, Westerhof N & Elzinga G (1980). Comparison of models used to calculate left
ventricular wall force. Medical and Biological Engineering and Computing 18, 133-144.
Jarvholm U (1990). On Shoulder Muscle Load. An Experimental Study ofMuscle Pressures, EMG and Blood
Flow, (PhD Thesis). Gothenburg: University of Gothenburg.
Jarvholm U, Palmerud G, Herberts P, Hogfors C & Kadefors R (1989). Intramuscular pressure and elec-
tromyography in the supraspinatus muscle at shoulder abduction. Clinical Orthopaedics and Related
Research 245, 102-109.
Jarvholm U, Palmerud G, Karlsson D, Herberts P & Kadefors R (1991). Intramuscular pressure and elec-
tromyography in four shoulder muscles. Journal of Orthopaedic Research 9, 609-619.
Jarvholm D, Palmerud G, Styf J, Herberts P & Kadefors R (1988). Intramuscular pressure in the supraspinatus
muscle. Medical and Biological Engineering and Computing 6, 230-238.
Jensen BR (1991). Isometric Contractions of Small Muscle Groups, (Ph.D. Thesis). Copenhagen: Danish
National Institute of Occupational Health.
Leenaerts P & Decramer M (1990). Respiratory changes in parasternal intercostal intramuscular pressure.
Journal ofApplied Physiology 68, 868-875.
Levick JR (1991). Capillary filtration absorption balance reconsidered in light of dynamic extravascular
factors. Experimental Physiology 76,825-857.
Mazella H (1954). On the pressure developed by the contraction of striated muscle and its influence on
muscular circulation. Archives Internationales de Physiologie 62, 334-347.
Mellander S & Albert D (1994). Effects of increased and decreased tissue pressure on haemodynamic and
capillary events in cat skeletal muscle. Journal of Physiology (London) 481, 163-175.
Mellander S, Maspers M, Bjornberg J & Andersson LO (1987). Autoregulation of capillary pressure and
filtration in cat skeletal muscle in states of normal and reduced vascular tone. Acta Physiologica
Scandinavica 129, 337-351.
Mubarak SJ, Owen tA, Hargens AR, Garetto LP & Akeson WH (1978). Acute compartment syndromes:
diagnosis and treatment with the aid of the wick catheter. Journal of Bone and Joint Surgery 60A,
1091-1095.
Nakhostine M, Styf JR, van Leuven S, Hargens AR & Gershuni DH (1993). Intramuscular pressure varies with
depth. The tibialis anterior muscle studied in 12 volunteers. Acta Orthopaedica Scandinavica 64,
377-381.
350 O. M. 8ejersted and A. R. Hargens

Parker PA, Komer LM & Kadefors R (1984). Estimation of muscle force from intramuscular pressure. Medical
and Biological Engineering and Computing 22, 453-457.
Pedowitz RA, Hargens AR, Mubarak 8J & Gershuni DH (1990). Modified criteria for the objective diagnosis
of chronic compartment syndrome of the leg. American Journal of Sports Medicine 18, 35-40.
Qvarfordt P, Christenson IT, EkliifB, Ohlin P & 8altin B (1983). Intramuscular pressure, muscle blood flow,
and skeletal muscle metabolism in chronic anterior tibial compartment syndrome. ClinicalOrthopae-
dics and Related Research 179, 284-291.
Reed RK & Wiig H (1981). Compliance of the interstitial space in rats. Acta Physiologica Scandinavica 113,
297-305.
Sadamoto T, Bonde-Petersen F & Suzuki Y (1983). Skeletal muscle tension, flow, pressure and EMG during
sustained isometric contractions in humans. European Journal of Applied Physiology and Occupa-
tional Physiology 51, 395-408.
Scholander PF, Hargens AR & Miller SL (1968). Negative pressure in the interstitial fluid of animals. Science
161,321-328.
Sejersted OM & Hargens AR (1986). Regional pressure and nutrition of skeletal muscle during isometric
contraction. In: Hargens AR (ed.), Tissue Nutrition and Viability. pp. 263-283. New York: Springer
Verlag.
Sejersted OM, Hargens AR, Kardel KR, Blom P, Jensen 0 & Hermansen L (1984). Intramuscular fluid pressure
during isometric contraction of human skeletal muscle. Journal of Applied Physiology 56,287-295.
Sejersted OM, V011estad NK & Medb0 JI (1986). Muscle fluid and electrolyte balance during and following
exercise. Acta Physiologica Scandinavica Supplementum 556, 119-127.
SheriffDD, Rowell LB & Scher AM (1993). Is rapid rise in vascular conductance at onset of dynamic exercise
due to muscle pump. American Journal of Physiology 265, HI227-HI234.
Sj0gaard G, Kiens B, J0rgensen K & Saltin B (1986). Intramuscular pressure, EMG and blood flow during
low-level prolonged static contraction in man. Acta Physiologica Scandinavica 128, 475-484.
Sj0gaard G, Savard G & Juel C (1988). Muscle blood flow during isometric activity and its relation to muscle
fatigue. European Journal of Applied Physiology and Occupational Physiology 57, 327-335.
Skalak R (1982). Approximate formulas for myocardial fiber stresses. Journal ofBiomechanical Engineering
104,162-163.
Styf J & Komer LM (1986a). Microcapillary infusion technique for measurement of intramuscular pressure
during exercise. Clinical Orthopaedics and Related Research 207, 253-262.
Styf J & Komer LM (1986b). Chronic anterior-compartment syndrome of the leg - results of treatment by
fasciotomy. Journal of Bone and Joint Surgery 68A, 1338-1347.
Styf J & Komer LM (1987). Diagnosis of chronic anterior compartment syndrome in the lower leg. Acta
Orthopaedica Scandinavica 58, 139-144.
Styf J, Komer LM & Suurkula M (1987). Intramuscular pressure and muscle blood-flow during exercise in
chronic compartment syndrome. Journal of Bone and Joint Surgery 69B, 301-305.
Supinski GS, DiMarco AF & Altose MD (1990). Effect of diaphragmatic contraction on intramuscular pressure
and vascular impedance. Journal of Applied Physiology 68,1486-1493.
Sylvest 0 & Hvid N (1959). Pressure measurements in human striated muscles during contraction. Acta
Rheumatologica Scandinavica 5, 216-222.
V0llestad NK, Wesche J & Sejersted OM (1990). Gradual increase in leg oxygen uptake during repeated
submaximal contractions in humans. Journal of Applied Physiology 68, 1150-1156.
WestgaardRH (1988). Measurement and evaluation of postural load in occupational work situations. European
Journal ofApplied Physiology and Occupational Physiology 57,291-304.
Wiig H (1985). Comparison of methods for measurement of interstitial fluid pressure in cat skin subcutis and
muscle. American Journal of Physiology 249, H929-H944.
Wiig H (1990). Evaluation of methodologies for measurement of interstitial fluid pressure (Pi): physiological
implications of recent Pi data. Critical Reviews in Biomedical Engineering 18, 27-54.
 26

TASK-DEPENDENT NATURE OF FATIGUE IN

SINGLE MOTOR UNITS

B. R. Botterman

Department of Cell Biology and Neuroscience

University of Texas Southwestern Medical Center
Dallas, Texas 75235

ABSTRACT

The loss of force production during sustained activity presents the eNS a unique control
problem. Different tasks stress the neuromuscular system at different sites and times, and
involve different cellular mechanisms. The functional organization of muscles and their
motor units has evolved to avoid fatigue processes that impair motor performance. The
purpose of this brief review is to examine the fatigue properties of type-identified motor
units and to speculate what these properties reveal about the organization and control of
muscle.

INTRODUCTION

Most muscles contain three distinct types of motor units (slow twitch or type S; fast
twitch fatigue resistant or type FR; and fast twitch, fatigable or type FF), each having
different susceptibilities to fatigue. Within each motor unit type, the range of fatigabilities
can also vary greatly, depending on the task that produces the fatigue. Prolonged activation
ofa motor unit, regardless of type, leads to a decline in force production. As Bigland-Ritchie
and colleagues have stressed (Bigland-Ritchie & Woods, 1984; Bigland-Ritchie et aI., 1986),
the fatigue process begins immediately after the onset of activity. The site of fatigue might
be due to failure anywhere along the path that ultimately results in force production, from
the descending command that activates a-motoneurons of a motor pool, to interaction of the
contractile proteins within single muscle fibers. During activity, whether it is brief or
sustained, each component of the motor unit (motor axon, neuromuscular junction, muscle
fibers) undergoes fatigue-related changes. These changes can be subtle or dramatic. The
component which fails depends on the task that the motor unit has to perform. An important
feature of motor units is how well their components are matched with respect to their fatigue
resistance and their recovery from fatigue. Fatigue-resistant units show both electrical and
mechanical endurance, while highly fatigable units show a rapid decline in electromyog-
raphic (EMG) activity and force after brief activation. This feature has important conse-

351
352 B. R. Botterman

quences for muscle control. Different motor unit types fail for different reasons. Because of
their different susceptibilities to fatigue, it may not be too surprising to find that the
predominant site and mechanism for fatigue may differ between active motor units, depend-
ing on the task (for review, see Enoka & Stuart, 1992). On the other hand, given the
appropriate stimulation parameters, it is likely that each type of motor unit can be made to
fail at the same site(s).
What advantage is there to studying single, isolated motor units during fatigue? After
all, a motor unit can be viewed as a very small muscle containing a fairly homogenous set of
muscle fibers when compared to muscle fibers of other motor units within a muscle. In human
studies, motor unit performance during fatiguing contractions has been largely inferred from
the changing EMG pattern of the whole muscle or from the discharge patterns of single motor
units. The advantage of the former approach is that the performance of the muscle can be
evaluated for the task chosen by the investigator, but has the disadvantage that the contribution
of individual motor units to the fatigue process cannot be evaluated. Using the latter approach,
single motor unit studies in humans have provided important insights into how muscle is
controlled during fatiguing contractions. For example, Bigland-Ritchie and colleagues
(Bigland-Ritchie et aI., 1983a; Bigland-Ritchie & Woods, 1984; Gandevia et aI., 1990) have
shown that the firing rates of single motor units decrease during maximal voluntary contrac-
tions. The decrease in firing rates is matched to the slowing of contractile speed associated with
fatiguing contractions, thereby avoiding unnecessarily high rates that could hasten fatigue.
However, these types of studies also have their limitations. It is not possible to control easily
the activation history of individual units, and there is always the uncertainty of which types of
units are active, the force that they produce, and the degree offatigue that they show during the
contraction. Single motor unit studies in animals overcome many of these limitations because
motor units can be identified as to type and studied in isolation. By using intramuscular or
intraneural microstimulation techniques, human motor units can also be isolated and their
contractile properties recorded with an acceptable degree of certainty (Garnett et aI., 1979;
Westling et aI., 1990). However, while isolated motor units can be subjected to a multitude of
stimulation protocols, their environment differs from that in human studies, where many motor
units are active simultaneously. Under these conditions, changes in the surrounding extracellu-
lar fluid may hasten fatigue of newly recruited motor units (Westerblad et aI., 1991; see
McComas & Sj0gaard, Chapter 4). The degree to which extracellular metabolites contribute to
the development of fatigue in different motor unit types is unknown.
The purpose of this chapter is to examine briefly the fatigue properties oftype-iden-
tified motor units and to speculate what these properties reveal about the organization and
control of muscle. Moreover, the tasks used to delineate these fatigue properties will be
considered, since the extent and site of motor unit fatigue depend on the task.

MOTOR UNIT TYPE AND FATIGUE

Muscle fatigue has been evaluated under a wide range of experimental protocols. In
most human studies, the output level of the muscle is voluntarily held at maximum or some
percentage of maximum for various lengths of time. By contrast, in most animal studies
investigators have used constant or intermittent stimulation, usually of fixed frequency, to
produce contractions of various duration. Fatigue is usually measured as a decline in force
production relative to some predetermined rested state of the muscle. The muscle is usually
stimulated under isometric conditions, and held at or near optimal length for force production
during the fatiguing contraction. The type of stimulus protocol or task used to induce fatigue
is often intended to produce significantly impaired function in only one or two processes
involved in force production.
Task-Dependent Nature of Fatigue in Single Motor Units 353

The most widely used fatigue test for single motor units is that of Burke and
colleagues (Burke et ai., 1973). Because motor unit type provided insight into the functional
organization of the cat medial gastrocnemius (MG) muscle (Burke, 1981), the test has been
used as a benchmark for comparison of motor units in other muscles and across species.
As applied to cat MG motor units, the test separated units into two groups. Motor
units were studied under isometric conditions at optimal length for tension development and
were stimulated intermittently (40 Hz for 330 ms of every second) for 2-4 min. The
intermittent stimulation protocol had a minimal effect on the amplitude of the EMG, so
fatigue was thought to be primarily restricted to the muscle fibers themselves. However,
more recent work has shown that EMG waveform reductions in amplitude and area do occur
in the more fatigable units (e.g., Sandercock et ai., 1985; Gardiner & Olha, 1987; Enoka et
ai., 1992). The test essentially divided fast-twitch (type F) motor units into two, non-over-
lapping groups (type FR and FF units). As applied to other limb muscles in cat and other
species (e.g., Botterman et ai., 1985; Gardiner & Olha, 1987; cf. Kugelberg & Lindegren,
1979; Thomas et ai., 1991a), the test produced a more continuous, albeit bimodal, distribu-
tion among fast-twitch units. While many FR units show only modest amounts of fatigue
after 4 min of stimulation (Burke & Tsairis, 1974), they are much more fatigable as a group
than S units. The test does not reveal the relative fatigabilities ofS units (Enoka et ai., 1992).
Another method of testing for the relative fatigability of motor units is by maintaining
their output at some fixed percentage of their maximal force, similar to what is done when
human subjects are asked to maintain force at a percentage of their maximal voluntary
contraction. Motor unit force is maintained or clamped at a desired level through computer
feedback control, by altering the stimulation rate of the motor axon (Fig. 1). When force can
no longer be maintained by increasing stimulation rate, the unit's endurance time is reached.
In the cat MG muscle, the endurance time of F units studied at an output level of 25% of
maximal tension ranged between 20 and 2000 s. In contrast to the Burke fatigue test, where
fatigue index had a bimodal distribution, endurance time was distributed continuously. If a
Burke fatigue test is given 15 min later, the expected percentages ofFR and FF units in MG
are still found (cf. Clamann & Robinson, 1985). So although the absolute force is diminished
following the first fatiguing contraction (force-clamp), units can still be separated into two
groups, one representing FR units and the other group consisting ofF units that have fatigue

A B
100 100
,-,
~N 80
'-'::r: 80
I'i'-'
.9 .s
'"
'" ...
I'i
2 I'i
60 60
e ..8 40 40
.~ ] 20
::l '"

'"
:::EU).- 20
0 0
0 100 200 300 400 0 100 200 300 400
Time (s) Time (s)
Figure 1. Typical force-clamp contractions of a FR unit in cat medial gastrocnemius (A) and an S unit in cat
soleus (S). Tension was maintained by altering the stimulation rate applied to a unit's axon through computer
feedback control. The endurance time of a unit was reached when target tension could no longer be maintained
with a stimulation rate ~ 100 Hz. Target levels were 25 and 85% of maximal tetanic tension of the FR and S
units, respectively. In both A and S, the thicker traces represent the averaged tension maintained by feedback
control, the thinner traces the instantaneous stimulation rate.
354 B. R. Botterman

indices below 0.25. In this latter group, some FR and FI units are undoubtedly classified as
FF because they were subjected to two fatigue tests.
Unlike F units, the endurance times of S units could not be established at an output
level of25% of maximum because of practical reasons; their endurance times exceeded 3000
s. When S units (MG and soleus muscles of the cat) were clamped at 85% of maximum (Fig.
1), endurance times were found between 14 and 800 s, roughly half the range found for F
units at 25% of maximum. As with F units, the distribution of endurance times was
continuous. The stimulation rates needed to maintain force at 85% of maximum was
generally below 35 Hz for S units, in comparison to rates less than 30 Hz for F units held at
25% of maximum. At both output levels, only during the last 5% of the fatiguing contraction
did stimulation rates increase dramatically in order to maintain force. Also during this period,
EMG waveforms were found to change rapidly, substantially broadening in width and
decreasing in peak amplitude (Botterman & Cope, 1988a; Cope et aI., 1991). It appears that
endurance time mainly reflects fatigue processes beyond sarcolemma excitability.
Although the exact cause for fatigue during either the Burke fatigue test or force-
clamp contractions is not known, the electrical and mechanical failure observed during each
test is clearly related to motor unit type. A number of studies have noted the relationship
between motor unit and susceptibility to electrical failure (EMG decline) during various
types of fatiguing contractions (e.g., Kugelberg & Lindegren, 1979; Clamann & Robinson,
1985; Sandercock et aI., 1985; Gardiner & Olha, 1987; see also Enoka et aI., 1992).
Recovery of tension after a fatiguing contraction, whether it is induced by high- or
low- frequency stimulation and delivered continuously or intermittently, is also related to
motor unit type. Fatigue produced by intense, high frequency (80 Hz) stimulation affects
both Sand F units, but more so type F units. This type of fatigue is thought to involve failure
in electrical excitation during the contraction (Clamann & Robinson, 1985; Sandercock et
aI., 1985), and both tension and EMG recover quickly after the contraction. In contrast, less
intense activity, such as continuous (10 Hz) or intermittent (40 Hz) stimulation lasting for
several minutes, produces a delayed fatigue that develops 30-60 min after the initial fatiguing
contraction (Jami et aI., 1983; Sandercock et aI., 1985). Even modest amounts of activity
(several 10-s trains at 20 Hz) reduce tension significantly (Powers & Binder, 1991). This
type of delayed fatigue or low-frequency fatigue (Edwards et aI., 1977) is most evident when
motor units are stimulated at lower rates (<40 Hz) and can be partly overcome with sustained
activity at these rates (e.g., Jami et aI., 1983). Low-frequency fatigue is most often attributed
to disturbances in excitation-contraction coupling, since EMG waveform characteristics
return to normal within several minutes (Edwards et aI., 1977; Sandercock et aI., 1985;
Powers & Binder, 1991; cf. Metzger & Fitts, 1987). It occurs with more regularity in FF than
FR units (Jami et aI., 1983; Sandercock et aI., 1985; Powers & Binder, 1991), while there is
very little evidence for it in S motor units (Sandercock et aI., 1985).

RATE MODULATION DURING FATIGUING CONTRACTIONS

A constant feature of sustained, maximal voluntary contractions in humans is the

decrease in motor unit firing rate that occurs during the contraction. Even during sustained
submaximal contractions, where the firing rates of active units may be expected to rise (e.g.,
Maton & Gamet, 1989), rates often either remain the same (Bigland-Ritchie et aI., 1986) or
decrease (Garland et aI., 1994). To compensate for fatigue during these types of contractions,
additional motor units are recruited to maintain target force. It has been proposed by
Bigland-Ritchie and Woods (1984) that the CNS adjusts the firing rates of motor units so
that the lowest possible rate produces the maximum required force. By optimizing force
production in this way, premature fatigue of the muscle may be avoided. The reduction in
Task-Dependent Nature of Fatigue in Single Motor Units 355

motoneuron discharge rates under inhibitory reflex control from the exercising muscle
(Woods et aI., 1987) occurs in the presence of increased excitatory drive to the motor pool.
An important issue is, to what extent do the contractile properties of motor units contribute
to this seemingly paradoxical decrease in their firing rates, while other motor units are being
recruited during the fatiguing contraction?
The relationship between the firing rate of a motor unit and its force output is strongly
influenced by its prior activation history. Without knowledge of the pattern and duration of
the preceding activity, predicting whether a given firing rate will produce an increase, a
decrease or no change in force output is not possible (Binder-Macleod & Clamann, 1989).
The force-firing rate relationship is influenced by two opposing processes; one that enhances
force (potentiation), while the other one diminishes force (fatigue). The mechanisms under-
lying these two opposing processes are different (Krarup, 1981; Gordon et aI., 1990b; Powers
& Binder, 1991), and their influence on force production can last for several seconds to
several hours. Another important influence on force production is the contractile slowing
that accompanies fatiguing contractions (Burke et ai., 1973; Bigland-Ritchie et ai., 1983b;
Dubose et ai., 1987; Gordon et aI., 1990a; Fitts, 1994).
The two mechanisms that promote decreased firing rates, potentiation and contractile
slowing, often occur simultaneously and shift the force-frequency curve to the left (e.g., Powers
& Binder, 1991; Thomas et aI., 1991 b). The presence of potentiation, a predominate factor early
in sustained contractions, can still be found after fatigue processes have overwhelmed most of
its effects (Kernell et aI., 1975; Jami et ai., 1983; Gordon et aI., 1990b). Potentiation occurs
chiefly in F units (FR > FF) and, depending on how its presence is measured, enhances force
more in FR than FF units (Jami et aI., 1983; Powers & Binder, 1991; cf. Gordon et aI., 1990b).
The susceptibility of motor units to contractile slowing, like potentiation, is also related to their
fatigability. During the first 30 s of the Burke fatigue test, Clamann and colleagues (Dubose et
ai., 1987) have observed significant increases in contraction and half-relaxation times for FF
units, whereas FR units responded with minor increases in half-relaxation times. In response to
the initial 4-5 trains, S units actually sped up and then remained fairly constant for the remainder
of the testing period. Also using the Burke fatigue test, Gordon and colleagues (1990a) found
no change in the rate of tetanic force development and relaxation for Sand FR units, while
relaxation increased for FF units. However, during long, sustained force-clamp contractions
(Fig. 1), FR units, and to a lesser extent S units, show evidence of contractile slowing
(Botterman and Cope, unpublished observations).
Another mechanism that may contribute to lower firing rates is the marked hysteresis
observed for the force-frequency relationship (Binder-Macleod & Clamann, 1989). When
the force output of a motor unit is rapidly increased with a brief high-frequency stimulus
train, higher force levels can subsequently be maintained with dramatically lower rates. This
effect was seen for all unit types. As Binder-Macleod and Clamann (1989) have pointed out,
this high to low strategy is an excellent way to generate maximal force at a given stimulation
frequency. This strategy also enhances force during a fatigue test similar to that of Burke
and colleagues (Bevan et aI., 1992). Stimulus trains that contained two brief interpulse
intervals (10 ms), followed by constant rate stimulation at a much lower rate (e.g., 25 Hz),
produced more force than constant-rate trains at the same lower rate over the course of the
test (360 s). Over a few seconds, optimized activation patterns are a means of delaying
fatigue. However, once this high to low strategy is invoked and the optimal rate is achieved,
the force produced by that rate can only be maintained by that same rate or higher rates as
the motor unit becomes more fatigued (e.g., Fig. 1).
Many of the features seen in the discharge behavior of human motor units during
sustained submaximal contractions (either a decrease or no change in firing rate) can also
be seen in the stimulation patterns of cat motor units studied under force-clamp conditions.
At an output level of 25% of maximum, F units showed an average reduction of 50% in
356 B. R. Botterman

stimulation rate during the first part ofthe contraction. In the more fatigue-resistant FR units,
the continuous decline in stimulation rate can last in excess of 100 s (Fig. 1). This initial
decline in stimulation rate was seen for all F units studied, and likely is in response to the
combined processes of potentiation and contractile slowing during this phase of the contrac-
tion. On the other hand, S units clamped at either 25, 70 or 85% of maximum rarely showed
decreases in stimulation rate and, for long periods, remained fairly constant (Fig. I;
Botterman & Cope, 1988a; Cope et ai., 1991; Botterman et ai., 1992). Therefore, for the
units (S and FR) most likely to be recruited during submaximal contractions, their firing
rates can remain the same (S) or must decrease (FR) if their force output is to remain constant.
Because of the adaptive properties of motoneurons (Kernell & Monster, 1982), only small
changes in the synaptic drive to Sand FR motor units would be needed to maintain their
force at a constant level. It is assumed, of course, that the rates used to maintain force were
comparable to those seen during submaximal contractions in humans, which may be the case
when differences in the contractile speed of cat and human motor units are taken into account
(Kernell et ai., 1983; Botterman et ai., 1986; Thomas et ai., 1991b).
The force-frequency relationship is strongly influenced by fatigue processes that reduce
force production. Fatigued motor units require higher stimulation rates to produce the same
force as they did before they were fatigued (Kernell et aI., 1975; Thomas et ai., 1991 b). In an
attempt to gain more information about the changes in the force-frequency relationship during
fatiguing contractions, Botterman and colleagues (1992) varied the force of cat soleus (type S)
units between three levels (50-70-90% of maximum), resulting in an averaged output of70%
of maximum (Fig. 2). Under computer feedback control, stimulation rate was altered to achieve
the desired output levei. During the contraction, the stimulation rate needed to maintain 50 and
70% of maximum remained fairly constant (-7 and 10Hz, respectively) or modestly increased
throughout the contraction, while the rate needed to maintain 90% of maximum steadily
increased from -30 Hz to 60 Hz, after which it increased rapidly to the experimentally set upper
limit of 100 Hz. This form of high-frequency fatigue was found for all units. However, in 40%
of the units the rate needed to maintain tension at 90% of maximum actually decreased from
the initial 90% step of the contraction, between 4 and 49%, before increasing during the later
stages of the contraction. Whether this phenomenon operating at such a high output level is due
to potentiation or contractile slowing or another mechanism is not known. It does, however,

A B
100 100
*
80 =:::::::~::_:_IT----::----~
60 ~D==::::::~TI
40

20~
o
o 6 12 18 24 30 40 s
Time (min)
Figure 2. Typical step-clamp contraction of a S unit in cat soleus. Target levels were 50-70-90% of maximal
tetanic tension of the unit, with a step duration of 40 s. Asterisk indicates one step cycle, which is shown in B
on an expanded time scale. B, dashed lines indicate the three target levels. In both A and B, the upper traces
represent the averaged tension maintained by feedback control, the lower traces the instantaneous stimulation
rate.
Task-Dependent Nature of Fatigue in Single Motor Units 357

probably provide another example of the need to reduce motoneuron firing rate to match the
changing contractile properties of the muscle unit.

FORCE-FATIGABILITY RELATIONSHIP IN SINGLE MOTOR

UNITS

Two important force-fatigability relationships can be identified in muscle, one that

describes the generalized relationship for any force producing element, whether it is a muscle,
motor unit or single muscle fiber, and the other that describes the relation of a motor unit's force
capacity and resistance to fatigue relative to other motor units within a muscle. The first
relationship can be intuitively appreciated: the greater the force exerted or work produced, the
more rapid the fatigue. The second relationship, also an inverse one, reveals that the more force
capacity a motor unit has, the more fatigable it is (see Thomas, Chapter 10).
That muscle endurance depends heavily on the absolute force produced by a muscle
can be shown by comparing sustained isometric contractions at different output levels, or
by varying the duty cycle (i.e., changing the amount of work) of intermittent isometric
contractions (reviewed by Enoka & Stuart, 1992). Even when muscles are fully activated at
different muscle lengths, endurance is found to be lower at the length that yields the most
absolute force (e.g., Fitch & McComas, 1985). While different tasks can alter the form of
the generalized force-fatigability relationship, there appears to be no structure or mechanism
that when stressed invalidates its inverse nature. As Enoka and Stuart (1992) have stated,
the underlying mechanisms responsible for fatigue interact in such a way as to scale with
force.
Many of the important observations on the generalized force-fatigability relationship
have been made using whole muscle contractions. Some of these observations have recently
been confirmed for cat soleus motor units studied under force-clamp conditions. Isometric
contractions of constant tension, clamped at 70 and 85% of maximum, revealed that units
studied at the lower level had on average significantly longer endurance times than units at
the higher level (1449 and 272 s, respectively). The mean force-time integral (proportional
to work done) was also greater for units studied at the lower compared to the higher level
by a factor of 4.38 to 1. Moreover, when contractions of constant force at 70% of maximum
were compared with those that varied between 50-70-90% of maximum, constant-force
endurance times were greater by nearly a factor of2 (1449 VS. 741 s). Although both types
of contractions had the same averaged output (70% of maximum), the energy cost appeared
to be much higher to vary force than to maintain it at a fixed level, in agreement with earlier
studies on single muscle fibers and whole muscles (Dawson et aI., 1978; Loiselle &
Walmsley, 1982; Duchateau & Hainaut, 1985; Bergstrom & Hultman, 1988).
While the relationship between the force capability (i.e., maximal tetanic tension) of
a motor unit and its fatigability has been appreciated for many years (Henneman & Olson,
1965; Burke et aI., 1973), the precision of the relationship has only recently been systemati-
cally studied (Botterman & Cope, 1988b; Cope et aI., 1991; Botterman et aI., 1992). For F
motor units in individual limb muscles of the cat, there is a remarkably high correlation
(Spearman rank correlation coefficent > -0.90) between tetanic tension and the endurance
time of contractions clamped at 25% of maximum (Botterman & Cope, 1988b). Kugelberg
and Lindegren (1979) also found a high correlation between unit tension and fatigability for
F units of rat tibialis anterior muscle. In their study, fatigue was produced by stimulating
units 20 times at 100 Hz twice a second for 4 min. In contrast, the strong tetanic tension-en-
durance time relationship seen for F units was not present for S units of the cat soleus when
their force outputs were clamped at 70 or 85% of maximum (Botterman et aI., 1992).
358 B. R. Botterman

However, when force output was varied between three levels (Fig. 2), yielding an averaged
output of70% of maximum, a significant correlation between tetanic tension and endurance
time was found. Thus, it appears that the tetanic tension-endurance time relationship has a
task-dependent component to it. Whether this explanation is also valid for the observations
of Nordstrom and Miles (1990), who found no correlation between twitch tension and
fatigability in human masseter, is unclear. Fatigue resistance was evaluated by comparing a
unit's twitch amplitude before and after 15 min of continuous activation at 10Hz. The lack
of a relationship may be due to the masseter's unusual physiological and histochemical
features (Nordstrom & Miles, 1990; see also Miles & Nordstrom, Chapter 31). Further
studies are clearly needed to clarify the effects that various tasks have on the force-fatigability
relationship, as well as the influence of motor unit type, muscle, and species on the
relationship.

CONCLUDING REMARKS

One of the more important tasks that the CNS performs is to engage the motor units
ofa muscle in a way that reduces or avoids fatigue of the muscle. From the pioneering work
of Henneman and colleagues (Henneman et aI., 1965a,b; Henneman & Olson, 1965) came
the simplifying principle that motoneurons of a pool are recruited in a fixed order, and the
speed, forcefulness and endurance of the muscle units that they supply are tightly coupled.
Since most muscles contain motor units that have widely different susceptibilities to fatigue
(Burke etaI., 1973; Botterman & Cope, 1988a,b; Cope etaI., 1991), the order in which motor
units are recruited and the pattern of their activation probably strongly influences the time
course of muscle fatigue (e.g., Botterman & Cope, 1988b; Bevan et aI., 1992). While it is
generally accepted that motor units of a muscle are recruited in order of increasing fatigabil-
ity and forcefulness, few studies have actually tested for the relationship directly (e.g.,
Stephens & Usherwood, 1977; Zajac & Faden, 1985; Nordstrom & Miles, 1990; Yee et aI.,
1990; Cope & Clark, 1991). Moreover, the degree to which recruitment order and fatigability
are related remains uncertain. In animal studies, investigators have used the Burke fatigue
test to evaluate fatigue resistance, which is of limited utility in determining the relative
fatigability of the more fatigue resistant units in a muscle. Human studies using traditional
recording approaches have their limitations as well, where the activation history of recruited
units is difficult to control and, by necessity, units are often restricted to the low-threshold
range. Human studies that characterize the discharge properties of single motor units during
voluntary contractions and subsequently record their contractile properties by stimulating
their axons within the nerve fascicle may overcome many of these limitations (Westling et
aI., 1990; Thomas et aI., 1991a,b). Because of the powerful effect that fatigue has on the
threshold and discharge behavior of motoneurons (Bigland-Ritchie et aI., 1986; Enoka et aI.,
1989), future animal and human studies are needed that examine the activity of well-char-
acterized single motor units to various fatiguing tasks. By doing so, it may be possible to
decipher the strategies that the CNS uses to manage muscle fatigue.

ACKNOWLEDGMENTS

I wish to thank my collaborators, Dr. Timothy Cope, Dr. Keith Tansey, Larry Graf
and Andy Yee, for their invaluable contributions to the force-clamp studies. The work was
supported by United States Public Health Service grant NS 17683. Attendance ofB.R.B at
the 1994 Bigland-Ritchie conference was supported, in part, by NS 17683.
Task-Dependent Nature of Fatigue in Single Motor Units 359

REFERENCES
Bergstrom M & Hultman E (1988). Energy cost and fatigue during intermittent electrical stimulation of human
skeletal muscle. Journal ofApplied Physiology 65, 1500-1505.
Bevan L, Laouris Y, Reinking RM & Stuart DG (1992). The effect of the stimulation pattern on the fatigue of
single motor units in adult cats. Journal of Physiology (London) 449, 85-108.
Bigland-Ritchie B, Cafarelli E & V",llestad NK (1986). Fatigue of submaximal static contractions. Acta
Physiologica Scandinavica Supplementum 556, 137-148.
Bigland-Ritchie B, Johansson R, Lippold OCJ, Smith S & Woods JJ (1983a). Changes in motoneurone firing
rates during sustained maximal voluntary contractions. Journal ofPhysiology (London) 340, 335-346.
Bigland-Ritchie B, Johansson R, Lippold OCJ & Woods JJ (1983b). Contractile speed and EMG changes
during fatigue of sustained maximal voluntary contractions. Journal ofNeurophysiology 50, 313-324.
Bigland-Ritchie B & Woods JJ (1984). Changes in muscle contractile properties and neural control during
human muscular fatigue. Muscle & Nerve 7,691-699.
Binder-Macleod SA & Clamann HP (1989). Force output of cat motor units stimulated with trains of linearly
varying frequency. Journal of Neurophysiology 61, 208-217.
Botterman BR & Cope TC (1988a). Motor-unit stimulation patterns during fatiguing contractions of constant
tension. Journal ofNeurophysiology 60, 1198-1214.
Botterman BR & Cope TC (1988b). Maximum tension predicts relative endurance of fast-twitch motor units
in the cat. Journal of Neurophysiology 60,1215-1226.
Botterman BR, GrafLB & Tansey KE (1992). Fatigability of cat soleus motor units activated at varying force
levels. Society for Neuroscience Abstracts 18, 1556.
Botterman BR, Iwamoto GA & Gonyea WJ (1985). Classification of motor units in flexor carpi radialis muscle
of the cat. Journal ofNeurophysiology 54,676-690.
Botterman BR, Iwamoto GA & Gonyea WJ (1986). Gradation of isometric tension by different activation rates
on motor units of cat flexor carpi radialis muscle. Journal of Neurophysiology 56, 494-506.
Burke RE (1981). Motor units: anatomy, physiology, and functional organization. In: Brookhart JM,
Mountcastle VB (eds.), Brooks VB (vol. ed.), Handbook of Physiology, sec. 1, vol. l/, pt. 1, The
Nervous System: Motor Control, pp. 345-422. Bethesda, MD: American Physiological Society.
Burke RE, Levine DN, Tsairis P & Zajac FE (1973). Physiological types and histochemical profiles of motor
units of cat gastrocnemius. Journal of Physiology (London) 234, 723-748.
Burke RE & Tsairis P (1974) The correlation of physiological properties with histochemical characteristics in
single muscle units. Annals New York Academy of Sciences 228, 145-158.
Clamann HP & Robinson AJ (1985). A comparison of electromyographic and mechanical fatigue properties
in motor units of the cat hindlimb. Brain Research 327, 203-219.
Cope TC & Clark BD (1991). Motor unit recruitment in the decerebrate cat: several unit properties are equally
good predictors of order. Journal of Neurophysiology 66, 1127-1138.
Cope TC, Webb CB, Yee AK & Botterman BR (1991). Nonuniform fatigue characteristics of slow-twitch motor
units activated at a fixed percentage of their maximum tetanic tension. Journal of Neurophysiology
66, 1483-1492.
Dawson MJ, Gadian DG & Wilkie DR (1978). Muscular fatigue investigated by phosphorus nuclear magnetic
resonance. Nature (London) 274, 861-866.
Dubose L, Schelhorn TB & Clamann HP (1987). Changes in contractile speed of cat motor units during activity.
Muscle & Nerve 10, 744-752.
Duchateau J & Hainaut K (1985). Electrical and mechanical failures during sustained and intermittent
contractions in humans. Journal ofApplied Physiology 58,942-947.
Edwards RHT, Hill DK, Jones DA & Merton PA (1977). Fatigue of long duration in human skeletal muscle
after exercise. Journal ofPhysiology (London) 272, 769-778.
Enoka RM, Robinson GA & Kossev AR (1989). Task and fatigue effects on low-threshold motor units in human
hand muscle. Journal of Neurophysiology 62,1344-1359.
Enoka RM & Stuart DG (1992). Neurobiology of muscle fatigue. Journal of Applied Physiology 72,
1631-1648.
Enoka RM, Trayanova N, Laouris Y, Bevan L, Reinking RM & Stuart DG (1992). Fatigue-related changes in
motor unit action potentials of adult cats. Muscle & Nerve 14, 138-150.
Fitch S & McComas A (1985). Influence of human muscle length on fatigue. Journal ofPhysiology (London)
362,205-213.
Fitts RH (1994). Cellular mechanisms of muscle fatigue. Physiological Reviews 74, 49-94.
Gandevia SC, Macefield G, Burke D & McKenzie DK (1990). Voluntary activation of human motor axons in
the absence of muscle afferent feedback. The control of the deafferented hand. Brain 113, 1563-1581.
360 B. R. Botterman

Gardiner PF & Olha AE (1987). Contractile and electromyographic characteristics of rat plantaris motor unit
types during fatigue in situ. Journal of Physiology (London) 385, 13-34.
Garland SJ, Enoka RM, Serrano LP & Robinson GA (1994). Behavior of motor units in human biceps brachii
during a submaximal fatiguing contraction. Journal ofApplied Physiology 76, 2411-2419.
Garnett RAF, O'Donovan MJ, Stephens JA, Taylor A (1979). Motor unit organization of human medial
gastrocnemius. Journal of Physiology (London) 287, 33-43.
Gordon DA, Enoka RM, Karst GM & Stuart DG (1990a). Force development and relaxation in single motor
units of adult cats during a standard fatigue test. Journal of Physiology (London) 421, 583-594.
Gordon DA, Enoka RM & Stuart DG (1990b). Motor-unit force potentiation in adult cats during a standard
fatigue test. Journal of Physiology (London) 421, 569-582.
Henneman E & Olson CB (1965). Relations between structure and function in the design of skeletal muscles.
Journal of Neurophysiology 28, 581-598.
Henneman E, Somjen G & Carpenter DO (1965a). Functional Significance of cell size in spinal motoneurons.
Journal of Neurophysiology 28, 560-580.
Henneman E, Somjen G & Carpenter DO (1965b). Excitability and inhibitability ofmotoneurons of different
sizes. Journal ofNeurophysiology 28,599-620.
Jami L, Murthy KSK, Petit J & Zytnicki D (1983). After-effects of repetitive stimulation at low frequency on
fast-contracting motor units of cat muscle. Journal of Physiology (London) 340, 129-143.
Kernell D, Ducati A & Sjoholm H (1975). Properties of motor units in the first deep lumbrical muscle of the
cat's foot. Brain Research 98, 37-55.
Kernell D, Eerbeek 0 & Verhey BA (1983). Relation between isometric force and stimulation rate in cat's
hindlimb motor units of different twitch contraction time. Experimental Brain Research 50, 220-237.
Kernell D & Monster AW (1982). Motoneurone properties and motor fatigue. An intracellular study of
gastrocnemius motoneurones of the cat. Experimental Brain Research 46, 197-204.
Krarup C (1981). Enhancement and diminution of mechanical tension evoked by staircase and by tetanus in
rat muscle. Journal of Physiology (London) 311, 355-372.
Kugelberg E & Lindegren B (1979). Transmission and contraction fatigue of rat motor units in relation to
succinate dehydrogenase activity of motor unit fibres. Journal ofPhysiology (London) 288, 285-300.
Loiselle DS & Walmsley B (1982). Cost of force development as a function of stimulus rate in rat soleus
muscle. American Journal of Physiology 243, C242-C246.
Maton B & Gamet D (1989). The fatigability of two agonistic muscles in human isometric voluntary
submaximal contraction: and EMG study. II. Motor unit firing rate and recruitment. European Journal
ofApplied Physiology & Occupational Physiology 58, 369-374.
Metzger 1M & Fitts RH (1987). Fatigue from high- and low-frequency muscle stimulation: contractile and
biochemical alterations. Journal ofApplied Physiology 62, 2075-2082.
Nordstrom MA & Miles TS (1990). Fatigue of single motor units in human masseter. Journal of Applied
Physiology 68, 26-34.
Powers RK & Binder MD (1991). Effects oflow-frequency stimulation of the tension-frequency relations of
fast-twitch motor units in the cat. Journal ofNeurophysiology 66, 905-918.
Sandercock TG, Faulkner JA, Albers JW & Abbrecht PH (1985). Single motor unit and fiber action potentials
during fatigue. Journal Applied Physiology 58, 1073-1079.
Stephens JA & Usherwood TP (1977). The mechanical properties of human motor units with special reference
to their fatiguability and recruitment threshold. Brain Research 125, 91-97.
Thomas CK, Bigland-Ritchie B & Johansson RS (1991b). Force-frequency relationships of human thenar
motor units. Journal ofNeurophysiology 65, 1509-1516.
Thomas CK, Johansson RS & Bigland-Ritchie B (1991a). Attempts to physiologically classify human thenar
motor units. Journal ofNeurophysiology 65, 1501-1508.
Westerblad H, Lee JA, Liinnergren J & Allen DG (1991). Cellular mechanisms of fatigue in skeletal muscle.
American Journal of Physiology 261, CI95-C209.
Westling G, Johansson RS, Thomas CK & Bigland-Ritchie B (1990). Measurement of contractile and electrical
properties of single human thenar motor units in response to intraneural motor-axon stimulation.
Journal of Neurophysiology 64,1331-1346.
Woods JJ, Furbush F and Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal ofNeurophysiology 58, 125-137.
Yee AI<, Tansey KE & Botterman BR (1990). Relative endurance and recruitment order among pairs offast-twitch
motor units in the cat medial gastrocnemius muscle. Society for Neuroscience Abstracts 16, 888.
Zajac FE & Faden JS (1985). Relationship among recruitment order, axonal conduction velocity, and
muscle-unit properties of type-identified motor units in the cat plantaris muscle. Journal of Neuro-
physiology 53, 1303-1322.
 27

TASK-DEPENDENT FACTORS IN FATIGUE

OF HUMAN VOLUNTARY CONTRACTIONS

B. Bigland-Ritchie,l C. L. Rice,2 S. J. Garland,3 and M. L. Walsh l

1 Department of Pediatrics, Yale University School of Medicine

New Haven, Connecticut 06520
2School of Physical and Occupational Therapy and Department of Anatomy,
McGill University
Montreal, Quebec, Canada
3 Department of Physical Therapy, University of Western Ontario
London, Ontario, Canada

ABSTRACT

This chapter explores the hypothesis that fatigue is not caused uniquely by any common set
of factors, but rather that the amount of stress placed on each site depends on the type of
exercise from which fatigue develops. Evidence supporting this idea is presented by com-
paring results from various studies in which fatigue was caused by different exercise
protocols. However, the way in which human endurance capacity changes with the type or
intensity of the task performed suggest a unitary process. Thus, perhaps the neuromuscular
system as a whole is so well adjusted that any task-related additional impairment at one site
is compensated by corresponding functional improvements at others. We suggest that nature
has had a long time in which to "get it right".

INTRODUCTION

Other chapters in this book provide detailed descriptions of functional changes at

each of the various sites within the neuromuscular system that contribute to fatigue. Each
process is treated individually. Our aim is to provide an overview of how each of these events
may be influenced by the different types of exercise used to generate fatigue and how all
these events combine. Our hypothesis is that the amount of stress placed on each site depends
on the type of exercise, or task, from which fatigue develops. To address this question we
compare results from different studies, deliberately selecting those that yield contradictory
results. In fact, many of the inconsistent findings appear to be related to differences in the
particular exercise protocols used to generate fatigue.
Fatigue is defined here as any reduction in a person's ability to exert force or power
in response to voluntary effort, regardless of whether or not the task itself can still be
performed successfully (Bigland-Ritchie & Woods, 1984; cf. Edwards, 1981; Enoka & Stuart

361
362 B. Bigland-Ritchie et al.

1992). Hence, this definition includes both central and peripheral factors. The term task,
used to compare task-related changes with fatigue, includes not only the performance of
different types of exercise by any given muscle group (isometric, dynamic, etc.), but also
similar types of exercise performance by various muscles that have different contractile
properties. We also compare muscle responses to stimulation with those in which fatigue is
caused by voluntary contractions.
To establish that fatigue is always caused by a unique process or combination of
processes requires that the time course of its changes match those of force or power output,
both during the period of exercise and also throughout recovery. A second criterion, which
is much less frequently addressed, is to establish that this match between force and the
process being considered also exists under all exercise conditions. In this article, we point
out some situations where these rules do not always apply. We have not tried to be
comprehensive and many of the examples we use to illustrate each point were selected only
because they occurred to us. In speculating as to how incompatible findings could perhaps
be explained, we recognize that our views may not all be correct. But if so, we hope they
may serve to stimulate new lines of research.
Different tasks are performed by using muscles in different ways (i.e., together in
groups or more rarely in isolation) to generate either high or low forces that may either be
sustained or repeated intermittently. In each contraction, the muscle may shorten or lengthen,
or its length may remain unchanged. Energy requirements differ between isometric and
dynamic contractions of comparable intensity. The rate at which muscular performance
declines is determined mainly by the degree to which substrate delivery or removal is
compromised by restriction of muscle blood flow. But performance must also decline,
regardless of what metabolic changes take place within the muscle, if the central nervous
system (eNS) fails to activate the muscle adequately. This can occur if there are changes in
central motor drive, or in reflex input to spinal motoneurons, or if impulse conduction at
peripheral sites becomes impaired. Many of these processes are extremely sensitive to the
stimulation rate at which a muscle is excited. Thus, the relative contribution of each process
to fatigue must depend on many factors, but particularly on the contraction intensity involved
in each task. These factors must all be considered when trying to account for force loss during
isometric contractions, but changes of muscle contractile speed must also be included when
reduced power-output is considered. Hence there are good reasons to suspect that fatigue is
caused by a combination of processes which occur at different sites, and the way these
contribute to the final outcome is likely to depend on the task.

-r
:J
Forebrain & Motor cortex-- -----
Effort & Motivation ~ t'
1..-----<... Brainstem __ - - - - -
~ :
Spinal Motoneurons
Neuromusc~ transmisssion .....___ Reflex control via
, muscle afferents
E/C coupling
~
Muscle metabolism -
~
Cross-bridge [ATP] turnover
H~
FORCE &lOR POWER OUTPUT
Figure 1. Sites of fatigue at which functional impairment may vary with the task.
Task-Dependent Factors in Fatigue of Human Voluntary Contractions 363

The human body functions remarkably well under a wide variety of exercise condi-
tions. In fact, the smooth transition of responses suggests that, when a task is changed, any
increased stress on one site may be compensated by relief at others, so that the overall
outcome appears to operate like a single process. This type of interaction between changes
at each site could allow the overall system to make subtle adjustments that may optimize
performance when different physical demands are placed upon it. At present, it is not clear
to what extent the overall performance a person can sustain is limited by additive interactions
between the various fatigue-related changes that occur simultaneously at different sites.
Alternatively, a change at anyone site may itself become rate-limiting, regardless of what
occurs elsewhere.
The various sites or processes we consider in this context are shown in Fig. 1. starting
with the biochemical events most immediately responsible for force generation.

CHANGES IN MUSCLE FUNCTION

Biochemical Events
Numerous studies have attempted to establish a close relation between changes in
the metabolic profile of muscles and force loss during fatigue, but inconsistent findings are
often evident when the data are compared. Important factors include: first, that force usually
declines in a near-linear manner, but metabolic changes are usually most rapid at the onset
of the exercise, their levels stabilizing thereafter. Second, the concentrations of many
substances sometimes fail to change in proportion to the force exerted, particularly for forces
above about 50% of the maximal voluntary contraction (MVC).
Figs. 2A and 2B show results adapted from Weiner and colleagues (1990) who
measured PCr and Pi changes by NMR from tibialis anterior during (A) sustained MVCs
each held for 4 min, and (B) intermittent MVCs continued for 20 min. The total force decline
was similar in each case, as were the overall profiles of PCr and Pi changes for each task,
despite the large differences in endurance times. Initially PCr fell rapidly, then leveled off
with a tendency to return toward resting values in the second half of each contraction
protocol. Thus, the relation between PCr and force loss is not a simple one. During sustained
MVCs, force loss was well correlated with pH changes, but after the first 2 min of intermittent
exercise pH remained stable while the force continued to fall.
These findings are compared in Fig. 2C with data from V I2Illestad and colleagues
(1988) in which fatigue was caused by intermittent isometric quadriceps contractions of
low-intensity (30% MVC). Every 5 min, the subject made a brief 3 s MVC to test changes
in the maximal force-generating capacity. This declined linearly throughout the exercise
period by almost 50% after 30 min of exercise. However, biopsy samples taken after 5, 15,
and 30 min showed that the force loss was not accompanied by significant changes in ATP
or PCr. Muscle glycogen was minimally depleted, and lactate levels remained extremely low.
In contrast, the metabolic changes (PCr, Pi, pH, lactate, etc.) reported during dynamic
exercise are much larger than those seen during corresponding periods of isometric contrac-
tions (9.3 mM ATP/s and 4.7 mM ATP/s, respectively; Jones, 1993).
The force a muscle can generate depends ultimately on the number of actomyosin
cross-bridges that form and on their turnover rates. Curtin and Edman (1994) found a
progressive decline in ATP turnover rate during fatigue when frog muscle fibers were
stimulated periodically for 1 sat 100 Hz. A similar decline in ATP turnover rate is also evident
during fatigue of human voluntary contractions (Jones et aI., 1993) in which the highest
turnover rates are recorded during dynamic shortening contractions (running, cycling, etc.).
While turnover rates can clearly be influenced by many factors (e.g., pH), any reduction in
 364 B. Bigland-Ritchie et al.

Sustained Q)

MVe :oJ
~
Cii
:;:::
:!:;
-oe

2 3 4
Time (min)

B
Q)
Intermittent :oJ
MVe Cii
>
2
:!:;
-oe

00
10 20
Time (min)

20 100
C OJ
~
.......................
0
Intermittent
30% MVe
E
.s
per
q
••••••••••••••••• •• 0 Force ::2
0til 10 50 ~
--I
Q)
(j ea
a.. LL
a: ATP
Lact
~
0 0
10 20 30 40
Time (min)

Figure 2. Metabolic changes in relation to force reduction during fatigue from (A) sustained and (B) intermit-
tent MVCs of tibialis anterior (adapted from Weineret aI., 1990), and (C) intermittent 30% MVCs of quadriceps
(adapted from V011estad et aI., 1988).

ATP turnover reduces a muscle's ability to generate maximal force or power. However, since
examples can be found that show no simple relation between metabolism and muscle
performance, other possible causes of fatigue must also be considered.

Excitation/Contraction Coupling
In the study by V011estad and coworkers (1988), the MVC force decline could not
be attributed to metabolic factors nor to other causes such as failure of neuromuscular
transmission or of CNS motor drive. Thus, by exclusion, force loss was attributed to failure
of excitation/contraction (EC) coupling. Indeed, recent evidence indicates that failure at this
Task-Dependent Factors in Fatigue of Human Voluntary Contractions 365

site may underlie and limit the rate of force or power loss in many laboratory conditions
(e.g., Uinnergren et aI., 1993; Fitts, 1994). But since it is not yet possible to make direct
measurements of functional changes at this site during human voluntary contractions, it is
unwise to speculate as to how failure ofEC coupling may be influenced by a change of task.
Nevertheless, it is now clear that conduction in t-tubules, and hence muscle activation,
becomes progressively impaired when excitation rates are high, but remains intact or can be
restored if stimulus rates are low (see Uinnergren et aI., 1993; Fitts, 1994). Impaired
conduction when excitation rates are high probably depends on the rates at which K+, H+,
and other metabolites accumulate within the t-tubules. However, fatigue from low force
contractions is more likely to reflect changes in the amount of Ca2+ released per impulse
from the sarcoplasmic reticulum (SR) into the cytosol. Thus, in voluntary contractions,
impaired conduction in t-tubules is more likely to occur when high forces are exerted.
Experiments on both intact human and isolated rat muscles show that a fatigued
muscle can generate more force when it is stimulated at low rather than high excitation rates
(Bigland-Ritchie et aI., 1979; Jones et aI., 1979). The stimulus protocol (50 Hz, turned down
abruptly to 20 Hz) is similar to that used by Allen and coworkers (1989) on single frog muscle
fibers to separate the influence of t-tubular conduction failure from that of reduced Ca2+
release per impUlse.
Reduced Ca2+ release per impulse appears responsible for the faster rate of force
decline of twitches, or response to low-frequency stimulation, compared to the correspond-
ing changes in either maximal tetanic or voluntary contraction force (often termed low-fre-
quency fatigue), a phenomenon seen readily in human muscle responses when fatigue is
induced by most types of contraction protocols. Thus, the term low frequency refers to the
stimulus rate used to test the response, rather than to that responsible for inducing it, as is
sometimes assumed. However, as at many other sites, EC coupling seems to include a large
safety factor, since the application of caffeine to either whole muscles (Jones & Bigland-
Ritchie, 1986) or single fibers (Allen et aI., 1989) causes a large release of additional Ca2+
with near full force recovery, even after the force from stimulation has been reduced to low
values.

Role of Occluded Muscle Blood Flow

The degree to which the contractile properties of fatiguing muscle are influenced by
depletion of energy substrates and accumulation also depends heavily on how muscle blood
flow is affected by the exercise. In a sustained MVC, for example, muscle blood flow is
totally occluded by the high intramuscular pressure generated by the contraction itself, and
force usually declines by about 30-50% in 60 s. However, this rate of MVC force loss is
substantially reduced if the same force-time integral is divided into segments with rest
periods between, presumably because the event(s) that cause fatigue are partially reversed
by reactive hyperemia in each rest period between contractions. If the same intermittent
MVC protocol is repeated with the blood supply occluded by a cuff, the rate offorce decline
now matches that of the sustained MVC (Rice et aI., 1991). If the intervals between
contractions are sufficiently increased or if the contraction force is sufficiently reduced, a
balance is reached between vascular supply and the increased metabolic demand of the
exercise such that no fatigue ensues. The ratio between contraction time and repetition rate
has been used to calculate afatigue threshold (e.g., Bellemare & Grassino, 1982; McKenzie
& Gandevia, 1988). Thus, this product is deemed to provide both a qualitative and quanti-
tative estimate of the degree to which blood flow meets metabolic demands.
Blood flow becomes occluded whenever the intramuscular pressure exceeds systolic
pressure. But systolic pressure has a relatively fixed value at any given time, while maximal
forces of large and small muscles vary widely. Thus, it seems unlikely that this rule can be
366 B. Bigland-Ritchie et al.

applied equally to all muscles. Estimates of the contraction intensity at which blood flow
fails to meet metabolic demands are complicated by differences in muscle architecture. The
site of greatest constriction is usually not in the blood vessels themselves, but due to their
compression by adjacent muscle fibers or bone (Hudlicka, 1973). Intramuscular blood
vessels lie mainly between and parallel to muscle fibers, so that blood flow restriction is
likely to be greatest in muscles with a pennate fiber arrangement. When muscles contract
their blood flow is probably unevenly distributed because intramuscular pressure is generally
greatest at its center. It is interesting that the highest density of oxidative, fatigue-resistant
fibers is often found in this central location, the first area to become ischemic, and that, under
ischemic conditions the fatigue properties of all muscle fiber types can be characterized as
fast-fatigable (Sahlin et ai., 1987).
The presence or absence of ischemia has a major influence not only on changes in
force generating capacity but also on muscle contractile speed, and probably also determines
whether inhibitory reflexes are initiated (see below). An open question is the extent to which
muscle blood flow is restricted in the types of contractions used in different types of voluntary
exercise executed at various submaxima1 intensities and duty cycles, as well as how these
estimates may vary between muscles. Early work (e.g., Edwards et ai., 1972; Saltin et ai.,
1981) found that intramuscular pressure in quadriceps increases linearly up to 100% MVC.
Arterial blood pressure also increases, but does so non-linearly and plateaus at about 50%
MVC. This muscle is therefore considered to be ischemic at forces above 50% MVC
(Sjogaard et ai., 1986). However, it is likely that blood flow becomes restricted and unevenly
distributed when the force rises above about 10 or 20% MVC (Folkow & Halicka, 1968;
Sjogaard et ai., 1986). Blood flow restriction is more severe in isometric contractions as
compared with dynamic contractions. These studies confirm a general rule of thumb that
blood flow is reduced at forces> 25% MVC (Barcroft & Millen, 1939).
While the importance of muscle blood flow is well recognized in principle, the
profound differences in how muscles behave under aerobic and ischemic conditions, and
how this behavior varies between muscles and with time, often tends to be overlooked.

Transmission at the Neuromuscular Junction and Muscle Surface

Membrane
Failure of impulse transmission across the neuromuscular junction, detected as a
reduction in the size of the muscle compound action potential (M-wave), has been recognized
as a potential cause offatigue for many decades. The rate at which M-waves decline depends
largely on the frequency at which impulses are delivered (Krnjevic & Miladi, 1958;
Bigland-Ritchie et ai., 1979). Hence, if transmission failure contributes to fatigue of
voluntary contractions, it would be most evident in tasks in which excitation rates are high.
M-wave size can be influenced independently of force by many factors such as
changes in the dynamics of the Na+/K+ pump and muscle fiber conduction velocity (Hicks
& McComas, 1989). Thus, there is either a wide safety margin between membrane depolari-
zation and muscle activation, or the events that occur at the muscle fiber membrane are not
always reflected accurately by M-waves recorded from the muscle surface. Indeed,
Grabowski and colleagues (1972) found twitch forces of amphibian muscle fibers remained
constant even when transmembrane action potentials were reduced 40% by lowering
extracellular Na+, a decline far larger than any reported during fatiguing voluntary contrac-
tions. An increased duration and amplitude reduction is also evident in M-waves from
unfatigued fibers of both rat and amphibian muscles when extracellular K+ is reduced to
levels similar to those recorded during fatigue of human muscles (Hnik et ai., 1972; Jones
Task-Dependent Factors in Fatigue of Human Voluntary Contractions 367

& Bigland-Ritchie, 1986; Sj0gaard et aI., 1986). Under these conditions, action potential
amplitudes decrease while twitch forces increase (Uinnergren & Westerblad, 1986).
One difficulty in comparing reports from different studies arises because no consen-
sus has been reached as to how M-wave size should most properly be measured. Some
measure peak-to-peak amplitudes and durations, others the area of either the whole M-wave
or only its first negative phase with respect to the isoelectric baseline. Each method is
vulnerable to some distortion due, for example, to interelectrode distance, changes in
conduction velocity, etc., and other factors that become important during fatigue. When
Thomas and colleagues (1989) measured the same M-wave data both ways they found
peak-to-peak amplitudes and durations always changed more than did the area of the first
half-waveform (see also Bigland-Ritchie et aI., 1982). The importance of reporting changes
in both M-wave amplitude and area is also emphasized by Milner-Brown and Miller (1986).
While it is easy to demonstrate that M-waves become reduced in size when muscles
are stimulated at rates above 10-20 Hz, no overall consensus has yet been reached about
changes in M-waves when measured during fatigue of human voluntary contractions. For
example, despite repeated efforts, Bigland-Ritchie and colleagues (1979, 1982) and Thomas
and colleagues (1989) found no decline in M-wave size during fatigue from sustained
maximal contractions of adductor pollicis (AP), tibialis anterior (TA), or during submaximal
contractions ofthe human diaphragm (Bellemare & Bigland-Ritchie, 1987; McKenzie et aI.,
1992). Similar findings have also been reported by Garland and coworkers (1988), Zijdewind
and Kernell (1994) and others. However, these reports differ from others in the literature,
including: Stephens and Taylor (1972), and Bellemare and Garzaniti (1988) in sustained
MVCs ofFDI and AP, respectively; Fuglevand and coworkers (1993) in sustained submaxi-
mal FDI contractions of various intensities; and Aldrich (1987) and others for diaphragm.
Thomas and coworkers (1989) found near-linear rates of force loss in TA and FDI during
sustained MVCs held for 5 min. While M-wave areas recorded from TA did not change
throughout the contraction, those from FDI did decline, but only during the first 2 min of
each contraction, remaining stable thereafter, while force continued to fall.
The extent to which impaired transmission may contribute to fatigue of voluntary
contractions is likely to depend on the motoneuron discharge rates and hence the contraction
intensity. Excitation rates recorded during voluntary activity are generally lower than those
used for stimulation and may decline with time (Bigland-Ritchie et aI., 1983a,b). Moreover,
the maximal firing rates delivered simultaneously to different units vary widely. A few units
may fire as fast as 50 Hz, while others never seem to exceed 10-15 Hz (DeLuca et aI., 1982;
Bellemare et al,. 1983). It is clearly beneficial if the CNS can keep excitation rates as low
as possible, without compromising force, since this reduces the likelihood of t-tubular
blockage, the rise of extracellular K+, and depletion of metabolic substrates, as well as
transmission failure at the neuromuscular junction. This goal is best achieved if the impulse
rates delivered to each unit vary in relation to its contractile properties, which seems to be
the case. This goal cannot be achieved by nerve stimulation whatever fixed frequency is
chosen. However, it would be surprising if impaired transmission does not occur during
fatigue from some type of exercise (e.g., Fuglevand et aI., 1993) since safety margins seldom
exceed those sometimes called upon in one task or another.
The comparison of M-wave behavior in different muscles (Thomas et aI., 1989)
suggests that perhaps there is a real difference between the EMG responses ofFDI and other
muscles. Fuglevand and coworkers (1993) found that M-wave amplitudes recorded from
FDI also decline during fatigue from sustained submaximal contractions. Surprisingly, the
rates at which they declined were faster for low compared to high forces. However, these
low force contractions could be sustained much longer. This unexpected behavior may
perhaps be explained, in part, by recent reports from Zijdewind and colleagues. (1995). They
found that, during fatiguing submaximal contractions ofFDI, the EMG responses may either
368 B. Bigland-Ritchie et al.

increase or decrease depending on the precise position of the recording electrodes. In

submaximal contractions, the amount of motor drive directed to each part of this muscle is
unevenly distributed and is strongly task related. Thus, the CNS can achieve a common goal
(e.g., a 50% MVC abduction of the index finger) by recruiting different unit subpopulations
that can be used in a variety of task-specific ways.

Relation between Muscle Force and Speed Changes

It is now clear that the force a muscle can exert does not necessarily change in parallel
with its contractile speed. The relation between force and contractile speed depends on the
type of exercise from which fatigue develops as well as on the muscle fiber types recruited.
Fig. 3A shows changes of contractile speed recorded during an MVC of AP sustained for 60
s, in which force typically declines by about 30-50%. In this case the maximal rate of
relaxation (MRR) declines almost in parallel with the force, but maximal contraction rates
(MCR) are much less affected (Bigland-Ritchie et aI., 1983b). MRR changes are similar
whether measured from voluntary or stimulated contractions. However, when the same MVC
force-loss is caused by low-force quadriceps contractions repeating for 30 min (Fig. 3B),
contraction rates get faster. Twitch contraction and half-relaxation times both get signifi-
cantly shorter while their maximal rates increase (Bigland-Ritchie et aI., unpublishsed data;

A
100

ClJ
MeR
05-0
~

§ -2:50
.:;:; 0
ro 0

~*
C3
a:
, ' ,
0
0 20 40 60
Time (s)
B
200

g150
c
o
o
~ 100

50L-~--~~~-7.~
o 10 20 30
Time (min)

C 100/",,=-_ _

80 '"
~ 60 Peak Power "-...
8
~~ /j300U;-
~
_ ,_."__ ./ 2001- Figure 3. Whole muscle contractile speed changes dur-

20~, !O~
ing: A, isometric MVC of adductor pollicis (adapted from
Bigland-Ritchie et aI., 1983b); B, repeated 30% MVCs of
Ofresh 1 2 3 4 quadriceps (Bigland-Ritchie unpublished data); C, dy-
Fatiguing sequence namic leg extension (adapted from Jones, 1993).
Task-Dependent Factors in Fatigue of Human Voluntary Contractions 369

Rice et aI., 1992). In contrast, Fig. 3C illustrates the contribution of contractile slowing to
the greater reduction in power output (force x velocity) compared to isometric force, which
occurs during fatigue from dynamic exercise (Jones, 1993).
Changes of contractile speed are also influenced by the fiber type composition of the
particular muscles or motor units recruited in each task. In experiments on cats, Botterman
and Cope (1988), Dubose and colleagues (1987) and others found that fast-fatigable (FF)
units became markedly slower when fatigued, while the speed offatigue resistant (FR) and
slow (S) units either remained unaltered or got faster (Botterman, Chapter 26). If different
fiber types are recruited sequentially, the contractile speeds recorded from whole muscles,
together with their rate of change during fatigue, must vary with the contraction intensity of
the particular exercise involved.

Differences between Human and Animal Models

The contractile properties of human muscle cannot always be closely predicted from
knowledge of their fiber-type composition determined histochemically. For example, the
human biceps brachii muscle usually has about equal numbers of Type I (slow) and Type II
(fast) units while both AP and soleus are reported to be about 70% Type I (Johnson et aI.,
1973). Thus, one would expect corresponding differences in their contractile speeds. How-
ever, Bellemare and coworkers (1983) and Round and colleagues (1984) both found little
difference between either the contraction or half-relaxation times measured for biceps brachii
and adductor pollicis; while those of the soleus were 50% longer. These findings raised the
question of how accurately histochemical staining relates to muscle contractile performance
measured physiologically (Rice et aI., 1988) and also how far contractile properties of motor
unit types in animals apply to human muscle. While large interspecies differences are known,
it has not been possible to obtain comparable measurements from humans.
The contractile properties of human motor units can be examined by intraneural
stimulation of single motor axons (Westling et aI., 1990). This method has the advantage
over spike-triggered averaging or stimulation of intramuscular nerve branches in that a wide
range of stimulus rates can be delivered, including those sufficiently high to elicit maximal
tetanic forces. Thus, force-frequency relations can be measured before and after fatigue
induced by various excitation patterns, including the standard Burke fatigue test and those
recorded during voluntary activity. Moreover, the conditions under which changes in motor
unit force and speed are measured in these experiments are identical with those in most
animal studies. However, the values measured in both animal and human stimulation studies
neglect any influence on force that may result from the simultaneous activity of other
surrounding units (Binder-Mcleod & Guerin, 1990). Nevertheless, Thomas and coworkers
(1991 b) were not able to classify human thenar motor units into three types using the criteria
normally used in most animal studies. The most noticeable differences were that: 1) while
the contractile speeds of different motor units varied widely, the individual values did not
correlate with either the unit's force (size) or its fatigue index, measured after a standard 2
min Burke fatigue test; and 2) their fatigue indices showed no sign of a bimodal distribution,
but were distributed continuously. In fact, human thenar motor units were remarkably
fatigue-resistant, with no units responding like FF units in animal muscles (i.e., no unit lost
more than 75% of its initial force). If FF units are present, we would expect to find them
since motor axons are encountered randomly in microneurograghic procedures. If FF units
are lacking in human intrinsic hand muscles, they may be rare in all human muscles.
Alternatively, their absence may be due to an evolutionary adaptation peculiar to human
intrinsic hand muscles. This idea seems reasonable since hands are used for precision grip
and other tasks not performed by other animals.
370 B. Bigland-Ritchie et at.

Force-frequency relations measured for individual units before and after fatigue
showed another puzzling feature; namely, that only about 113 of all units required an increase
in excitation rates to elicit the same force after fatigue. Equal numbers of other units needed
either no change in rates or lower rates compared to those required before fatigue (Thomas
et aI., 1991a). At first sight, this presents a major problem for the CNS ifit is to regulate
motoneuron firing rates so that a constant load, such as a heavy suitcase, can continue to be
supported as fatigue develops.

Force-Velocity Relations in Dynamic Contractions

The term force and power are often used interchangably to describe fatigue, thus
assuming these two parameters change in parallel. However, it is clear that force and power
can follow quite different time courses as fatigue progresses (see Fig. 3C). Curtin and Edman
(1994) found that isometric force of isolated frog muscle fibers was reduced by 26%,
compared to only a 10% reduction in maximal shortening velocity when I s tetani were
delivered every 2, 3, or 5 min. During muscle lengthening, force-velocity relations showed
that the ability to resist stretch was well preserved. The fibers also became less stiff,
indicating fewer attached cross-bridges in fatigue. These changes in V max could also be
induced by intracellular acidification with CO 2 although this caused smaller changes in
isometric force than those seen during fatigue. They conclude that changes in isometric force
may be due to an increase in both [Pi] and [H+], while V max is affected by [H+] alone.
For human quadriceps or AP muscles, Jones (1993) found contractile slowing under
ischemic conditions is greater when fatigue is induced by isokinetic shortening, compared
to that from sustained or intermittent isometric contractions. For example, they found that
when the isometric force of AP was reduced by 25% by repeated bouts of 50 Hz supramaxi-
mal nerve stimulation, relaxation times increased; thus, power was reduced by about 50%.
Under other conditions, isometric force fell by only 10%, while Vmax values were reduced
by about 50% thus dissociating any parallel change in force and power during fatigue. Hence,
tests for fatigue that depend on examining only changes in either MVC or stimulated forces
are misleading if applied to changes in human performance capacity during dynamic
exercise. These authors also provide evidence that voluntary effort can fully activate human
muscles during shortening as well as isometric contractions (see also Gandevia et aI., Chapter
20).

RELATION BETWEEN MUSCLE PROPERTIES AND

MOTONEURON FIRING RATES

Role of Excitation Frequency

The rates of motor unit discharge vary widely from one unit to another, with maximal
rates ranging from 10-50 Hz, distributed relatively normally between units. Mean rates
recorded during brief MVCs of different muscles generally vary between about 25-30 Hz,
sometimes showing consistent variations between subjects. For soleus, they did not exceed
about 15 Hz. These values vary roughly in proportion to differences in each muscle's
contractile speed (Bellemare et aI., 1983; Bigland-Ritchie et aI., 1983b, 1991). A similar
relation between excitation rates and speed is also evident within a single motor unit
popUlation (Thomas et aI., 1991b).
Since the force from maximal voluntary contraction usually matches that from
supramaximal nerve stimulation, it is generally concluded that voluntary effort can recruit
Task-Dependent Factors in Fatigue of Human Voluntary Contractions 371

all units and drive them to generate maximal tetanic force, and also that the rates needed to
elicit maximal force are higher for fast units. Conversely, if firing rates exceeded the
minimum needed to generate maximal force, the slope of the EMG/force relation would
increase sharply at high forces, something that is not seen (Woods & Bigland-Ritchie, 1983).
Thus, it is reasonable to suppose that, even in the absence of fatigue, the CNS adjusts spinal
motoneuron firing rates to match differences between the contractile properties of the
individual motor units they supply.
When an MVC is prolonged, relaxation rates slow and this reduces the excitation rate
needed to achieve maximal force. Thus, a decline in firing rates (or surface EMG) recorded
during a prolonged MVC does not necessarily imply that the motor drive to a muscle now
fails to activate it fully (i.e., it is not necessarily a sign that central fatigue is present). The
balance between force and excitation rates is well illustrated in experiments that show that
both the typical force and EMG decline, seen in a sustained MVC, can be mimicked if the
motor nerve is stimulated initially at high rates and then these rates are progressively reduced
(Bigland-Ritchie et aI., 1979; Jones et aI., 1979). The precise excitation rates are critical. If
they are reduced either too fast or not fast enough, force falls below its optimal value. One
benefit of keeping motoneuron firing rates as low as possible relates to the frequency-de-
pendence of transmission failure and other peripheral causes of fatigue.
At present, it is not clear whether the firing rates of all motor units decline equally
with time during a sustained MVC or whether the fall in mean rates is caused by a more
precipitous decline for those units that fire initially at the highest rates (Fuglevand et aI.,
1994). We know that, in sustained submaximal isometric contractions, the change in motor
unit discharge rate during the fatigue task was related to the initial discharge rates (Garland
et aI., 1994). That is, those units that increased discharge rate were among those motor units
with low initial discharge rates «10 Hz) and no motor units with high initial discharge rate
(> 20 Hz) displayed an increase in discharge rate during the fatiguing contraction. Further-
more, in that experiment, the discharge rate of most motor units that were active from the
beginning of the contraction declined, whereas the discharge rates of most newly recruited
units were either constant or increased slightly. However, when Garland and coworkers
utilized a fatigue task incorporating dynamic contractions, only a minority of motor units
exhibited a significant decline in firing rate despite the similarity in duty cycle between the
two fatigue paradigms. Thus the decline in discharge rate of motor units was not uniform
across all units and diminished with a dynamic fatigue task.
Intermittent contractions, executed at either maximal or submaximal forces with
blood supply intact, reduce a muscle's capacity to generate force and endurance times are
inversely related to contraction intensity and duty cycle. When a subject makes repeated
target force contractions of either 30% or 50% MVC (duty cycle, 0.6) the integrated EMG
(IEMG) increases progressively until, at the limit of endurance, it matches that recorded in
an MVC (Bigland-Ritchie et aI., 1986a,b). Muscle speed did not get slower during the 30%
MVC contraction protocol. While motor unit firing rates recorded during the target-force
contractions increased gradually with fatigue, there was no decline in those recorded during
brief test MVCs made every 5 min throughout the exercise period (Bigland-Ritchie &
Vellestad, unpublished data). Similarly, repeated MVCs, each held for only lOs (duty cycle,
0.5), produced only small changes in IEMG, firing rate, and contractile speed. But when
these were repeated with the blood supply occluded, all indices of fatigue changed rapidly,
and were not significantly different from those of a sustained MVC held for an equal total
integrated force x time period (Rice et aI., 1991). Garland and McComas (1990) found that
the IEMG of soleus also declined when fatigue was caused by periodic stimulation at 15 Hz
under ischemic conditions. Thus, a decline in motoneuron firing rates with fatigue seems
only to occur in the presence of contractile slowing, and when metabolic factors are involved.
Hence, both contractile speed and firing rate changes depend on whether or not sufficient
372 B. Bigland-Ritchie et a!.

rest periods are allowed between contractions to avoid ischemia, and to replenish or remove
the metabolic consequences of the previous contractions. However, again it makes good
sense that a common factor should control both these parameters.
In human voluntary contractions, the match between changes in motoneuron firing
rate and contractile speed (mainly relaxation rate), although close, is not always exact. Thus
it is possible that the large differences between endurance times for different individuals and
their rates of MVC force decline may be due, in part, to variations in the sensitivity with
which their CNS adjusts motor unit firing rates to changes in contractile speeds. The
difficulties presented to the CNS become even greater when contractile speed changes due
to fatigue are compounded by speed changes imposed by other factors, such as changes in
muscle length or its velocity of shortening.

Firing Rates and Muscle Afferent Feedback

It is not yet clear how changes in motoneuron firing rates are regulated during fatigue.
If the excitatory input to spinal motoneurons remains constant, then firing rates are probably
dictated by changes in the time course of their excitability (Kernell & Monster, 1982).
Hagbarth and colleagues (1986) suggest that the decline in discharge rates with fatigue results
mainly from withdrawal ofIa facilitation from muscle spindles; others favor an inhibitory
reflex carried by groups III and IV muscle afferents initiated by metabolic changes within
the fatigued muscle (Bigland-Ritchie et aI., 1986c; Woods et aI., 1987; Garland & McComas,
1990). Evidence for all three mechanisms is persuasive and it seems probable that all
participate in different ways.

Stimulation vs Voluntary Contraction

F or studies of fatigue in intact or isolated animal preparations, contractions are
elicited by nerve or muscle stimulation. Human muscles are often stimulated in experiments
in which the degree of muscle activation must be controlled, free from fluctuations imposed
by spinal reflexes or changes in a subject's effort. They are also stimulated for training
purposes and in studies designed to restore function to paralyzed muscles of paraplegics,
etc. Clearly, the rates and patterns of impulse delivery have a major impact on the outcome.
Although electrical stimulation has been advocated to increase muscle strength,
Hainaut and Duchateau (1989) found that individuals trained with submaximal 100 Hz
electrical stimulation became more fatigable and showed less response to training (when
tested with MVCs) than those subjects trained with submaximal voluntary contractions. This
provides an example of the well-known phenomenon of task specificity of training
(Duchateau & Hainaut, 1984) and illustrates the potential for lack of transfer of training from
electrical stimulation to voluntary contractions. Binder-MacLeod and colleagues have
demonstrated a reduction in the fatigability of quadriceps femoris muscle when stimulation
parameters were adjusted to mimic voluntary motoneuron discharge (Binder-MacLeod &
Guerin, 1990; Binder-MacLeod & Barker, 1991). Elucidating the best electrical stimulation
protocol, however, has obvious implications for the field of functional electrical stimulation
that may be used in the rehabilitation of individuals with spinal cord injury, for example.
Electrical stimulation may fail to address the task-dependent nature of muscle fatigue.
For instance, the task-specific training response is affected by the motor command rather
than just the type of muscle contraction (Behm & Sale, 1993); electrical stimulation would
fail to address the effectiveness of the motor command. In addition, motor unit recruitment
and rotation of motor units have been implicated in fatiguing contractions (for review, Enoka
& Stuart, 1992). Electrical stimulation excites different motor units from voluntary contrac-
Task-Dependent Factors in Fatigue of Human Voluntary Contractions 373

tions (Trimble & Enoka, 1990) and may not afford the possibility for rotation of motor unit
activity within the motoneuron pool.

CNS STRATEGIES: CENTRAL FACTORS

Motivation, Force, and Effort Sensations

A subject's ability to judge the intensity of the force exerted is thought to be mediated
by an efferent copy to the cortex of the motor command that activates the spinal motoneuron
pool. The sense of force or effort increases with fatigue, for example, during a submaximal
contraction held at constant force. This over-estimation probably results from the concurrent
increase in neural drive (EMG) required to compensate for contractile failure in units already
active (Jones & Hunter, 1983; Gantchev et aI., 1989; see Jones, Chapter 22).
The sense of effort is a major factor influencing motivation. As fatigue ensues, the
increased sense of effort tends to make subjects think they can no longer drive their muscles
maximally or continue to exercise at all. The ability to activate a muscle during high force
contractions or to prolong endurance time is clearly improved by practice. While this process
is attributed mainly to peripheral events, the greater willingness of subjects to put up with
discomfort and push themselves harder is important. This willingness is clearly enhanced if
the event takes place in a competitive environment. Hence, sports records are more often
beaten in competitions, in which participants make a greater effort to get psyched up, than
in practice sessions. The problem of motivation is also vital in laboratory tests, particularly
for protocols that continue until exhaustion or where fatigue is measured by endurance times.
In the laboratory, this atmosphere can be mimicked if subjects are encouraged to compete
against their own previous performance or work in groups, vying with each other as to who
can do best.

Voluntary Muscle Activation

Methods for assessing central motor drive and for measuring the magnitude of central
and peripheral fatigue are described elsewhere (see Gandevia, 1993; Gandevia et aI., Chapter
20). Here we consider only whether the ability of the CNS to activate muscles during fatigue
varies between tasks.
In 1954, Merton showed that the force of a maximal voluntary effort matches that
elicited by supramaximal tetanic stimulation, and also that no additional force can be elicited
when single shocks are applied at any time during a 3 min sustained MVC (twitch occlusion).
Similar results have since been shown for many other muscles, including quadriceps, biceps
brachii, TA, and FDI. Thus, the CNS is generally thought to be capable of recruiting all motor
units and continuing to excite them at rates sufficient to elicit near-maximal force, despite
the presence offatigue (see McKenzie & Gandevia, 1991; Rice et aI., 1991; cf. Gandevia et
aI., Chapter 20). It is not surprising that, in the early stages of prolonged exercise, muscles
can be near-maximally activated, but later this ability often declines. However, most studies
report that force loss due to reduced central motor drive is small « 10%) compared to that
due to other causes.
However, when either soleus or the diaphragm were fatigued by repeated submaximal
contractions, central factors appeared to playa larger role (Bigland-Ritchie et aI., 1986a;
Bellemare & Bigland-Ritchie, 1987). Unlike quadriceps responses, twitches elicited from
the diaphragm did not become occluded, and for both muscles the EMG associated with brief
test MVCs showed a marked decline with time. It is unlikely that the difference between the
behavior of quadriceps and that of either the diaphragm or soleus can be attributed to lack
374 B. Bigland-Ritchie et al.

of motivation because most experiments were performed by the same subjects. However,
later experiments showed that the amount of central fatigue evident in the diaphragm is
task-dependent (see below). Central fatigue is clearly evident when a subject is asked to
contract their diaphragm against a fatiguing inspiratory load (opposed by the weaker
intercostal muscles), but almost absent when it is fatigued by expulsive contractions, in
which maximal diaphragmatic force is less than that ofthe abdominal muscles which oppose
it (McKenzie et ai., 1992).

Role of Synergists
Experiments on limb muscles are often hampered by an unavoidable recruitment of
unfatigued synergists, particularly when fatigue is elicited by intermittent voluntary contrac-
tions (cited in Bigland-Ritchie et ai., 1986a, for contractions of AP). Similarly, in the
respiratory system, when one muscle becomes fatigued the motor drive is always switched
to elicit force from other less fatigued synergists whenever possible (Yan et. ai., 1993). This
strategy is evident in the paradoxical breathing often seen in patients in respiratory failure
in which periods of contractions during inspiration alternate between the diaphragm and
intercostal muscles. Similarly, Johnson and coworkers (1993) showed that the relative
contribution of the diaphragm to total respiratory motor drive is progressively reduced during
fatigue to exhaustion from dynamic exercise performed at 95% and 85% VOzmax. During
the first 5-10 min of exercise, the decline in trans diaphragmatic pressure (Pdi) was propor-
tional to that of esophageal pressure (Pe), after which fpdi /min stayed constant, while
fPe/min, Ve, and inspiratory flow rates continued to rise. In marathon runners, Chevrolet and
colleagues (1993) found that Pdi declined but with no change in Pe, maximal voluntary
ventilation (MVV), or lung mechanics, indicating that other respiratory muscles play an the
increasing role in breathing when the diaphragm becomes fatigued.

Possible Overall eNS Responses

The intensity and duration of exercise that can be maintained is determined mainly
by the contractile properties and fatigue resistance of the motor units in the particular muscles
involved. This can also be influenced by interactions between different muscles. Some
subjects have difficulty in excluding co-contractions by muscle synergists during fatigue,
even of those not closely related to the prime mover. Howard and Enoka (1991) provide
evidence that full muscle activation is possible when an isometric MVC is performed by one
leg, but not when both legs are involved, indicating that the amount of muscle mass may be
important. While it is generally assumed that the CNS is organized to control respiratory and
limb muscles independently, and that limb exercise is not limited by the capacity for
respiration, recent evidence suggests that this may not be so under some exercise conditions
(Boutellier & Piwko, 1992).
The concept that respiration does not limit exercise relies mainly on the fact that
voluntary efforts can achieve maximal rates of ventilation that are significantly higher than
the maximal rates elicited involuntarily by any kind of normal exercise. For example, when
cycling to exhaustion at 80% of the maximal work capacity, ventilation was found to be only
50% of that recorded during 12 s periods ofMVV. Thus, the chemical drive from exercise
fails to activate respiratory muscles maximally; in fact, they always seem to retain a plentiful
reserve (Mador et ai., 1993). This design may be essential for preservation of life in that it
protects respiratory muscles from excessive fatigue. For example, at the end of a race no
harm is done if a runner's skeletal muscles are exhausted to the point where the runner can
no longer stand, but their respiratory muscles must continue to function well above resting
levels for some time against a markedly increased COzload.
Task-Dependent Factors in Fatigue of Human Voluntary Contractions 375

Hypoxia
Some data presented from the Everest II experiments showed interesting interactions
between fatigue and chronic hypoxia. During chronic hypoxia 1) the CNS was usually
capable of fully activating a small muscle mass (triceps surae) and of keeping it maximal
activated when fatigued by intermittent MVCs for 180 s; and 2) in contrast, during dynamic
exercise performed by both legs simultaneously (cycling), maximal power output was
markedly depressed; and 3) under these hypoxic conditions, there was no similar depression
of power output by respiratory muscles, since maximal exercise ventilation rates were not
reduced at extreme altitudes. Why should hypoxia have such a profound effect on limb
muscle performance, while the work capacity of respiratory muscles seemed unaffected,
when both muscle groups were performing similar types of dynamic exercise? These
observations seemed incompatible with accepted theory. To explain them, Bigland-Ritchie
and V 0llestad (1988) postulated that it is essential for survival that somehow respiratory
muscles must avoid the extremes of fatigue experienced by limb muscles; they suggested
that this could be achieved if CNS strategy involves some kind of reciprocal inhibition
between the motor drive to limb and respiratory muscles, with that from the respiratory
system dominating. According to this scheme, the motor drive to limb muscles can achieve
and retain maximal muscle activation, provided this does not increase the metabolic demand
above that which the respiratory muscles can deliver. However, if either the muscle mass is
too large or the exercise sufficiently demanding, such that metabolism exceeds the capacity
of the oxygen delivery system, a balance between them is restored by an automatic reduction
of motor drive to limb muscles. In this case, the limb muscles can no longer be fully activated
and centralfatigue ensues. They suggested that this may be brought about by reflex inhibition
arising mainly from the fatigued diaphragm, but it could equally well be caused by reciprocal
inhibition between the two systems acting centrally within the brainstem.
Direct evidence in support of this hypothesis comes from experiments by Kayser and
coworkers (1993) who found that the rate ofIEMG increase, measured during fatigue of arm
muscles, was the same at sea level and at high altitude. A similar IEM G increase was recorded
from leg muscles while cycling at sea level. Thus, in all three cases the IEMG behaved in
the expected way, indicating that additional motor units were recruited and those already
active were driven harder as fatigue developed. However, when cycling at high altitude, the
IEMG responses ofleg muscles did not change throughout the exercise. The usual increase
in motor drive which compensates for contractile failure did not occur. These authors also
confirmed that chronic hypoxia reduces the normal increase in muscle lactate and other
fatigue-related compounds. Thus, fatigue developed under these conditions may have been
caused more by a reduced motor drive than by peripheral factors. This concept also fits the
observations that the IEMG of the diaphragm only increased marginally before stabilizing
when this muscle was fatigued by repeated contractions at 50% MVC (Bellemare &
Bigland-Ritchie, 1987). Furthermore, Walsh and Banister (1995) showed that, when cycling
to exhaustion in acute normobaric hypoxia, both the maximal ventilation rate and respiratory
delivery of oxygen per unit work rate were similar to the normoxic values, although acute
hypoxia markedly reduced both V0 2max and the maximal work rate of leg muscles.
Taken together, these observations support the concept that the motor drive to limb
muscles is reduced whenever the metabolic demand of skeletal muscle exceeds that which
the respiratory muscles can supply. Consequently, leg muscles accumulate less lactate, [H+],
and other signs of fatigue than those that accumulate in smaller muscles kept fully activated
for a similar period. It is logical to conclude that this neural design has evolved because small
muscles can never impose dangerously high metabolic demands on the respiratory system.
Indeed, such a neural interaction of this kind between the skeletal and respiratory motor
systems would provide a mechanism whereby respiratory muscles are protected from
376 B. Bigiand-Ritchie et al.

extremes of fatigue commonly experienced by leg muscles. If this operates under chronic
hyperbaric hypoxic conditions, it seems likely that similar interactions are also available
under other stressful conditions which are more commonly encountered.

AL TERNATIVE EXPLANATIONS

Many reasons have been presented to suggest why muscles are likely to respond
differently when fatigued by different tasks, and why the responses of different muscles may
vary from each other, even when they perform the same tasks. However, we must also
recognize that the human body does, in fact, perform remarkably well when challenged by
tasks that require a wide range of strength and speed. Indeed, the endurance time for which
any task can be sustained varies hyperbolically in relation to either the force or power exerted,
whether the task be a sustained (Romert, 1960) or intermittent isometric contraction (Monod
& Scherrer, 1965), or power-output measure in dynamic exercise (Moritani et aI., 1981;
Poole et aI., 1990).
At first sight, the smooth transitions of fatigue-related responses seen when the type
of exercise is varied suggests that fatigue depends on only one single process. For example,
Walsh and Bigland-Ritchie (1995) present a model of fatigue, based on a single process, that
generates the well-documented hyperbolic relation between power and fatigue. This simple
model accurately predicts exhaustion times and training influences for a variety of exercise
conditions. It assumes that fatigue can either accumulate or partially recover during exercise,
depending on the intensity. Initially, the hyperbolic relations calculated in this model were
based on only the amount of K+ release per impulse, measured in experiments reported by
Sj0gaard (1991), but similar curves have also been constructed based on other factors, such
as lactate accumulation, pH changes, etc. However, current data also suggest that the single
limiting factor that dominates the rate at which fatigue develops in all types of exercise may
prove to be failure of EC coupling, regardless of whatever functional changes may occur in
parallel when measured individually at other sites during any particular type of exercise.
Current knowledge about EC coupling failure and its influence on fatigue is accumulating
rapidly, but is not yet sufficient to be able to judge how its magnitude may be affected by
changes in the task (see Stephenson et aI., Chapter 2).
It seems likely that the changes in performance during different tasks do depend on
many factors, and that nature's experiment, extending over several million years, has got the
balance between them just right. This concept would require that changing the type of
exercise shifts the major stress from one site to another, but in such a way that, in so doing,
the added stress on one site results in improved function at another. In that case, the net
outcome would cause only small but progressive change in overall performance that simulate
a single process, regardless of which type of exercise is undertaken.

ACKNOWLEDGMENTS

The authors' laboratories have been supported by the Natural Sciences and Engineer-
ing Research Council (NSERC) of Canada (S.J. G. and C.L.R.), and United States Public
Health Service grants NS 14756 and HL 30026 (B.B-R. and M.L. W). Attendance of the
authors at the 1994 Bigland-Ritchie conference was supported by NSERC (S.J.G. and
C.L.R.), McGill University (C.L.R.), NS 14756 (B.B-R. and M.L. W), and the University of
Miami (B.B-R., S.J.G. and C.L.R.).
Task-Dependent Factors in Fatigue of Human Voluntary Contractions 377

REFERENCES
Aldrich TK (1987). Transmission failure of the rabbit diaphragm. Respiration Physiology 69,307-319.
Allen DG, Lee JA & Westerblad H (1989). Intracellular calcium and tension during fatigue in isolated single
muscle fibers from Xenopus laevis. Journal of Physiology (London) 415, 433-458.
Barcroft H & Millen JLE (1939). The blood flow through muscle during sustained contraction. Journal of
Physiology (London) 97, 17-31.
Behm CG & Sale DG (1993). Intended rather than actual movement velocity determines velocity-specific
training response. Journal of Applied Physiology 74, 359-368.
Bellemare F & Bigland-Ritchie B (1987). Central components of diaphragmatic fatigue assessed from bilateral
phrenic nerve stimulation. Journal of Applied Physiology 62, 1307-1316.
Bellemare F, Bigland-Ritchie B, Johansson R, Smith S & Woods JJ (1983). Motor unit discharge rates in
maximal voluntary contractions of three human muscles. Journal ofNeurophysiology 50, 1380-1392.
Bellemare F & Garzaniti N (1988). Failure of neuromuscular propagation during human maximal voluntary
contraction. Journal of Applied Physiology 64,1084-1093.
Bellemare F & Grassino A (1982). Effect of pressure and timing of contraction on human diaphragm fatigue.
Journal ofApplied Physiology 53, 1190-1195.
Bigland-Ritchie B, Cafarelli E & Vollestad NK (1986b). Fatigue of submaximal static contractions. Lars
Hermanson Memorial Symposium: Exercise in Human Physiology. Acta Physiologica Scandinavica
128 (suppI556), 137-148.
Bigland-Ritchie B, Dawson NJ, Johansson RS & Lippold OCJ (1986c). Reflex origin for the slowing of
motoneurone firing rates in fatigue of human voluntary contractions. Journal ofPhysiology (London)
379,451-459.
Bigland-Ritchie B, Furbush F, Gandevia SC & Thomas CK (1991). Voluntary discharge frequencies of human
motoneurons at different muscle lengths. Muscle & Nerve 15, 130-136.
Bigland-Ritchie B, Furbush F & Woods JJ (1986a). Fatigue of intermittent, submaximal voluntary contrac-
tions: central and peripheral factors. Journal of Applied Physiology 61, 421-429.
Bigland-Ritchie B, Johansson R, Lippold OCJ, Smith S & Woods JJ (1983a). Changes in motoneurone firing
rates during sustained maximal voluntary contractions. Journal ofPhysiology (London) 340, 335-346.
Bigland-Ritchie B, Johansson R, Lippold OCJ & Woods JJ (1983b). Contractile speed and EMG changes
during fatigue of sustained maximal voluntary contractions. Journal ofNeurophysiology 50, 313-324.
Bigland-Ritchie B, Jones DA & Woods JJ (1979). Excitation frequency and muscle fatigue: Electrical
responses during human voluntary and stimulated contractions. Experimental Neurology 64, 414-427.
Bigland-Ritchie B, Kukulka CG, Lippold OCJ & Woods JJ (1982). The absence of neuro-muscular transmis-
sion failure in sustained maximal voluntary contractions. Journal of Physiology (London) 330,
265-278.
Bigland-Ritchie B & Vollestad NK (1988). Fatigue and hypoxia: how are they related? In: Sutton JR, Houston
CS, Costes G (eds.), Hypoxia: The Tolerable Limits, pp. 337-352. Indianapolis, IN: Benchmark Press.
Bigland-Ritchie B & Woods JJ (1984). Changes in muscle contractile properties and neural control during
human muscular fatigue. Muscle & Nerve 7, 691-699.
Binder-MacLeod SA & Barker CB (1991). Use of catchlike property of human skeletal muscle to reduce
fatigue. Muscle & Nerve 14, 850-875.
Binder-MacLeod SA & Guerin T (1990). Preservation of force output through progressive reduction of
stimulation frequency in human quadriceps femoris muscle. Physical Therapy 70, 619-625.
Botterman BR & Cope TC (1988). Motor-unit stimulation patterns during fatiguing contractions of constant
tension. Journal ofNeurophysiology 60, 1198-1214.
Boutellier U & Piwko P (1992). The respiratory system as an exercise limiting factor in normal sedentary
subjects. European Journal of Applied Physiology 64, 145-152.
Chevrolet J-C, Tschopp J-M, Blanc Y, Rochat T & Junod AF (1993). Alterations in inspiratory and leg muscle
force and recovery pattern after a marathon. Medicine and Science in Sports and Exercise 25, 501-507.
Cooper RG, Edwards RHT, Gibson H & Stokes MJ (1988). Human muscle fatigue: frequency dependence of
excitation and force generation. Journal of Physiology (London) 397, 585-599.
Curtin NA & Edman KAP (1994). Force-velocity relation for frog muscle fibres - effects of moderate fatigue
and of intracellular acidification. Journal ofPhysiology (London) 475, 483-494.
DeLuca CJ, Lefever RS, McCue MP & Xenakis AP (1982). Control scheme governing concurrently active
human motor units during voluntary contractions. Journal of Physiology (London) 29, 129-142.
Dubose L, Scherlhorn TB & Clamann HP (1987). Changes in contractile speed of cat motor units during
activity. Muscle & Nerve 10, 744-752.
378 B. Bigland-Ritchie et al.

Duchateau J & Hainaut K (1984). Training effects on muscle fatigue in man. European Journal of Applied
Physiology 53, 248-253.
Edwards (1981). Human muscle function and fatigue. In: Porter R, Whelan J (eds.), Human Muscle Fatigue:
Physiological Mechanisms. pp 1-18. London: Pitman Medical.
Edwards RHT, Hill DK, Jones DA & Merton PA (1977). Fatigue of long duration in human skeletal muscle
after exercise. Journal of Physiology (London) 272, 769-778.
Edwards RHT, Hill DK & McDonald M (1972). Myothermal and intramuscular pressure measurements during
isometric contractions of the human quadriceps. Journal ofPhysiology (London) 224, 58P-59P.
Enoka RM & Stuart DG {I 992). Neurobiology of muscle fatigue. Journal of Applied Physiology 72,
1631-1648.
Fitts R (l994). Cellular mechanisms ofmusc1e fatigue. Physiological Reviews 74, 49-94.
Folkow B & Halicka HD (1968). A comparison between red and white muscle with respect to blood supply,
capillary surface area and oxygen uptake during rest and exercise. Microvascular Research 1, 1-14.
Fuglevand AI, Macefield VG, Howell IN & Bigland-Ritchie B (1994). Adaptation of motor unit discharge rate
and variability during sustained maximal voluntary contractions. Society for Neuroscience Abstracts.
20,1759.
Fuglevand AJ, Zackowski KM, Huey KA & Enoka RM (1993). Impairment of neuromuscular propagation
during human fatiguing contractions at submaximal forces. Journal of PhYSiology (London) 40,
549-572.
Gandevia SC (1993). Central and peripheral components to human isometric muscle fatigue. In: Sargeant AJ,
Kernell, D (eds.), Neuromuscular Fatigue. pp 156-164. Amsterdam: Royal Netherlands Academy of
Arts and Sciences.
Gantchev GN, Gatev P & Ivanova T (1989). Muscle force sensation during sustained voluntary contraction.
Biomedica Biochimica Acta 48, S515-S520.
Garland SJ, Enoka RM, Serrano LP & Robinson GA (1994). Behavior of motor units in human biceps brachii
during a submaximal fatiguing contraction. Journal ofApplied Physiology 76,2411-2419.
Garland SJ, Gamer SH & McComas AJ (1988). Reduced voluntary electromyographic activity after fatiguing
stimulation of human muscle. Journal ofPhysiology (London) 401, 547-556.
Garland SJ & McComas AJ (1990). Reflex inhibition of human soleus muscle during fatigue. Journal of
Physiology (London) 429,17-29.
Grabowski W, Lobsiger EA & Luttgau, HC (1972). The effect of repetitive stimulation at low frequencies upon
the electrical and mechanical properties of single muscle fibres. PjIugers Archiv 334,222-239.
Hagbarth K-E, Kunesch EJ, Nordin M, Schmidt R & Wallin EU (1986). Gamma loop contributions to maximal
voluntary contractions in man. Journal ofPhysiology (London) 380: 575-591.
Hainaut K & Duchateau J (1989). Muscle fatigue, effects of training and disuse. Muscle & Nerve 12, 660-669.
Hicks A & McComas AJ (1989). Increased sodium pump activity following repetitive stimulation of rat soleus
muscles. Journal of Physiology (London) 414, 337-340.
Howard JD & Enoka RM (1991). Maximal bilateral contractions are modified by neurally mediated interlimb
effects. Journal ofApplied Physiology 70,306-316.
Hnik P, Vyskocil F, Kriz N & Holas M (1972). Work-induced increase in extracellular potassium concentration
in muscle measured by ion selective electrodes. Brain Research 40, 559-562.
Hudlicka D (1973). Muscle Blood Flow: Its Relation to Muscle Metabolism and Function. Amsterdam: Swets
& Zeitlinger.
Johnson BD, Babcock MA, Suman DE & Dempsey JA (1993). Exercise-induced diaphragmatic fatigue in
healthy humans. Journal ofPhysiology (London) 460, 385-405.
Johnson MA, Polgar J, Weightman D & Appleton D (1973). Data on the distribution of fibre types in thirty-six
human muscles: An autopsy study. Journal of the Neurological Sciences 18, 111-129.
Jones DA (1993). How far can experiments in the laboratory explain the fatigue of athletes in the field? In:
Sargeant AJ, Kernell D (eds.), Neuromuscular Fatigue. pp 100-108. Amsterdam: Royal Netherlands
Academy of Arts and Sciences.
Jones DA & Bigland-Ritchie B (1986). Electrical and contractile changes in muscle fatigue. In: Saltin B (ed.),
Biochemistry ofExercise VI, pp. 377-392. Champaign, IL: Human Kinetics Publishers.
Jones DA, Bigland-Ritchie B & Edwards RHT (l979). Excitation frequency and muscle fatigue: Mechanical
response during voluntary and stimulated contractions. Experimental Neurology 64,401-413.
Jones L & Hunter I (1983). Effect of fatigue on force sensation. Experimental Neurology 81,640-650.
Kayser B, Narici M, Binzoni T, Grassi B & Cerretelli P (1993). Fatigue and exhaustion in chronic hypobaric
hypoxia: influence of exercising muscle mass. Journal ofApplied Physiology 76, 634-640.
Kernell D & Monster AW (1982). Time course and properties of late adaptation in spinal motoneurones of cat.
Experimental Brain Research 46, 197-204.
Task-Dependent Factors in Fatigue of Human Voluntary Contractions 379

Krnjevic K & Miledi R (1958). Failure of neuromuscular propagation in rats. Journal ojPhysiology (London)
140, 440-461.
Lannergren J & Westerblad H (1986). Force and membrane potential during and after fatiguing, continuous
high-frequency stimulation of single Xenopus muscle fibers. Acta Physiologica Scandinavica 128,
359-368.
Lannergren J, Westerblad H & Allen DG (1993). Mechanisms of fatigue as studied in single muscle fibers. In:
Sargeant AJ, Kernell D (eds.), Neuromuscular Fatigue, pp 3-9. Amsterdam: Royal Netherlands
Academy of Arts and Science.
Mador MJ, Magalang UJ, Rodis A & Kufel TJ (1993). Diaphragmatic fatigue after exercise in healthy human
subjects. American Review oj Respiratory Disease 148, 1571-1575.
McKenzie DK, Bigland-Ritchie B, Gorman RB & Gandevia SC (1992). Central and peripheral fatigue of
human diaphragm and limb muscles assessed by twitch interpolation. Journal oJPhysiology (London)
454, 643-656.
McKenzie DK & Gandevia SC (1988). Human diaphragmatic endurance during different maximal respiratory
efforts. Journal ojPhysiology (London) 395, 625-638.
McKenzie DK & Gandevia SC (1991). Recovery from fatigue of human diaphragm and limb muscles.
Respiration Physiology 84, 49-60.
Merton PA (1954). Voluntary strength and fatigue. Journal ojPhysiology (London) 123, 553-556.
Miller RG, Giannini D, Milner-Brown HS, Layzer RB, Koretsky AP, Hooper D & Weiner MW (1987). Effects
of fatiguing exercise on high-energy phosphates, force, and EMG: Evidence for three phases of
recovery. Muscle & Nerve 10, 810-821.
Milner-Brown HS & Miller RG (1986). Muscle membrane excitation and impulse propogation velocity are
reduced during muscle fatigue. Muscle & Nerve 9,367-374.
Monod H & Scherrer J (1965). Work capacity ofa synergistic muscle group. Ergonomics. 8, 329-338.
Moritani T, Nagata A & De Vries HA (1981). Critical power as a measure of work capacity and anaerobic
threshold. Ergonomics 24, 339-350.
Newham DJ & Cady EB (1990). A 31p study of fatigue and metabolism in human skeletal muscle with
voluntary, intermittent contractions at different forces. NMR in Biomedicine 3,211-219.
Poole DC, Ward S & Whipp BJ (1990). The effects of training on the metabolic and respiratory profile of
high-intensity cycle ergometer exercise. European Journal ojApplied Physiology 59, 421-429.
Rice CL, Andrews MA & Bigland-Ritchie B (1991). Ischemic and non-ischemic fatigue from intermittent
maximal voluntary contractions. Medicine and Science in Sport and Exercise (Supp1.) 23, S126.
Rice CL, Cunningham DA, Taylor AW & Paterson DH (1988). Comparison of the histochemical and contractile
properties of human triceps surae. European Journal ojApplied Physiology 58, 165-170.
Rice CL, V0llestad NK & Bigland-Ritchie B (1992). Dissociation of fatigue-related neuromuscular events.
Medicine and Science in Sport and Exercise (Supp1.) 24, S56.
Romert W (1960). Ermittung von Erholungspausen fur statische Arbeit des Menschen. Internationale Zeit-
schriftJur Angewandte Physiologiel8, 123.
Round JM, Jones DA, Chapman SJ, Edwards RHT, Ward PS & Fodden DL (1984). The anatomy and fiber
type composition of the human adductor pollicis in relation to its contractile properties. Journal oj
the Neurological Sciences 66, 263-292.
Sahlin K, Edstrom L & Sjoholm H (1987). Force, relaxation and energy metabolism of rat soleus muscle during
aerobic contractions. American PhYSiological Society 129,1-7.
Saltin B, Sj0gaard G, Gaffney FA & Rowell LB (1981). Potassium, lactate, and water fluxes in human
quadriceps muscle during static contractions. Circulation Research 48 (Supp1. I), 118-124.
Sj0gaard G (1991). Role of exercise-induced potassium fluxes underlying muscle fatigue: a brief review.
Canadian Journal Physiology Pharmacology 69, 238-245.
Sj0gaard G, Kiens B, Jorgensen K & Saltin B (1986). Intramuscular pressure, EMG and blood flow during
low level prolonged static contractions in man. Acta Physiologica Scandinavica 128, 475-484.
Stephens JA & Taylor A (1972). Fatigue of maintained voluntary muscle contraction in man. Journal oj
Physiology (London) 220,1-18.
Thomas CK, Bigland-Ritchie B & Johansson RS (1991a). Force-frequency relationships of human thenar
motor units. Journal oJNeurophysiology 65, 1509-1516.
Thomas CK, Johansson R & Bigland-Ritchie B (199Ib). Attempts to classify human thenar motor units.
Journal oJNeurophysiology 65, 1501-1508.
Thomas CK, Woods JJ & Bigland-Ritchie B (1989). Neuromuscular transmission and muscle activation in
prolonged maximal voluntary contractions of human hand and limb muscles. Journal oj Applied
Physiology 67, 1835-1842.
380 B. Bigland-Ritchie et al.

Trimble MH & Enoka RM (1990) Mechanisms underlying the training effects associated with neuromuscular
electrical stimulation. Physical Therapy 71, 273-282.
Vl'lllestad NK & Saugen E (1993). Gradual increase in metabolic rate during repeated isometric contractions.
In: Sargeant AJ, Kernell D (eds.), Neuromuscular Fatigue, 96-97. Amsterdam: Royal Netherlands
Academy of Arts and Sciences.
Vl'lllestad NK, Sejersted OM, Bahr R, Woods JJ & Bigland-Ritchie B (1988). Motor drive and metabolic
responses during repeated submaximal voluntary contractions in man. Journal ofApplied Physiology
64,1421-1427.
Walsh ML & Banister EW (1995). Acute ventilatory responses to ramp exercise while breathing hypoxic,
normoxic, or hyperoxic air. In: Semple SJG, Adams L (eds.), Advances in Experimental Medicine and
Biology: Modelling and Control of Ventilation, pp. 00-00. New York: Plenum Publishing, In press.
Walsh ML & Bigland-Ritchie B (1995). A unitary model of fatigue and exhaustion. Abstracts, International
Symposium on Neural and Neuromuscular Aspects ofMuscle Fatigue (Miami, FL, November 10-13,
1994, pp. 27. Miami, FL: Miami Project to Cure Paralysis, University of Miami.
Weiner MW, Moussavi RS, Baker AJ, Boska MD & Miller RG (1990). Constant relationships between force,
phosphate concentration, and pH in muscles with differential fatigability. Neurology 40, 1888-1893.
Westling G, Johansson RS, Thomas CK & Bigland-Ritchie B (1990). Measurement of contractile and electrical
properties of single human motor units in response to intraneural motor axon stimulation. Journal of
Neurophysiology 64, 1331-1338.
Woods JJ & Bigland-Ritchie B (1983). Linear and non-linear surface EMG/Force relations in human muscles:
An anatomical/functional argument for the existence of both. American Journal ofPhysical Medicine
62, 287-299.
Woods JJ, Furbush F & Bigland-Ritchie B (1987). Evidence for a fatigue-induced reflex inhibition of
motoneuron firing rates. Journal ofNeurophysiology 58, 125-136.
Yan S, Lichros I, Zakynthinos S & Macklem PT (1993). Effect of diaphragmatic fatigue on control of
respiratory muscles and ventilation during C02 breathing. Journal of Applied Physiology 75,
1364-1370.
Zijdewind I & Kernell D (1994). Fatigue associated EMG behavior of the first dorsal interosseous and adductor
pollicis muscles in different groups of subjects. Muscle & Nerve 17,1044-1054.
Zijdewind I, Kernell D & Kukulka CG (1995) Spacial differences in fatigue-associated EMG behaviour of the
human first dorsal interosseous muscle. Journal of Physiology (London) 483, 499-509.
SECTION VIII
Integrative Systems Issues

One aspect of fatigue that is sometimes ignored by neurobiologists in their rush to

examine smaller and smaller units within the neuromuscular system (or even within a single
muscle cell) is how the whole body uses oxygen and delivers it to the mitochondria. One
potent method to examine this integrative issue is by assessment of how the problem has
been solved in different parts of the evolutionary tree. In Chapter 28 (Lindstedt & Hoppeler)
oxygen delivery is quantified from the muscle back to the lung, and ultimately to the central
respiratory controllers that generate the rate and depth of breathing. Design in oxygen
delivery via the cardiorespiratory system seems to be driven by the necessity for consumption
of oxygen by the muscle fiber. Some intriguing design problems have occasionally been
encountered in evolution; one ofthem is how the humming bird uses so much oxygen without
trading too much of its myofibrillar space for the necessary mitochondria.
Next (Chapter 29), Dempsey and Babcock apply some of the principles of the
preceding chapter to consider the likelihood of a ventilatory limit to normal human exercise.
Fortunately for us, but not the thoroughbred racehorse, our apparatus for oxygen delivery is
overbuilt and a respiratory limit is achieved in only a few elite athletes. Studies reviewed in
this chapter have revealed an important effect of whole-body exercise on the development
of fatigue in the diaphragm, a muscle with much evolutionary and survival value. Whether
this fatigue is mediated by suboptimal perfusion of the diaphragm, when the demands for
blood flow are increased elsewhere, remains to be determined.
Does fatigue of respiratory muscles playa major role in patients with respiratory
disease? This question has long been controversial despite a realization that the diaphragm
has adaptations which suit it for endurance, such as a high capillary and mitochondrial
supply. Chapter 30 (McKenzie & Bellemare) reviews this area. Appropriate techniques,
such as phrenic nerve stimulation, have been applied to assess the extent to which a failure
to produce force by the diaphragm is mediated peripherally or centrally by the eNS.
However, given what we know about fatigue of limb muscles during voluntary exercise, it
would be naive to think that fatigue would occur at only one of these two sites; or, that there
was no interaction between them.
Finally, morphometric and histochemical specializations exhibited by the muscles
involved in chewing and speech are described (Chapter 31, Nordstrom & Miles). Human
jaw-closing muscles exhibit extraordinary strength and their endurance in sustained volun-
tary contractions is probably high and limited as much by pain as by peripheral fatigue.
Interestingly, the fibers in jaw-closing muscles are smaller than those in limb muscles, which
would be advantageous for oxygen delivery.
382 Integrative Systems Issues

Overall, the concept emerges that while individual muscles may be specialized for
particular physiological tasks, their performance may sometimes be compromised by the
need to use other muscles. In this way, the design features of some muscles are subservient
to factors that must have influenced the evolution of the whole animal. Understanding the
fatigue properties of every muscle studied in isolation will not necessarily reveal how fatigue
will develop in a coordinated task that involves multiple muscle groups.
 28

FATIGUE AND THE DESIGN OF THE

RESPIRATORY SYSTEM

S. L. Lindstedt l and H. HoppelerZ

1 Department of Biological Sciences

Northern Arizona University
Flagstaff, Arizona 86011-5640
2 Department of Anatomy
University of Bern
CH-3000 Bern, Switzerland

ABSTRACT

One source ofmusc1e fatigue may be the failure to provide the required oxygen by any step
in the oxygen transport cascade or alack of the necessary machinery to utilize that oxygen.
We favor abandoning the concept of a single rate-limiting step for the concept of tuned
resistors, each contributing to the overall resistance to oxygen flow. However, because some
of these steps have considerably less phenotypic plasticity than others, these are the
component parts of the respiratory system that must be built with adequate "reserve" to
accommodate adaptive increases in the other steps (Lindstedt et ai., 1988; Weibel et ai., 1992;
Lindstedt et ai., 1994). These structures will usually appear to be over built except in those
rare individual animals at the species-specific limit of V0, in which these less malleable
structures may be limiting.

INTRODUCTION

When animals perform the vast majority of the tasks encountered during their lives,
fine motor control is generally more critical than are extremes of force generation, power
output or duration of muscular activity. Hence, it may be rare when an animal's performance
is limited by inadequate force generation by its muscles. However, there will be times in all
animals' lives in which one of these extremes of performance is necessary; in those instances,
muscular fatigue defines the limit of performance and may be of survival value. We adopt
the definition of fatigue as a failure to maintain the required or expected force.
Our focus is on those factors relating to the delivery and utilization of oxygen and
hence we restrict our discussion to aerobically functioning muscles only. This restriction is
not intended to imply that fatigue occurs only in these muscles, rather that when muscles are
active over extended durations (i.e., durations beyond which the on-board fuel can supply
the required ATP anaerobically), an additional potential source offatigue is introduced that

383
384 S. L. Lindstedt and H. Hoppeler

trained
c
o
°a I
E
::J
I
IJJ
C
I
o
u I
I
c
<Il I I
Cl
>. I I
x I
o I
I I

V
I
lactate V
I

Power

Figure 1. Maximum oxygen uptake (VO,max) is defined as the plateau ofVo, independent of increasing power
output. Because additional power output above that point is not fueled aerooically, it must be accompanied by
an abrupt increase in lactate production. Endurance training results in an increase in V° max and a shift to the
right of the power output at which lactate levels increase. 2

resides outside the neuromuscular machinery itself. When the supply of oxygen is inadequate
to meet the demand for oxidative phosphorylation, the muscle will be compromised in its
ability to produce force. How often does this occur and is there an identifiable weak link in
the delivery and utilization of oxygen to fuel muscular activity? These questions form the
focus of this review.
Aerobic fatigue may take two different forms. First, when an animal exercises, its
whole body oxygen uptake increases greatly. In average mammals, this increase is equal to
about ten times the resting or basal oxygen uptake and it may be much greater in some athletic
species. However, in all animals there is a point at which an increase in power output (e.g.,
running, swimming or flying speed) is no longer accompanied by an increase in oxygen
uptake. This plateau of oxygen consumption is defined as Vo max (Fig. 1). When exercise
involves power output (or force generation) at or beyond this p~int it can be maintained for
short durations only. However, this point of aerobic fatigue is adaptable; i.e., it is a
phenotypically plastic trait that will respond to, for example, endurance training. As more
oxygen is supplied and utilized by the musculature, the point of aerobic muscle fatigue is
shifted upward or downward in detraining (Fig. 1; see also Gordon, Chapter 32).
Aerobic fatigue can take another form when high levels of muscular activity are
required over extended periods of time. The sustainable level of O2 supply and utilization
decreases in whole body exercise as a function of the duration of exercise. Hence, V0 ,max
may be maintained for a few minutes only while longer durations of exercise are coupled to
ever decreasing fractions of V0 max' as is demonstrated in the speed and duration of human
running records (Fig. 2). We ~ill briefly address this interesting form of aerobic fatigue,
which likely is related to the supply offuel rather than oxygen delivery per se.
To examine these aspects of fatigue in whole body exercise, we exploit two natural
models and an experimental one (see Weibel et aI., 1992). Across the range of mammalian
body sizes, both basal and maximum oxygen uptake vary predictably with body size. Thus,
allometric variation provides a range of about 20 fold in (weight specific) V0 1max values
Fatigue and the Design of the Respiratory System 385

40

20

15min 30min 1hr 2hr

Time

Figure 2. World record standings in running (men). Sustained power output decreases with the length of time
the power is maintained, as is seen in current world record performances in men's track events. While the
sprints are entirely anaerobic, all distance events require aerobic resynthesis of ATP.

comparing a 2 g shrew with a 5000 kg elephant. Comparing athletic and sedentary animals
of the same body size, there is a 2.5 to 5 fold range of Vo,maX' the consequence of adaptive
variation, that can provide additional insights into possible design constraints of the
respiratory system. Finally, we draw upon the 50% increase in Vo max that can be achieved
within a given animal in response to specific training. This induc~d variation provides the
final experimental model discussed below.
In this chapter, we define the respiratory system broadly, as all of the structures
through which oxygen flows en route from atmospheric air to the mitochondria of exercising
muscles. Because it is only at VO,max when these structures are used to their maximum
capacities, we focus on this single physiological state to examine the structure-function links
in the respiratory system. As a working hypothesis, we assume that at each step, the capacity
to transport oxygen should be quantitatively tuned to the demand for oxygen of the working
muscles (Fig. 3). We then ask when and under what conditions is the supply of oxygen
inadequate; are there consistent or occasional weak links in the cascade of respiratory system
structures?

O 2 UPTAKE IN THE LUNGS

The transfer of O2 from alveoli to blood in the lungs is driven by the partial pressure
difference and the conductance for O2 (D L02 ) of the intervening tissue of the gas exchanger
(line I in Fig. 3). DL02 in tum can be decomposed into two sequential conductances, the lung
membrane diffusing capacity, DM o2 , and the erythrocyte diffusing capacity, De02' DM02 is

v~ = (~- P~)· DL~

Figure 3. As oxygen flows through the respiratory system, the v~ =(Co02- Cvo2)' 6
oxygen flow (Vo,) through each of the components must be equal.
If anyone or all of these structures is unable to supply oxygen at Vo2 =( Pbo2 - Pc(2)' DTo2
a rate equal to its demand in the working muscles, the muscle will
be unable to maintain force . We consider each step independently
below.
386 S. L. Lindstedt and H. Hoppeler

determined by structural parameters, specifically the total lung volume as well as the density
and characteristics of the gas exchange units packed into the lungs. The variables that may
modulate DM02 are therefore, the lung volume, the packing density of the alveolar surface,
the capillary loading of the alveolar surface and the mean harmonic thickness of the barrier
separating the air from the blood. The lung volume is a relatively constant fraction of body
size ranging from 35 ml·kgo! in a 3 g shrew to 60 ml·kg o! in large mammals such as humans
or steers. With adaptive variation, athletic species such as horses and pronghorn antelopes
may reach considerably larger weight-specific lung volumes (l00 and 200 ml . kg'! ,
respectively). The packing of alveolar septa, alveolar surface density, is relatively constant
(400-500 cm2·cmo3 ) in animals weighing between 1 and 100 kg. In larger animals, such as
horses and steers, the alveolar surface density falls to 250 cm2·cmo3 while in shrews it is
increased to 1500 cm2·cmo3 ). The packing density of the alveolar surface is similar in athletic
and sedentary species of the same size and probably does not change with training (Weibel
et aI., 1992). There is a weak allometric variability of capillary loading ofthe alveolar surface
as well as of the harmonic mean barrier thickness both of which increase with body mass
(Mb0.05). Capillary loading varies from 0.7 to 1.7 ml·mo2 from shrews to steers while barrier
thickness varies from 0.27 to 0.65 /lm in the same species. The erythrocyte diffusing capacity,
characterized by the rate of oxygen uptake of the pulmonary capillary unit, is dependent on
both body size and athletic capability. It ranges from 42 ml02min°[Link]! in animals weighing
500 g to 12 ml O2 . min°[Link]! in animals weighing 250 kg, athletic animals having higher
De02 values than their weight matched sedentary counterparts.
Intuitively one might expect total pulmonary diffusing capacity, D L02 , to vary
directly with 'if02max in allometric and adaptive variation as well as possibly in response
to induced variation. However, the experimental evidence is at odds with intuition as
DL02IMb varies with Mb oo .!!; a 500 kg animal has 3 times as much lung diffusing capacity
than a 50 g animal. Likewise, athletic species have a larger DL02 1Mb than sedentary
species of the same body weight; however, the difference in DL02 does not match the
difference in 'if02max' As•
a consequence, the driving force, ~P02' is larger both in small
animals (Lindstedt, 1984) as well as in highly active species (Weibel et aI., 1992). The
partial pressure difference between the alveolar air and the lung capillary blood depends
on the alveolar ventilation as well as on the perfusion of capillaries by the circulation.
In this context, the capillary transit time (tc) is an important variable indicating the time
available for loading of O 2 onto red blood cells. Combining morphological estimates for
lung capillary blood volume and measurements of cardiac output at 'ifO,max' the available
transit time is sufficient for full saturation of blood during its passage through the lungs
in all but the smallest mammals (Lindstedt, 1984; Weibel et aI., 1992). However, athletic
species such as dogs use up to 80% of the available transit time; whereas, sedentary
species such as goats may reach full saturation after less than 50% of tc' In some highly
athletic species such as thoroughbred horses (Jones et aI., 1989) and in top human
endurance athletes (Dempsey et aI., 1984), there is evidence for a drop of arterial P0 2
and oxygen saturation of arterial blood during maximal aerobic work indicating that the
lungs may limit aerobic performance.
In summary, the gas exchanger of the lung is overbuilt with regard to the functional
requirements during maximal aerobic whole body exercise in all mammals except for a few
exceptional aerobic performers. Moreover, there seems to be a lack of epigenetic malleability
of the lungs with stimuli that increase 'if02max such as endurance exercise training or chronic
cold exposure (Weibel, personal communication). We hypothesize that a consequence of
athletic training in humans is an increase in 'if02max until the diffusion capacity of the lungs
is reached, at which point arterial oxygen pressure and saturation begin to drop, limiting
oxygen availability to the periphery.
Fatigue and the Design of the Respiratory System 387

CONVECTIVE TRANSPORT OF O 2 BY THE HEART

The transport of oxygen to the periphery by the heart is described by the Fick equation
(line 2, Fig. 3). The heart as a pump has conventionally been thought as the major determinant
of aerobic performance capacity, limiting •'V0 2max in most species and certainly in humans
(Saltin & Strange, 1992). Cardiac output, Q, is the product of the functional variable, heart
rate (fH), and the structural variable, stroke volume (V s), the latter being directly proportional
to heart volume or mass. The arteriovenous oxygen concentration difference depends on the
partial pressure of O2 in the arterial and mixed venous blood, the hematocrit and the O 2
capacitance of the blood (assumed to be identical in mammalian species). Heart size is a
similar fraction of close to 0.6% of the body mass in all mammals (Stahl, 1967). As a
consequence, the stroke volume at 'V0 max is found to be invariant with body size (Fig. 4).
Likewise, hematocrit at 'V0 2max is similar in mammals of all sizes. However, both stroke
volume and hemoglobin concentration are consistently increased in athletic relative to
sedentary species. Furthermore, endurance training increases stroke volume but not usually
hemoglobin concentration. Because both heart size and hemoglobin concentration are
invariant with body size, it is obvious that allometric differences in circulatory oxygen
transport must entirely be brought about by varying heart rate. Resting heart rates vary with
basal metabolic rate and relate to Mb·0 .25 (Stahl, 1967; Fig. 4). Heart rate at 'V0 2max decreases
at a somewhat more shallow slope (Mb·o.17), because small animals can increase their heart
rate less during heavy exercise than large animals. In adaptive pairs, the maximal heart rate
is similar while athletic animals may have lower resting heart rates. The same is observed
with endurance exercise training in humans which increases stroke volume and decreases
resting heart rate.

r-------~-------r------_r------_,--------r_------T10 ~

VslMb = 2.49· MbO = 0.4· MH/Mb a

S
-- ------_K ___ ---- --------!-----!----1- C5
C
3
CD

--s:
~

0"

t
'E
1000 0.1 :f
;;t
.S
~
~
~
r::::
(])
100 fHrest = 241·Mb·O.25
::J

I •
(Stahl 1967)
•
ffi
(])
[Link] rat
dog
goat
guinea pig
pony
calf
horse
steer
I 10+-------~------~-------+------~r_------+_------~
0.001 0.01 0.1 1.0 10 100 1000
Body mass Mb in kg
Figure 4. Stroke volume (Vs) is a nearly constant fraction of body mass in all mammals while heart rate (fH)
falls as a regular function of body mass. As a consequence, cardiac output, the product of Vs and fH' is
systematically higher in small than large animals relative to body size. Quantitatively, this relation is identical
to that ofthe metabolic oxygen requirement.
388 S. L. Lindstedt and H. Hoppeler

The question as to whether cardiac output limits aerobic exercise performance can
be answered negatively; however, it may be the wrong question to ask in the first place. One
might rather ask how much of the resistance in the pathway of oxygen from lung to skeletal
muscle mitochondria is related to convective oxygen transport by the heart under a given set
of conditions (see di Prampero, 1985). Experiments using blood doping as an experimental
tool to vary oxygen supply to the periphery acutely and independently from peripheral
oxygen demand indicate that there is considerable resistance residing in the microcirculation,
the transfer of oxygen to mitochondria and/or within the metabolic machinery in mitochon-
dria (Turner et aI., 1993).

MICROVASCULAR O2 DELIVERY TO MITOCHONDRIA

02 flow from capillaries to mitochondria is functionally determined by the pressure

head between average skeletal muscle capillary partial pressure and the 02 partial pressure
in the immediate vicinity of mitochondria believed to be as low as a few torr, and the
diffusing capacity of the intervening tissue (Fig. 3, line 3; Gayeski & Honig, 1986). The
tissue diffusing capacity, DT 02, depends on a number of variables such as the capillary
blood volume, the capillary hematocrit, the capillary and sarcolemmal diffusion surfaces,
the average distance from the capillary to mitochondria, and the influence that the presence
of myoglobin might have on the 02 conductance of myofibers. There is, currently, no
comprehensive model available that integrates these variables into a model for calculating
skeletal muscle DT 02. The design of the microvascular system will, therefore, be described
by relating its main structural parameters to mitochondrial oxygen demand. Capillary
diameter is between 4.1 and 4.6 /lm with no systematic variation due to body size or
athletic ability. The key structural parameter, therefore, is capillary length J(c). Capillary
length can be calculated from counts of capillary profiles on tissue sections with appro-
priate morphometric methods. Once capillary length is known, capillary volume (relevant
for mass flow of gases and solutes) and capillary surface area (relevant for transcapillary
transport phenomena) can easily be obtained given the constant capillary diameter.
Capillary length (and hence capillary surface and volume) per unit mass decreases with
increasing body mass with a scaling exponent of -0.11. In adaptive variation, active
species have a larger weight-specific capillary length than inactive species. This difference
is increased by active species having a higher hematocrit at \fO,max than inactive species
as discussed previously. The higher hematocrit of active species is, therefore, critical for
diffusive transport of 02 both in the lungs and in muscles as well as for convective
transport of 02 in the circulation. When both capillary volume and capillary blood flow
are known, mean capillary transit time can be calculated; skeletal muscle capillary time
increases with increasing body size from 0.4 s in rats to 0.8 s in horses with no systematic
difference between active and inactive species. Time for unloading in the periphery is
about double that for loading in the lungs and is likely sufficient for unloading in all but
the smallest mammals (Lindstedt & Thomas, 1994). Capillary density and hence J(c)
increases with endurance exercise training (Hudlicka, 1985), such that there is a constant
proportion relationship between capillary length, (characterizing oxygen supply) and
mitochondrial volume (characterizing oxygen demand) in skeletal muscle tissue (Fig. 5).
The question as to how resistance to oxygen flow over the sequential steps between
erythrocytes and mitochondria must be attributed cannot presently be answered. It is unclear,
therefore, whether capillary, mitochondria, or the tissue between them is the more likely site
of tissue diffusion limitation.
Fatigue and the Design of the Respiratory System 389

100000

10000

1000

~ 100

10
Figure S. Total capillary length (J(c» • HEART
varies systematically and linearly with .. DIAPHRAGM
total mitochondrial volume (V(mt» in • LOCOMOTOR MUSCLES
heart, diaphragm and locomotor muscles
among the mammals. This linkage sug- 0.1 = 13.31 . x 102
y
gests that capillaries and mitochondria = 0.99
r2
form a single functional unit in skeletal 0.01 +----r---,-----,--.----.----,-----,
muscle, mitochondria setting the de- 0001 001 01 1 10 100 1000 10000
mand and capillaries, the supply. V(mt) cm 3

MITOCHONDRIAL OXYGEN SINK

The mitochondria are solely responsible for the resynthesis of ATP by oxidative
phosphorylation. Therefore, if the amount of mitochondrial machinery in the muscle is
inadequate to meet the ATP demand by the myosin and sarcoplasmic reticulum ATPases, the
muscle will be unable to maintain its force production (line 4, Fig 3). The source of muscle
fatigue in this instance can be directly and quantitatively linked to the total volume of
mitochondria within the skeletal muscle. Among mammals, the density of inner mitochon-
drial membrane per unit volume of mitochondria is relatively invariant at 35 J-lm 2 • m- 3
(Schwerzmann et al., 1989). Thus, the total volume of mitochondria within any single muscle
or muscle group is a reliable marker of the total quantity of respiratory chain enzymes within
the muscle (Weibel et al., 1992). Because 90-95% of the oxygen consumed at the lungs at
V0 max is destined for the skeletal muscle mitochondria, the muscle mitochondria are a single
furtctional oxygen sink during exercise. We would expect the size of this sink to be not only
directly proportional to Vo max; but indeed, it is the very source of the oxygen demand. It is
the mitochondria that ultirhately set the demand for oxygen, the upstream structures dis-
cussed above must be subservient to that demand.
We have been interested in quantifying the size of the sink for the past decade
(Hoppeler and Lindstedt, 1985). When V 0 2max is expressed as a function of the size of the
oxygen sink, the total skeletal muscle mitochondrial volume, these two variables are linked
via a consistent quantitative coupling, each ml of mitochondrial volume consumes roughly
4.5 ml of O2 per minute. Recently, this relationship has been expanded to include other
vertebrates in addition to mammals. Ifhummingbirds operated with the same mitochondrial
oxygen uptake as mammals, their flight muscles would have to be composed of two-thirds
mitochondria and only one-third myofibrils! To solve this problem, hummingbirds have
managed a trick that is likely unique among the vertebrates; they have packed twice the
density of mitochondrial membranes in their mitochondria. As a consequence, mitochondrial
oxygen uptake in these animals is identical to that of other vertebrates when expressed per
unit mitochondrial membrane normalized for differences in body temperature (Wells, 1990;
Suarez, 1991; Schaeffer & Lindstedt, 1992) and the volume density of the mitochondria
remains around one-third. Finally, there is evidence to suggest that mitochondria are among
390 S. L. Lindstedt and H. HoppeJer

the first structures to respond to endurance training within an individual animal (e.g.,
Reichmann et aI., 1985).
In summary, mitochondria (and the capillaries supplying them) respond adaptively
to shifts in the aerobic functioning of muscles. For that reason, we propose that the density
of mitochondria (i.e., the size of the oxygen sink) will reflect the current aerobic demand of
the muscle. Such a proposal does not implicate this single step as the rate limiting step in
the oxygen transport cascade; rather, that mitochondrial volume is quantitatively tuned to
the ATP demand of aerobically functioning muscles.

ACKNOWLEDGMENTS

The authors thank Profs. C. Richard Taylor and Ewald Weibel for their numerous
contributions. The authors' laboratories have been supported by United States Public Health
Service grant HL 41986, and National Science Foundation grant IBN 9317527 (S.L.L.), and
the National Science Foundation of Switzerland (HH). Attendance of the authors at the
1994 Bigland-Ritchie conference was supported, in part, by the Muscular Dystrophy
Association (USA; HH) and the University of Arizona Regents' Professor funds of Douglas
Stuart (S.L.L.).

REFERENCES
Dempsey JA, Hanson PG & Henderson KS (1984). Exercise-induced arterial hypoxia in healthy human
subjects at sea level. Journal of Physiology (London) 355, 161-175.
Dempsey JA, Johnson BD & Saupe KW (1990). Adaptations and limitations in the pulmonary system during
exercise. Chest Supplement 97, 81S-87S.
di Prampero PE (1985). Metabolic and circulatory limitations to Va,max at the whole animal level. Journal of
Experimental Biololgy 115, 319-33l.
Gayeski TEJ & Honig CR (1986). O2 gradients from sarcolemma to cell interior in red muscle at maximal
Va,max' American Journal of Physiology 251, H789-H799.
Hoppeler H & Lindstedt SL (1985). Malleability of skeletal muscle tissue in overcoming limitations: Structural

°
elements. Journal of Experimental Biology 115, 355-364.
Hudlicka (1985). Development and adaptability of microvasculature in skeletal muscle. Journal ofExperi-
mental Biology 115, 215-228.
Jones JH, Taylor CR, Lindholm A, Straub R, Longworth KE & Karas RH (1989). Blood gas measurements
during exercise: errors due to temperature correction. Journal ofApplied Physiology 67,879-884.
Lindstedt SL (1984). Pulmonary transit time and diffusing capacity in mammals. American Journal of
Physiology 246, R384-R388.
Lindstedt SL & Thomas RG (1994). Exercise performance in mammals: An allometric perspective. In: Jones,
JH (ed.) Advances in Veterinary Science and Comparative Medicine 38B, pp 325-348. New York:
Academic.
Lindstedt SL, Thomas RG & Leith DE (1994). Does peak inspiratory flow contribute to setting Va,max? A test
of symmorphosis. Respiration Physiology 95, 109-118.
Lindstedt SL, Wells DJ, Jones JH, Hoppeler H & Thronson HA, Jr (1988). Limitations to aerobic performance
in mammals: Interaction of structure and demand. International Journal of Sports Medicine 9,
210-217.
Reichmann H, Hoppeler H, Mathieu-Costello 0, von Bergen F & Pette D (1985). Biochemical and ultrastruc-
tural changes of skeletal muscle mitochondria after chronic electrical stimulation in rabbits. Pflugers
Archiv 404, 1-9.
Saltin B & Strange S (1992). Maximal oxygen uptake: "old" and "new" arguments for a cardiovascular
limitation. Medicine and Science in Sports and Exercise 24, 30-37.
Schaeffer P & Lindstedt S L (1992). Structure function coupling in the fastest contracting vertebrate muscle:
the rattlesnake tail shaker muscle. The Physiologist 35,224 (abstract).
Fatigue and the Design ofthe Respiratory System 391

Schwerzmann K, Hoppeler H, Kayar SR & Weibel ER (1989). Oxidative capacity of muscle and mitochondria:
Correlation of physiological, biochemical, and morphometric characteristics. Proceedings oj the
National Academy oJSciences 86,1583-1587.
Stahl WR (1967). Scaling of respiratory variables in mammals. Journal ojApplied Physiology 22,453-460.
Suarez RK (1991). Oxidative Metabolism in Hummingbird Flight Muscles. In: Bicudo JEPW (ed.), The
Vertebrate Gas Transport Cascade, pp. 279-285. Boca Raton: CBC Press.
Turner DL, Hoppeler H, Noti C, Gurtner HP, Gerber H, Schena F, Kayser B & Ferretti G (1993). Limitations
to V02max in humans after blood retransfusion. Respiration Physiology 92, 329-341.
Weibel ER, Taylor CR & Hoppeler H (1992). The concept of symmorphosis: A testable hypothesis of
structure-function relationship. Proceedings oJ the National Academy oj Sciences, USA 88,10357-
10361.
Wells DJ (1990). Hummingbird Flight Physiology: Muscle Performance and Ecological Constraints. Ph.D
Dissertation, Univ. Microfilm Int. Cit. No. 9105412. Laramie: University of Wyoming.
 29

AN INTEGRATIVE VIEW OF LIMITATIONS

TO MUSCULAR PERFORMANCE

1. A. Dempsey and M. A. Babcock

The John Rankin Laboratory of Pulmonary Medicine

University of Wisconsin-Madison
Department of Preventive Medicine
Madison, Wisconsin

ABSTRACT

First we describe the changing site of limitation to maximal O2 transport with increasing
fitness in mammals. The capacity for diffusion and airway/parenchymal flow rate and
volume are markedly overbuilt in the sedentary subject's lung, but undergo little change with
increased training/fitness; accordingly, as demand for O 2 transport increases in the highly
fit, the limits for maximal diffusion and ventilation are surpassed or met at maximal exercise.
Secondly, low-frequency diaphragmatic fatigue occured with by heavy endurance exercise.
This fatigue resulted from increased diaphragmatic work together with the major contribu-
tion from the secondary effects of increased locomotor muscle activity; namely, metabolic
acidosis and increased requirement for blood flow.

INTRODUCTION

The physiologic mechanisms responsible for limiting muscular performance have

long been debated. In exercise involving large muscle mass, oxygen delivery to locomotor
muscle capillaries is the dominant key determinant of locomotor muscle performance and
the heart's maximal stroke volume and cardiac output are the major limiting factors to sustain
O2 delivery (Andersen & Saltin, 1985; Wagner et aI., 1991). So this is the general concept;
but in an integrated physiological system, there are many exceptions even in health to these
generalizations.
We make two points concerning limitations to exercise to underscore the complexity
of the integrative system. First, we illustrate using the respiratory system that the site of the
primary limitation to oxygen transport and to performance of healthy subjects is changeable
because the capacities of the different organ systems adapt at different rates to physical
training and to the aging. Secondly, we show that the incurred fatigue in respiratory skeletal
muscles during whole-body exercise depends not only on the amount of mechanical work
performed by the respiratory muscles themselves but also on the milieu created by the limb

393
394 J. A. Dempsey and M. A. Babcock

locomotor muscles. The occurrence of clinical diaphragmatic fatigue is discussed by McKen-

zie and Bellemare, Chapter 30.

CHANGING SITES OF LIMITATION TO MAXIMAL O 2

TRANSPORT

As the external work rate is increased, total \1 0 , increases linearly and then begins to
plateau with further increases of work rate heralding the attainment of maximal \10
(\1 o,maJ. The determinants of \1 02 max are classically described by the Fick equation: 2

\102 max = {maximal cardiac output (stroke volume * heart rate)} *{arterial O2 content-
mixed venous O2 content}

The product of cardiac output and arterial O2 content define the O2 delivery and the
magnitude of the arterial to mixed venous O2 content difference measures the degree of O2
extraction, primarily by the locomotor muscles. At \1 0 max the maximal muscle blood flow
and maximal arteriovenous oxygen difference (a-vO;) are achieved simultaneously and
plateau with the \10, so it is difficult to distinguish between the relative importance of these
functions as limiting factors based purely on these correlative data. Additional data strongly
implicate O2 delivery to the muscles and in turn the maximal stroke volume as the critical
limiting factors (Andersen & Saltin, 1985). In most mammals arterial O2 content is main-
tained near resting levels throughout exercise. Accordingly, much has been made of the
substantial reserve enjoyed by the healthy pulmonary gas exchanger and its overbuilt nature
in terms of several characteristics:
I. Large diffusion surface area and minimal diffusion distances,
2. Large capacity of pulmonary capillary blood volume with respect to the maximal
velocity of pulmonary blood flow,
3. The capacity of the lymphatic drainage system of the lung would far exceed the
increased turnover of extravascular lung water during exercise,
4. The distribution of mechanical time constants (resistance x compliance) through-
out the airways is sufficiently uniform so that \1Aand \1 A/Q distributions remain
relatively uniform even at the high rates of perfusion and shortened inspiratory
times present during maximal exercise,
5. The high elastic recoil in healthy lungs allows the airways to maintain patency
and avoids flow limitation during forced expirations at high levels of hyperpnea,
6. The capacity of the inspiratory muscles to generate negative intrapleural pressure
far exceeds that demanded by the normal maximal ventilation.
This is a situation in the normal human in whom most data has been generated and
from whom generalizations are made concerning performance limitations. However, the
critical limiting factor is not fixed under all circumstances and as maximal work capacity
and \1 0 max change, critical determinants of gas transport will also change.
We have examined physical training/fitness and healthy aging as adaptive processes
during which the relative importance of the pulmonary system to maximal gas transport
changes. Achieving greater than normal aerobic capacity via physical training and/or greater
genetic endowment requires that appropriate increases in capacity occur at those links in O2
transport which were previously limiting factors. Thus, maximal stroke volume and cardiac
output are increased as are muscle capillarization and locomotor muscle oxidative capacity.
An Integrative View of Limitations to Muscular Performance 395

Whether the lung and chest wall continue to be overbuilt depends upon their relative
adaptabilities as demand in the form ofVo2max is altered.

Alveolar to Arterial O2 Transport

In mammals, some aspects of lung and chest wall structure and function clearly lag
far behind cardiovascular and skeletal muscle adaptations to increased energy demands while
others appear to parallel the increasing 02max. For example, in many highly fit humans,
with V0 2max 40 to 50% or more greater than normal, arterial P02 is no longer maintained at
resting levels (Dempsey et aI., 1984). The ultimate expression of this imbalance between
demand vs. capacity in pulmonary gas exchange is the thoroughbred horse who because of
centuries of genetic engineering is capable of exercising at maximal work rates and O2
uptakes which are double that in the elite human athlete. These animals show substantial
exercise-induced hypoxemia (Bayly et aI., 1989). One key reason for these failures in
pulmonary gas exchange in heavy exercise in the highly fit is the structural limits to the
dimensions of the pulmonary vascular bed which are not changing with training even though
the systemic vasculature is highly adaptable. This results in a pulmonary capillary blood
volume which is not increased in the highly trained while pulmonary blood flow keeps
increasing - thus red cell transit time in the lung's capillaries is shortened substantially.
Furthermore, the high pulmonary blood flow also leads to high pulmonary arterial pressure
and pulmonary vascular resistance. In extreme cases of increased capillary perfusion
pressure it is speculated that the alveolar-capillary interface may actually undergo stress
failure leading to the accumulation of extravascular lung water during heavy exercise causing
diffusion limitation (West et aI., 1991).

Air Flow Limitation

Intrathoracic airway structure and lung elastic recoil are also not affected by train-
ing/fitness in the human. The highly trained individual retains the same susceptibility as the
untrained to airway closure during active, forced expiration and the area of their maximal
flow:volume envelope is also normal. Accordingly, as maximal metabolic rate increases and
V E rises proportionately, signification limitation to expiratory flow begins to occur in the
very highly trained human and ventilation reaches its maximal limit at maximal work rate
(Johnson et aI., 1992). Consequences of this increased flow limitation are to increase the
work and the O2 cost of breathing (Aaron et aI., 1992) which is offset to some extent during
submaximal exercise because the magnitude of the ventilatory response to maximal exercise
is reduced in the highly fit. Nevertheless, while the magnitude of the hyperventilatory
response to maximal exercise may be constrained slightly by mechanical limitations to flow
and volume in the highly trained, frank CO 2retention does not occur in the healthy, fit human,
although it does in the thoroughbred horse (Bayly et aI., 1989).

Respiratory Muscle Capacity

The capacity of inspiratory muscles to generate pressure at any given length or
velocity of shortening is only slightly affected by physical training/fitness. Accordingly,
because of the high ventilatory requirement at V0 2max' the pressure generating capacity of
the inspiratory muscles is also achieved in the highly fit (Johnson et aI., 1992). On the other
hand, the endurance capacity of inspiratory muscles may well adapt with increased train-
ing/fitness, as evidenced by the finding that even though the highly trained endurance athlete
does show clear evidence of diaphragm fatigue as a result of heavy endurance exercise, these
trained subjects are capable of sustaining much higher levels of ventilation and diaphrag-
396 J. A. Dempsey and M. A. Babcock

matic work during exercise for the same level of diaphragm fatigue as the untrained subject
(Babcock et aI., 1994).
These examples demonstrate how major determinants of maximal exercise capacity
will change with increased training/fitness, specifically because of the relative lack of
plasticity in the diffusion capacity of the lung, in the maximal airflow rates acceptable by
the airways and in pressure generating capabilities of the inspiratory muscles. In other words,
in the highly fit at VO,rnax demand on each of these sub-systems rises to reach capacity in the
case of expiratory flow rate, maximal ventilation and inspiratory muscle pressure generation
and to exceed capacity in the case of diffusion equilibrium at the level of the alveolar-cap-
illary interface. In the latter case, the resulting exercise-induced arterial hypoxemia presents
a significant limitation to VO,rnax.

RESPIRA TORY MUSCLE FATIGUE IN THE PHYSIOLOGICAL

ENVIRONMENT OF WHOLE BODY EXERCISE

Whole body endurance exercise at very high intensity causes diaphragmatic fa-
tigue, as shown by the 15-40% reduction in the trans-diaphragmatic pressure developed
in response to supramaximal phrenic nerve stimulation (BPNS) immediately following
exercise (Fig. 1). This reduction in diaphragmatic force 1) occurs at several different
lung volumes (i.e., diaphragmatic lengths); 2) occurs at stimulation frequencies of 1-20
Hz; 3) persists for 60-90 min following exercise; 4) is potentiated by hypoxia; 5) occurs
most consistently at endurance work loads that are greater than 85% of Vo2max (Fig. 2)
and lasting greater than 8-10 min; 6) is correlated with the relative intensity of the
exercise; 7) occurs to a similar extent in human subjects with a wide variety of fitness
levels from normal to twice riormal VOrnax (Johnson et aI., 1993; Babcock et aI., 1995a;
Babcock et aI., 1995b). 2

PRE-EXERCISE POST-EXERCISE
[90-95% \102ma.]
Pe

Rt. Di EMG

Lt. Oi EMG

Pdi __
/
,~----/-
Pg - - '

Figure 1. Typical responses of esophageal (P e), gastric (P g) and transdiaphragmatic (P di) pressures and
diaphragm EMG (i.e., M-wave) to 10 Hz bilateral phrenic nerve stimulation (BPNS) before and immediately
after whole body exercise at 90% V0 max' Transdiaphragmatic pressure (P di) was significantly reduced
following the whole-body exercise. The're was no change in the M-wave response of the diaphragm to the
supramaximal stimulation (modified from Johnson et aI., 1993).
An Integrative View of Limitations to Muscular Performance 397

10

)( 0
•
CD

"
C
-
CD
·10 •
::::s
~ ·20
Figure 2. This figure illustrates that the level of exer- .f
cise-induced diaphragm fatigue increased with in- E ·30
creased intensity of the endurance exercise (measured Cl
e •
as a percentage of 'iIo,maJ. Diaphragm fatigue index =
% change in Pdi during supramaximal BPNS at 1, 10 and
.c
a.
-40
•
20 Hz, pre- vs. post-endurance exercise. The exercise .!!
C
·50
•
time to exhaustion at 85% 'iIO,max was 26.0 ± 2.6 min
and at 95% 'ilo,max was 13.0 ± 1.7 min. Fitness level ·60 -t--,----,---,..--;---.---,----,
70
varied widely among the subjects, 'iIO,max was 58.7 ± 2.2 105

ml·kg- 1·min- 1 (range 39.8 to 78.6 ml·kg- 1·min- 1, n =24). Exercise Intensity ( % V02 max)

What causes this exercise-induced diaphragmatic fatigue? Is it dependent on the

amount of work done by the diaphragm? ... or do the milieu and/or demands created by the
locomotor muscles exert significant independent influences?
To determine the effects of diaphragmatic work, per se, on diaphragmatic fatigue we
had normal subjects voluntarily produce high levels ofnormocapnic ventilation and transdia-
phragmatic pressure while at rest (Babcock et aI., 1995b). Then we used the supramaximal
BPNS (Bellemare & Bigland-Ritchie, 1984; see also Gandevia & McKenzie, 1985) to
determine the change in maximal pressure generation by the diaphragm as affected by these
sustained hyperpneic trials. Two levels of diaphragmatic work loads were voluntarily
produced for 10 to 20 minutes, i.e., equal to the subject's whole-body exercise performance
time. We used visual feedback for volume and pressure and audio feedback for breath timing.
First, we mimicked the Vt and fp di • fb . (diaphragmatic work) achieved throughout whole
body endurance exercise. Secondly, the subjects voluntarily produced (for the same duration)
a 50-70% increase in the fp di . fb achieved during the exercise.
Two findings arose which pointed to a major effect of the whole body exercise, per se,
on diaphragmatic fatigue. First, we found (using BPNS) that a fatigue threshold for diaphrag-
matic work (in the resting subject) was present during sustained inspiratory efforts requiring an
increase in fp di . fb of 4 to 5 times resting levels (Fig. 3). This fatigue threshold also coincided
with the time:tension index of the diaphragm (TTdi) being less than 0.15 - a value previously
associated with task failure during voluntary inspiratory efforts against high resistive loads
(Bellemare & Grassino, 1982); or when the peak Pdi during the sustained hyperpnea exceeded
60-70% of the dynamic capacity of the diaphragm for pressure generation. This fatigue
threshold was only rarely achieved by the diaphragm during endurance exercise. Rather,
exercise-induced diaphragmatic fatigue most commonly occurred when the diaphragmatic
work was substantially less than the fatigue threshold as determined in the resting subjects (Fig.
3). We interpret these findings to mean that the whole-body endurance exercise itself must be
contributing significantly to the exercise-induced diaphragmatic fatigue.
Several exercise-related mechanisms appear feasible to explain this finding. Firstly,
extracellular metabolic acidosis created by locomotor muscles during whole body exercise
may induce intracellular acidosis in the diaphragm and promote fatigue. This is indicated
indirectly by the high levels oflactate accumulated in the rat diaphragm as a result of heavy
endurance exercise (Fregosi & Dempsey, 1986). Secondly, the high levels of diaphragmatic
work achieved during the hyperpnea do require high blood flows. Meeting this flow
requirement is not a problem during the voluntary hyperpnea trials at rest when the
398 J. A. Dempsey and M. A. Babcock

Figure 3. Relationship between the level of diaphragmatic pressure generating work as reflected by Jpdi • fb
and the level of diaphragm fatigue measured by supramaximal BPNS (mean % .[Link] from 1, 10 and 20 Hz.
BPNS). The hatched area represents the effects
10
of sustained voluntary hyperpnea (at rest) gen-
erating a continuum of Jp di • fb levels, on dia-
phragm fatigue. Individual results (n = 24)
from whole body endurance exercise are su-
perimposed (_) and show that in most cases !!. ·10
CD
during exercise the Jp di . fb was below the :::I
fatigue threshold for the diaphragm « 550 cm en
:0= ·20
H20 [Link]·]). The fact that significant dia- nI
LL.
phragmatic fatigue was found following whole
body exercise means that some factor is con- E -30 • •• • •
en
•
-a
I!
tributing to the fatigue in addition to the dia-
phragm work per se. In exceptional cases,
some individuals generate Jp di • fb levels nI
40
• ~Hyperpnea
• Exercise

greater than 550 cm H20 s·min·] throughout C .50 •

exercise and in these situations the diaphrag-
matic pressure generation by itself may be re- o 200 400 600 800 1000 1200
sponsible for all or much of the
exercise-induced diaphragm fatigue.

inspiratory muscles represent the major, unopposed requirement for increased perfusion.
However, it might be a problem during whole body exercise when the diaphragm must
compete with the locomotor muscles for the available cardiac output.
We cannot distinguish between these postulated mechanisms but we did determine
that the amount of work that non-locomotor (including respiratory) muscles do during
exercise is important in determining whether fatigue does occur. We used supramaximal
ulnar nerve stimulation to show that whole body exercise did not cause fatigue of the first
dorsal interosseous muscles. Thus, even though a resting skeletal muscle is exposed to the
circulating metabolic acids and must compete for blood flow during heavy exercise, these
factors alone will not cause fatigue. These data are consistent with recent findings in dogs
which show glycogen depletion by the active muscles during leg exercise is critically
dependent upon the intensity of contraction carried out by the muscle not primarily engaged
in locomotion (Gladden et aI., 1994). Furthermore, it has been shown that increasing
contraction intensity enhances lactate uptake by human forearm muscle during leg exercise
(Catchside & Scroop, 1993).
These data underscore the importance of the interaction occurring between the
independent effects oflocomotor muscle activity (which causes a competition for blood flow
and increased circulating metabolic acidosis) and the work produced specifically by the
diaphragm, as determinants of exercise-induced diaphragmatic fatigue. These data also
indicate that fatigue during whole body exercise is a multifaceted process which involves
responses beyond the primary muscles of interest. Furthermore, our first example concerning
changing sites of limitation to maximal O2 transport in the face of changing fitness levels,
emphasizes that the various determinants of transport are not equally malleable in the face
of changing performance capabilities. Accordingly, the major site of limitation may well
change radically, and even come to include the organ systems and mechanisms which at one
time were the most over-built.
In summary, our study epitomizes our gratitude to Brenda Bigland-Ritchie for her
many contributions to our understanding of muscle fatigue. We are especially grateful to
Brenda for her development of the BPNS technique - as a unique method for quantifying
diaphragmatic force development, in vivo.
An Integrative View of Limitations to Muscular Performance 399

ACKNOWLEDGMENTS

We wish to thank Gundula Birong for preparation of the manuscript and David
Pegelow for his invaluable technical assistance. The authors also wish to thank Dr. Simon
Gandevia for his helpful comments on the manuscript. The author's laboratory has been
supported by grants from United States Public Health Service, National Heart, Lung, and
Blood Institute. M.B. is a Parker B. Francis fellow. Attendance of J.A.D. at the 1994
Bigland-Ritchie conference was supported, in part, by the University of Arizona Regents'
Professor funds of Douglas Stuart.

REFERENCES
Aaron EA, Seow K, Johnson BD & Dempsey JA (1992). Oxygen cost of exercise hyperpnea: implications for
performance. Journal ofApplied Physiology 72, 1818- 1825.
Andersen P & Saltin B (1985). Maximal perfusion of skeletal muscle in man. Journal ofPhysiology (London)
366, 233-249.
Babcock MA, Johnson BD, Pegelow DF, Suman OE, Griffin D & Dempsey JA (1995a). Hypoxic effects on
exercise-induced diaphragmatic fatigue in normal healthy humans. Journal of Applied Physiology,
78(1),82-92.
Babcock MA, Pegelow D & Dempsey JA (1994). Aerobic fitness effects on exercise-induced diaphragm
fatigue. American Journal of Respiratory and Critical Care Medicine 149, A 799.
Babcock MA, Pegelow DF, Suman 0, McClaran SR & Dempsey JA (1995b). Contribution of diaphragmatic
work to exercise-induced diaphragm fatigue. Journal ofApplied Physiology 78, 1710-1719.
Bayly WM, Hodgon DR, Schulz DA, Dempsey JA & Gollnick PD (1989). Exercise-induced hypercapnia in
the horse. Journal ofApplied Physiology 67, 1958-1966.
Bellemare F & Bigland-Ritchie B (1984). Assessment of human diaphragm strength and activation using
phrenic nerve stimulation. Respiration Physiology 58, 263-277.
Bellemare F & Grassino A (1982). Effect of pressure and timing of contraction on human diaphragm fatigue.
Journal ofApplied Physiology 53, 1190-1195.
Catchside PG & Scroop GC (1993). Lactate kinetics in resting and exercising forearms during moderate-in-
tensity supine leg exercise. Journal ofApplied Physiology 74, 435-443.
Dempsey JA, Hanson P G & Henderson KS (1984). Exercise-induced arterial hypoxemia in healthy human
subjects at sea level. Journal of Physiology (London) 355, 161-175.
Fregosi RF & Dempsey JA (1986). Effects of exercise in normoxia and acute hypoxia on respiratory muscle
metabolites. Journal ofApplied Physiology 60, 1274-1283.
Gandevia SC & McKenzie DK (1985). Activation of the human diaphragm during maximal static efforts.
Journal of Physiology (London) 367, 45-56.
Gladden LB, Crawford R & Webster M (1994) Effect oflactate concentration and metabolic rate on net lactate
uptake by canine skeletal muscle. American Journal of Physiology: Regulatory, Integrative and
Comparative Physiology 35, R1095- RIIOI.
Johnson BD, Babcock MA, Suman OE & Dempsey, JA (1993). Exercise-induced diaphragmatic fatigue in
healthy humans. Journal of Physiology (London) 460, 385-405.
Johnson BD, Saupe KW & Dempsey JA (1992). Mechanical constraints on exercise hyperpnea in endurance
athletes. Journal ofApplied Physiology 73, 874-886.
Wagner PD, Hoppler H & Saltin B (1991). Determinants of maximal oxygen uptake. In: Crystal RG, West JB
(eds.), The Lung: Scientific Foundations, vol. 2, pp. 1585-1593. New York: Raven Press.
West JB, Tsukimoto K, Mathieu-Costello 0 & Prediletto, R (1991). Stress failure in pulmonary capillaries.
Journal ofApplied Physiology 70, 1731-1742.
 30

RESPIRATORY MUSCLE FATIGUE

D. K. McKenzie' andF. Bellemare2

IDepartment of Respiratory Medicine and Prince of Wales Medical

Research Institute
Prince of Wales Hospital and Division of Medicine
University of New South Wales, Sydney, Australia
2 Departement d' Anesthesie
Hotel-Dieu de Montreal
3840 St. Urbain, Montreal, Quebec, Canada

ABSTRACT

Ventilatory failure may accompany a variety of pulmonary and neuromuscular diseases.

There has been much controversy about whether this failure is due to respiratory muscle
fatigue at peripheral sites or a failure of drive at sites within the central nervous system. The
chapter reviews this topic.

INTRODUCTION

Is respiratory muscle fatigue an important cause of ventilatory failure? In 1919, the

renowned physiologists Davies, Haldane and Priestley noted that "no very definite investi-
gations have hitherto been made as to how the adaptation (of breathing to increased
resistance) is brought about and at what point it begins to fail". They performed a series of
experiments in which they subjected themselves to breathing through external resistances
of varying severity. With extreme resistances they noted an inability to maintain ventilation
with the onset of rapid shallow breathing and clinical features of hypoxaemia and hypercar-
bia (Fig. 1). They concluded that the respiratory centre had become fatigued and that the
development of hypoxaemia "hastened this failure of the respiratory centre". About sixty
years later Roussos, Macklem and colleagues (1977, 1979) performed similar experiments
on themselves but imposed a set breathing pattern and a target respiratory pressure. When
the target pressure exceeded a critical value the subject could not maintain the pressure and
ventilation decreased. These authors concluded that this failure resulted largely from fatigue
of respiratory muscles, especially the diaphragm, a conclusion which appeared to be
supported by a shift in the EMG power spectrum to lower frequencies (Gross et aI., 1979;
Bellemare & Grassino, 1982a; their Fig. 2).
The concept of respiratory muscle fatigue as a major contributor to ventilatory failure
subsequently gained wide acceptance and stimulated much research. Many pulmonary

401
402 D. K. McKenzie and F. Bellemare

iii I Iii iii I I Iii i I I I I Iii iii iii iii Iii I I I I iii I I I Iii I
IRon IRoff

Effects of resistance

Figure 1. A resistive breathing experiment with J.S Haldane as the subject. With excessive resistance,
respiratory rate fell initially from 27 to 12 per minute. Then tidal volume decreased and rate increased
progressively until "the resistance had to be removed on account of threatening symptoms". The authors noted
that the same resistance was tolerated by J.G. Priestley "without signs of fatigue ... (or) distress". (adapted
from Davies et aI., 1919).

rehabilitation units implemented specific training programs with the object of increasing the
strength and/or endurance of the respiratory muscles. However, in more recent years
attention has again focused on respiratory control and the possibility that failure of central
drive may be an inescapable consequence of the imposition of fatiguing loads to breathing.
A number of researchers are examining the possible interaction and roles of volitional and
involuntary central motor pathways to respiratory motoneuron pools and their relative
importance in situations of critical loading. This chapter reviews the evidence for failure at
central and peripheral sites both from human and animal studies (see also Dempsey &
Babcock, Chapter 29). It highlights some recent studies which have employed new tech-
niques in an attempt to resolve discrepancies. The first section outlines the clinical relevance
of these studies and defines some of the terminology.

THE DEVELOPMENT OF VENTILATORY FAILURE

Ventilatory failure may occur as a result of any neuromuscular or musculoskeletal

disorder which impairs the action of the ventilatory pump or as a result of insuperable loads
to breathing imposed by disease of the lungs or chest wall.
Chronic airflow limitation due to diffuse airway narrowing is the commonest cause
of respiratory failure. In these patients the work of breathing is increased substantially as a
result of increased resistance to breathing, impaired gas exchange which necessitates an
increase in minute ventilation, and the elastic loads imposed by hyperinflation of the lungs.
The hyperinflation also reduces the operating length of inspiratory muscles and decreases
the mechanical advantage of the diaphragm and rib-cage for producing volume expansion
(for review see Rochester, 199Ia,b).
The extent to which the inspiratory muscles of such patients adapt to these mechanical
loads has not been studied exhaustively. It is likely that length adaptation occurs by shedding
sarcomeres. It remains to be determined conclusively in humans whether the endurance
capacity of these muscles is enhanced by the chronic loading ofiung disease (McKenzie &
Gandevia, 1986; Newell et ai., 1989). Although the ability to develop inspiratory pressure
is impaired at high lung volumes, the endurance capacity of the muscles is well preserved
in relative terms (McKenzie & Gandevia, 1987; Clanton et ai., 1993). It has also been
proposed that adverse metabolic and nutritional consequences of advanced pulmonary
disease may impair the function of the inspiratory muscles (see Rochester, 1991a,b). There
is no information on the influence of the various consequences of airflow obstruction on
blood flow to the diaphragm.
Respiratory Muscle Fatigue 403

In view of all the factors which might compromise the function of the inspiratory
muscles, it was logical to conclude that the muscles would develop peripheral fatigue.
Fatigue is defined as any reduction in the force (or velocity) generating capacity of a muscle,
regardless of whether the muscle is capable of sustaining a given task (NHLBI, 1990).
Weakness refers to a reduced force generating capacity of rested muscle and the term should
not be used when force is reduced by some change in the geometry of the muscle and/or its
attachments. Failure of voluntary activation is used here to denote suboptimal force genera-
tion in a maximal effort with a rested muscle. Central fatigue denotes a decline in motoneuro-
nal activation as force declines during voluntary muscle contractions (see Gandevia et aI.,
Chapter 20).

ENDURANCE CAPACITY OF THE DIAPHRAGM

The diaphragm is the principal muscle of respiration during quiet breathing and
moderate hyperpnoea. Under most circumstances of quiet breathing expiration is largely
passive relying on the elastic recoil of the lungs. The diaphragm contracts intermittently from
mid-gestation throughout life with no extended periods of rest. This pattern of phasic activity
should endow the diaphragm with a high capacity for endurance and teleological argument
would attach evolutionary importance to such properties.
As ventilation increases during exercise or as the loads to inspiration increase
inspiratory synergists (primarily the inspiratory intercostals and scalenes) are progressively
recruited. Eventually the expiratory muscles are also recruited. In terms of pressure genera-
tion during static effort these muscles can be considered as acting partly in series but this
approach has led to conflicting results. At least in the mid-volume range, the diaphragm is
the limiting muscle for pressure development - i.e., it approaches maximal activation while
its synergists may not (Gandevia et aI., 1990; cf. Hershenson et aI., 1988). This finding could
reflect either the greater muscle mass of the synergists or a mechanical advantage. The
transmural pressure recorded across the diaphragm can exceed that across the rib cage during
maximal inspiratory efforts due to recruitment of the abdominal muscles and elevation of
abdominal pressure. This variation in transmural pressure, independent of diaphragmatic
activation, can be explained largely by the length-tension and force-velocity properties of
the diaphragm (Gandevia et aI., 1992). If the diaphragm approaches maximal activation first
its synergists are relatively protected from developing fatigue during inspiratory loading and
any decline in output from the group of muscles is likely to reflect impaired diaphragmatic
performance (Gandevia & McKenzie, 1988; cf. Hershenson et aI., 1989).
Physiological, morphometric and biochemical studies indicate that the diaphragm is
uniquely adapted for aerobic work. As early as 1916 Lee, Guenther and Melaney concluded
from in vitro studies that the diaphragm was more resistant to fatigue than other skeletal
muscles, a finding confirmed in humans by Gandevia, McKenzie and Neering (1983). In
subsequent studies the latter group documented that the diaphragm recovered from fatigue
induced by static inspiratory efforts at 10 times the rate of elbow flexors performing maximal
static efforts (McKenzie & Gandevia, 1991a; their Fig. 3).
The upper limit of blood flow to the diaphragm is two to four times that available to
limb muscles and there is a corresponding increase in capillary density around diaphragmatic
muscle fibers compared with a range oflimb muscles across a number of mammalian species
(for review see McKenzie & Gandevia, 1991b). The volume density of mitochondria, the
oxidative capacity of the muscle fibers and the maximal oxygen consumption of the
diaphragm exceed those of other limb muscles by two to six times.
The diaphragm also appears to be situated advantageously such that its perfusion may
be relatively preserved by the large negative intrathoracic pressures which are developed
 404 D. K. McKenzie and F. Bellemare

100 Figure 2. Relationship between final force

of the elbow flexors after 18 maximal efforts
90 _........ _.. _.. __ ... _._._ .. _........Insp
_.. _...... 2 of 10 s duration and duty cycle (filled cir-
cles). A logarithmic curve was fitted to the
Final force 80 Expul ~ data (r =0.97). For comparison, data for the
(% max) diaphragm during inspiratory efforts (open
70 circles) and expulsive efforts (open squares)
Limb with a duty cycle of 50% have been added.
60 The dashed line indicates the final force of
elbow flexors with a duty cycle of 5%, a
50 performance similar to that of the diaphragm
0 10 20 30 40 50 with a 10-fold increase in duty cycle (from
Duty cycle (%) McKenzie & Gandevia, 199Ia).

when the loads to breathing are high (Buchler et al., 1985), possibly contributing to its
enhanced endurance capacity (Gandevia & McKenzie, 1988). This contrasts with the
situation for most limb muscles and even the myocardium in which intramuscular perfusion
is progressively impaired as tension increases. Diaphragmatic perfusion is reduced during
expulsive contractions which increase abdominal pressure (Buchler et aI., 1985) and, not
surprisingly, endurance is impaired compared with that observed for inspiratory efforts
performed with little change in abdominal pressure (Fig. 2).

CENTRAL AND PERIPHERAL COMPONENTS OF FATIGUE

Much of the earlier literature, both in man and mammalian species, lacked the
methodological precision to enable the central and peripheral components of fatigue to be
quantitated separately. Indeed, many studies did not provide any estimates offorce producing
capacity but relied on indirect estimates such as ventilation or a failure to sustain submaximal
pressures. Studies of experimental animals, healthy subjects and patients are presented
separately and the methodological limitations of each approach highlighted.

Animal Studies
Studies of respiratory muscle fatigue in animals have provided conflicting results,
possibly explained by differences in level of anesthesia, the degree of loading and the
physiological parameters measured. Species differences may also have contributed.
Although the inspiratory muscles are resistant to the development of fatigue, a
number of studies have documented some failure of the contractile mechanism using
supramaximal phrenic nerve stimulation (e.g., Mayock et aI., 1987; De Vito & Roncoroni,
1993). When cardiac output was severely compromised by pericardial tamponade profound
diaphragmatic fatigue was documented (Aubier et aI., 1981).
By contrast Watchko and colleagues (1988) reported that awake infant monkeys
exposed to high inspiratory resistances developed profound hypoventilation with no evi-
dence of peripheral fatigue. A similar result has also been reported for adult dogs exposed
to cardiogenic shock (Nava & Bellemare, 1989). Respiratory frequency declined and apnoea
ensued with no evidence of peripheral diaphragmatic fatigue.
In awake adult sheep exposed to resistive loads hypoventilation appeared to be
associated with evidence of both peripheral and central failure (Bazzy & Haddad, 1984;
Sadoul et aI., 1985) but in these studies contractile failure was not assessed by phrenic nerve
stimulation and there was no direct estimate of central drive. In a subsequent study
diaphragmatic action potentials were shown to decline late in extreme loading suggesting
Respiratory Muscle Fatigue 405

."
Flow
1ft ',"('t of'! YI1 ...... , ··(1· ....
[ 25cmH20

Figure 3. Typical record of air flow, tracheal pressure (Ptr), Pdi' and Edi during resistive loading in an intact
anesthetized dog just before respiratory arrest (RA). Frequency falls just before RA and is associated with
preserved iEMGdi. Pdi obtained by electrical stimulation at 60 and 10 Hz was 42.2 and 55.7%, respectively,
of basal values immediately after RA (adapted from De Vito & Roncoroni, 1993).

that failure of neuromuscular transmission might also contribute to the development of

ventilatory failure (Bazzy & Donnelly, 1993; see also Aldrich, 1987).
After allowing for differences in anesthesia and species a possible unifying explana-
tion for these results is that the response of the respiratory system varies with severity of the
obstruction. With critical airway narrowing respiratory arrest may occur early with little or
no evidence of peripheral fatigue whereas ventilation may be sustained indefinitely with
minimal peripheral fatigue in spite of moderately severe loads. There is presumably a
intermediate range ofloads which can be sustained for a prolonged period with a compromise
of peripheral and central fatigue before respiratory arrest ultimately ensues. Fig. 3 shows the
sudden development of respiratory arrest at a time when the stimulated diaphragm was still
capable of generating 55% of maximal pressure.
The mechanisms underlying the profound central fatigue which can lead to respira-
tory arrest are unclear but several interesting observations have been made. Firstly, the
administration of oxygen prevents secondary hypoxic depression of the respiratory centre
during progressive hypercarbia (Chapman et aI., 1980). Thus, more pronounced peripheral
fatigue will occur with supplemental oxygen. Secondly, administration of naloxone to
reverse the depressant effect of rising endorphin concentrations can restore ventilation
(Scardella et aI., 1986). Bilateral cervical vagotomy can also restore ventilation, suggesting
an inhibitory role for pulmonary afferents (Adams et aI., 1988). Road and colleagues (1987)
have also suggested an inhibitory role for nociceptive phrenic afferents.

Studies in Human Volunteers

A number of early studies of respiratory muscle endurance employed isocapnic
hyperpnoea and measured the maximal sustainable voluntary ventilation as a proportion of
maximal voluntary ventilation (MVV). They mostly reported that very high levels of
ventilation could be sustained. However, it has not been proven that MVV involves maximal
activation of the inspiratory muscles. A decline in maximal inspiratory pressure (MIP) may
occur in athletes after exertion but it is not known whether this is due to central or peripheral
muscle fatigue. Dempsey and co-workers (Johnson et aI., 1993; Dempsey et aI., Chapter 29)
demonstrated a small decline in diaphragmatic twitch and low frequency tetanic pressures
406 D. K. McKenzie and F. Bellemare

Pre-exercise Post-exercise
(90-95% Y02max)
Poes
~ ~

1+-
Figure 4. Individual examples of

~
bilateral phrenic nerve stimulation
REMG in a subject who demonstrated sig-
nificant fatigue post-exercise at

+-
95% of Y 02max. The figure shows
esophageal pressure (Po), gastric

~
pressure (P g) and transdiaphrag-
matic pressure (Pdi) in addition to
LEMG
the compound muscle action poten-
tials from the left and right dia-
phragm (LD and RD). All stimuli
Pdi
Pga ~ ~ ] IOcmH20
were performed at FRC (adapted
from Johnson et aI., 1993).

after maximal intensity exercise but the subjects were still capable of developing their
pre-exercise maximal Pdi during Mueller maneuvers (Fig. 4).
Many studies of respiratory muscle endurance have employed either inspiratory
resistive loading (e.g., Roussos & Macklem, 1977; Roussos et aI., 1979) or threshold loading
(Nickerson & Keens, 1981; see also Clanton et aI., 1988, 1993). These studies involve the
subject sustaining a given percentage of maximal pressure for as long as possible (cf. Clanton
et aI., 1993). Most investigators failed to document whether the maximal pressure was truly
optimal for each subject and did not document the degree of failure of central drive when
the target pressure could no longer be sustained. An important exception was the study of
Bellemare and Big1and-Ritchie (1987) in which twitch interpolation was used to document
the development of significant central fatigue.
Bellemare and Bigland-Ritchie (1984, 1987), Gandevia and McKenzie (1985,
1988), Bellemare and colleagues (1986), McKenzie and Gandevia (1986), and Gandevia
and colleagues (1990) independently developed a method for quantitating the degree of
voluntary activation of the diaphragm using twitch interpolation during maximal static
efforts. They showed that the diaphragm could be activated maximally during inspiratory,
expulsive and combined inspiratory/expUlsive maneuvers. They also used twitch inter-
polation to investigate the development of central fatigue but reached contrasting con-
clusions. Whereas Gandevia and McKenzie (1988) reported little evidence of central
fatigue during 6 min of repeated maximal inspiratory and expulsive efforts (duty cycle
50%), Bellemare and Bigland-Ritchie (1987) reported that, during submaximal expulsive
efforts (30% maximal, duty cycle 67%) the ability to develop Pdi declined by about 50%
and half this loss of force could be attributed to central fatigue. In a collaborative study
(McKenzie et aI., 1992), the latter result was confirmed but the diaphragm appeared to
be resistant to central fatigue during inspiratory efforts; i.e., it appeared to be task
dependent (Fig. 5). That significant central fatigue occurred during inspiratory resistive
breathing (Bellemare & Bigland-Ritchie, 1987) suggests a fundamental difference in
neural control compared with the intermittent brief maximal static efforts used to test
activation in the latter study (McKenzie et aI., 1992). This study also showed that
low-frequency fatigue was relatively small for the diaphragm compared with the elbow
flexors performing a similar task and that there was no failure of neuromuscular trans-
mission (Fig. 6). Other aspects of task dependence are considered by Bigland-Ritchie et
aI., Chapter 27.
Respiratory Muscle Fatigue 407

100

95
~ 90
c
o
'0
<.s 85
>
'0
o
« 80
Limb
75
~ Diaphragm
70 (inspiration)
Test 1 Test 2 Test 3 Test 4

100

95
~
'-' 90
c::
o
.~ 85
>
'0
~ 80 • Limb
75 ~ Diaphragm
70 (expulsion)
Fatigue I Fatigue 2 Fatigue 3

Figure 5. Index of voluntary activation (group data) for the diaphragm and the limb muscle during the four
test sequences (above) and 3 fatigue sequences (below). The test sequences consisted of3-5 brief MVCs with
interpolation of twin stimuli (see also Fig. 6) while the intervening fatigue sequences consisted of 10 MVCs
of 10 s duration (50% duty cycle). Means for all trials in each test and fatigue sequence shown. Diaphragm
tested with maximal inspiratory maneuvers during test MVCs and expulsive maneuvers duringfatigue MVCs.
Stimuli were delivered without warning during fatigue sequences. Note the gradual reduction in voluntary
drive with time and the reduced activation during the fatigue sequence, especially for diaphragm performing
expulsive maneuvers (from McKenzie et aI. , 1992).

RESPIRATORY MUSCLE FATIGUE AND VENTILATORY

FAILURE
In patients with chronic airflow limitation the balance between increased work of
breathing in the face of resistive and elastic loads and the reserve capacity of the respiratory
muscles to generate inspiratory pressures is shifted in an unfavorable direction. This balance
can be expressed as the ratio of the pressure for a given breath (Pbreath) and the maximal
inspiratory pressure (MIP). As shown by Bellemare and Grassino (I 982b, 1983) the likeli-
hood of development of fatigue is predicted by the product of Pbreath/MIP and the duty cycle
for inspiratory work (i.e., the ratio of the duration of inspiration to total breath time: T/T tot ).
They referred to this product as the pressure-time index (PTI) which can be increased by any
combination of increased resistive or elastic work and reduced capacity to generate inspira-
tory pressure (hyperinflation, weakness, fatigue).
Hypercapnia results from a reduction in alveolar ventilation which is determined by
the relative magnitudes of tidal volume and physiological dead space and the respiratory
frequency. As the PTI increases, the ability to maintain tidal volume is impaired with the
inescapable consequence that frequency must rise (Loverdige et aI. , 1986). Although this
may reduce respiratory muscle efficiency and increase the relative proportion of dead space
408 D. K. McKenzie and F. Bellemare

A
EMG

~t
~t
• Test 1

]20CmH20

10ms 50ms

B
150

Relative
amplitude 125
(%)
T? f?

Test 1 Test 2 Test 3 Test 4

Figure 6. A, EMG (left traces) and twitch responses of relaxed diaphragm (P di, right traces) to single stimuli
(above) and twin stimuli (below) obtained during the four test sequences in a typical study (see Fig. 5). EMG
responses ofleft and right costal diaphragm were obtained with surface electrodes. Note the decline in twitch
amplitude following each fatigue sequence. The smallest EMG responses were obtained during the first test
prior to development of fatigue. Vertical calibration for EMG I mY. B, amplitude of compound muscle action
potentials recorded from the hemidiaphragm during relaxation after each test sequence to a single stimulus.
Pooled data from four experiments, normalized to value from first test (mean ± S.E.M.). The amplitude of the
negative phase of the compound muscle action potential (closed symbols) and the peak-to-peak amplitude
(open symbols) increased (from McKenzie et aI., 1992).

ventilation it may reduce the load to breathing and minimize peripheral fatigue (Tobin et aI.,
1986). The perception of breathlessness and inspiratory effort are also predicted by the ratio
Pbreath/MIP (Killian & Jones, 1988). Thus the reduction of tidal volume may be as much a
mechanism to reduce respiratory distress as to minimize inspiratory muscle fatigue.
There is no definitive evidence that inspiratory muscles are weakened as a result of
chronic airflow limitation although their ability to develop inspiratory pressure is severely
compromised by the length-tension and geometric consequences of hyperinflation. How-
ever, the metabolic consequences of advanced disease and acute ventilatory failure may
include hypoxaemia, hypercapnia, acidosis and other electrolyte disturbances. Hypercapnic
acidosis reduces twitch tension of adductor pollicis and reduces its endurance capacity
(Vianna et aI., 1990). The diaphragm may also be susceptible (Juan et aI., 1984). Hydrogen
ion accumulation may impair calcium release from the sarcoplasmic reticulum and also exert
a negative influence on the contractile mechanism (e.g., Westerblad et aI., 1991).
Respiratory Muscle Fatigue 409

The diaphragm is resistant to utilizing anaerobic metabolism even under conditions

of severe hypoxaemia and extreme loading, presumably because of its enhanced oxidative
capacity and perfusion (see above). There is little data on this for the other respiratory
muscles and lactate accumulation has been found in human subjects breathing through high
inspiratory loads (Jardim et aI., 1981; Freedman et aI., 1983). Endurance times for resistive
breathing are significantly decreased during hypoxaemia (Roussos & Macklem, 1977;
Jardim et aI., 1981) but it is not clear whether this finding reflects increased peripheral or
central fatigue (see also Koulouris et aI., 1984).
During prolonged submaximal work the development of fatigue is accompanied
by glycogen depletion in the exercising muscle fibers. This has been documented in
some animal studies of loaded breathing (Ferguson et aI., 1990) but not all (Mayock
et aI., 1991). The discrepancy could reflect species and age differences or differences
in loading protocols. One report has appeared suggesting a reduction in energy sources
in biopsies of respiratory muscles from patients with respiratory failure (Gertz et aI.,
1977).
There is little evidence that respiratory failure in patients with chronic stable
hypercapnia is associated with any degree of respiratory muscle fatigue. Begin and
Grassino (1991) studied 300 patients with chronic airflow limitation. Although the PTI
for these patients was much higher than normal it did not lie in the fatiguing range
(Bellemare & Grassino, 1982b), unless they were asked to breathe more slowly and
deeply than they chose.
The study of Begin and Grassino (1991) also demonstrated that the degree of
hypercarbia was not tightly correlated with any single measure of airway narrowing or
respiratory muscle compromise. Even when five variables were considered only half the
variance of arterial carbon dioxide tension could be explained. This is not surprising in view
of the many factors that can affect alveolar ventilation and CO 2 production and the influence
ofwake/sleep states. A major factor in determining the set point for arterial carbon dioxide
is nocturnal hypoventilation due to repetitive central or obstructive apneas. The severity of
respiratory disturbance during sleep is not well predicted by any daytime measure of
function. Supporting this hypothesis, Begin and Grassino reported that overweight patients
were disproportionately represented among those with increasing levels of hypercarbia
during wakefulness.
In summary, the major mechanism for chronic hypercarbia in these patients is a
failure to increase central drive sufficiently to maintain alveolar ventilation. Central drive is
not reduced compared with normal, except perhaps in the late stages ofhypercarbia (Aubier
et aI., 1980; Gribben et aI., 1983). This point has been recently confirmed directly by single
motor unit recordings from the parasternal intercostals (De Troyer, Gandevia, Leeper and
McKenzie, unpublished observations).
Although it seems that peripheral muscular fatigue is not prominent in stable
hypercapnic patients the situation could be different in patients with acute exacerbations
of critical airway narrowing. Useful data in this setting are scarce although there have
been several studies of patients weaning from ventilators. The ratio Pbreath/MIP is a useful
guide and when the ratio is high weaning is likely to fail with development of rapid
shallow breathing (Cohen et aI., 1982; Tobin et aI., 1986; Goldstone et aI., 1994). Recently
it has been shown that the maximal relaxation rate of the inspiratory pressure-time profile
slows in patients who fail a trial of weaning while it remains unchanged in successful
trials (Goldstone et aI., 1994). While this finding points to a peripheral component of
inspiratory muscle fatigue it does not rule out the possibility of a concomitant decline in
central drive.
410 D. K. McKenzie and F. Bellemare

THE ROLE OF VOLUNTARY PATHWAYS IN LOADED

BREATHING

For most of this century investigations of respiratory control mechanisms have

centered on the groups of respiratory-related neurons located in the bulbo-pontine region of
the brainstem. It has been widely accepted that this respiratory centre was the primary
regulator of respiration in both sleep and wakefulness with only occasional intrusions from
supra-pontine regions for specific tasks. For example it has long been assumed that a
voluntary deep inhalation could be initiated by motor cortical neurons but direct confirmation
has only recently been provided. Macefield and Gandevia (1991) demonstrated in recordings
from scalp electrodes definite pre-motor D.C. potential shifts (Bereitschaftspotentials) with
appropriate latencies prior to inhalation. Motor potentials were recorded 25 ms before the
onset ofEMG activity. PET scanning techniques have been used during volitional breathing
to demonstrate increases in cerebral blood flow in the primary motor cortical area, the basal
ganglia and the cerebellum (Colebatch et aI., 1991).
It has also been accepted that respiratory center output can be inhibited during speech
and breath-holding. In 1963 Agostoni documented cessation of diaphragmatic EMG for
about 20 s during breath-holding and a potential neural substrate for this inhibition has been
identified in cats trained to breath-hold (Orem & Netick, 1986).
Emotions have long been known to influence respiration but this is thought to involve
subcortical pathways. For example, patients with lesions in the corticospinal tract which
ablate volitional control of breathing still show alterations in respiration with emotions
(Munschauer et aI., 1991).
In recent years several groups have entertained the possibility that suprapontine
pathways may predominate in the control of breathing during wakefulness while the
brainstem respiratory neurons play the major role during sleep (and anesthesia). Evidence
for this is circumstantial but mounting, especially for exercise hyperpnoea (see Eldridge &
Waldrop, 1991). Until recently it has been assumed that the hyperpnoea of exercise was
driven by afferent neural information and metabolites from exercising muscles but there have
always been paradoxical results which were difficult to explain. Subjects with defective
automatic control of respiration who hypoventilate or stop breathing during sleep have
normal ventilatory responses to exercise (e.g., children with congenital hypoventilation:
Shea et aI., 1993; adults with medullary lesions: Heywood et aI.. 1993). There is even
evidence that ventilation can be stimulated by anticipation or imagination of exercise (Tobin
et aI., 1986; Decety et aI., 1991).
Voluntary control of breathing could be achieved either via direct corticospinal
pathways or indirectly via putative cortico-bulbar projections (Bassal et aI., 1981). The
presence of powerful, rapidly-conducting corticospinal projections to a variety of neck and
trunk muscles including the diaphragm has been documented by Gandevia and colleagues
using high-voltage transcutaneous stimulation of the motor cortex and spinal outflow tracts
(Gandevia & Rothwell, 1987; Gandevia & Plassman, 1988).
The partitioning between voluntary and automatic regulation of breathing may be of
crucial importance in determining ventilatory responses to loaded breathing, especially in
patients with critical airway narrowing (e.g., acute asthma or chronic airflow limitation). It
is a common clinical observation that patients with critical airway narrowing cannot sleep.
When such patients do sleep it means either that they have drifted into hypercapnic coma or
their airflow obstruction has been relieved by the emergency therapy. As yet, however, there
is no direct evidence to support the contention that volitional breathing is the predominant
driving force in such patients. On the other hand there is ample evidence that such patients
are at risk of developing hypoventilation during sleep (see Saunders & Sullivan, 1994).
Respiratory Muscle Fatigue 411

Moreover, there is evidence from animal studies that the respiratory centre is not capable of
maximal activation of the phrenic motoneuron pool, even under conditions of extreme
hypercarbia and/or hypoxemia (Sieck & Fournier, 1989).
In conclusion, there has been a rapid growth in knowledge about the function of
respiratory muscles in response to loading over the past fifteen years. Much of this progress
has been stimulated by the development and application of sound neurophysiological
methodology in the investigation of a complex set of muscles which act in series and in
parallel. In contrast to the situation in the organ bath and for many limb muscles, it is difficult
to measure the force, length and velocity of most respiratory muscles non-invasively
compounding the difficulties in understanding the mechanics of contraction (cf. Gandevia
et aI., 1992; McKenzie et aI., 1994). The challenge of the future will be to document the
precise roles and interactions of the volitional and involuntary control systems in everyday
activities and in response to increased loads.

ACKNOWLEDGMENTS

The authors wish to thank Dr. Simon Gandevia and Ms. Gabrielle Allen for comments
on the manuscript, and G.A. and Ms. Jane Butler for its preparation. Our laboratory has been
supported by the National Health & Medical Research Council of Australia and the Asthma
Foundation of New South Wales, Australia. Attendance of the authors at the 1994 Bigland-
Ritchie conference was supported, in part, by the Muscular Dystrophy Association (USA;
[Link].) and the University of Miami (FB.).

REFERENCES
Adams JM, Farkas GA, Rochester DF (1988). Vagal afferents, diaphragm fatigue and inspiratory resistance
in the anesthetized dog. Journal of Applied Physiology 64, 2279-2286.
Agostoni (1963). Diaphragm activity during breath-holding: factors related to its onset. Journal of Applied
Physiology 18, 30-36.
Aldrich TK (1987). Transmission failure of the rabbit diaphragm. Respiration Physiology 69, 307-319.
Aubier M, Murciano D, Fournier M, Milic-Emili J, Pariente R & Derenne JPh (1980). Central respiratory drive
in patients with chronic obstructive lung disease in acute respiratory failure. American Review of
Respiratory Disease 122, 191-200.
Aubier M, Trippenbach T & Roussos Ch (1981). Respiratory muscle fatigue during cardiogenic shock. Journal
ofApplied Physiology 51, 499-508.
Bassal M, Bianchi AL, Dussardier M (1981). Effets de la stimulation des structures nerveuses centriiles sur
I'activite des neurones rt!spiratoires chez Ie chat. Journal of Physiology (Paris) 77, 779-795.
Bazzy AR & Donnelly DF (1993). Diaphragmatic failure during loaded breathing: role of neuromuscular
transmission. Journal of Applied Physiology 74, 1679-1683.
Bazzy AR & Haddad GG (1984). Diaphragmatic fatigue in unanesthetized adult sheep. Journal of Applied
Physiology 57, 182-190.
Begin P & Grassino A (1991). Inspiratory muscle dysfunction and chronic hypercapnia in chronic obstructive
pulmonary disease. American Review of Respiratory Diseases 143, 905-912.
Bellemare F & Bigland-Ritchie B (1984). Assessment of human diaphragm strength and activation using
phrenic nerve stimulation. Respiration PhYSiology 58,263-277.
Bellemare F & Bigland-Ritchie B (1987). Central components of diaphragmatic fatigue assessed by phrenic
nerve stimulation. Journal of Applied Physiology 62, 1307-1316.
Bellemare F, Bigland-Ritchie B & Woods JJ (1986). Contractile properties of the human diaphragm in vivo.
Journal ofApplied Physiology 61, 1153-1161.
Bellemare F & Grassino A( 1982a). Evaluation of the human diaphragm fatigue. Journal ofApplied Physiology
53, 1196-1206.
Bellemare F & Grassino A (1982b). Effect of pressure and timing of contraction on human diaphragm fatigue.
Journal ofApplied Physiology 53, 1190-1195.
412 D. K. McKenzie and F. Bellemare

Bellemare F & Grassino A (1983). Force reserve of the diaphragm in patients with chronic obstructive
pulmonary disease. Journal ofApplied Physiology 55,8-15.
Buchler B, Magder, S, Katsardis H, Jammes Y & Roussos C (1985). Effects of pleural pressure and abdominal
pressure on diaphragmatic blood flow. Journal of Applied Physiology 58,691-697.
Chapman RW, Santiago TV & Edelman NH (1980). Brain hypoxia and the control of breathing: neuromechani-
cal control. Journal of Applied Physiology 49, 497-505.
Clanton TL & Ameredes BT (1988). Fatigue of the inspiratory muscle pump in humans: an isoflow approach.
Journal ofApplied Physiology 64, 1693-1699.
Clanton TL, Hartman E & Julian MW (1993). Preservation of sustainable inspiratory muscle pressure at
increased end-expiratory lung volume. American Review of Respiratory Disease 147, 385-39\.
Cohen C, Zagelbaum G, Gross D, Roussos Ch & Macklem PT (1982). Clinical manifestations of inspiratory
muscle fatigue. American Journal of Medicine 73, 308-316.
Colebatch JG, Adams L, Murphy K, Martin AJ, Lammertsma AA, Tochon-Danguy HJ, Clark JC, Friston KJ
& Guz A (1991). Regional cerebral blood flow during volitional breathing in man. Journal of
Physiology (London) 443, 91-103.
Davies HW, Haldane JS & Priestly JG (1919). The response to respiratory resistance. Journal of Physiology
(London) 53, 60-69.
Decety J, Jeannerod M, Germaine M & Pastene J (1991). Vegetative response during imagined movement is
proportional to mental effort. Behavioural Brain Research 42, 1-5.
De Vito EL & Roncoroni AJ (1993). Central and peripheral diaphragmatic fatigue in loaded normal and
vagotomized dogs. Journal ofApplied PhYSiology 74, 2820-2827.
Eldridge FL &Waldrop TG (1991). Neural control of breathing during exercise. In: Whipp BJ, Wasserman K
(eds.), Exercise Pulmonary Physiology and Pathology, pp. 309-370. New York: Marcel Dekker.
Ferguson GT, Irvin CG & Cherniack RM (1990). Effect of corticosteroids on diaphragm function and
biochemistry in the rabbit. American Review of Respiratory Disease 141, 156-163.
Freedman S, Cooke NT & Moxham J (1983). Production of lactic acid by respiratory muscles. Thorax 38,
50-54.
Gandevia SC, Gorman RB, McKenzie DK & Southon FCG (1992). Dynamic changes in human diaphragm
length: maximal inspiratory and expulsive efforts studied with sequential radiography. Journal of
Physiology (London) 457,167-176.
Gandevia SC & McKenzie DK (1985). Activation of the human diaphragm during maximal static efforts.
Journal of Physiology (London) 367, 45-56.
Gandevia SC & McKenzie, DK (1988). Human diaphragmatic endurance during different maximal respiratory
efforts. Journal of Physiology (London) 395, 625-638.
Gandevia SC, McKenzie DK & Neering IR (1983). Endurance properties of respiratory and limb muscles.
Respiration Physiology 53, 47-6\.
Gandevia SC, McKenzie DK & Plassman BL (1990). Activation of human respiratory muscles during different
voluntary manoeuvres. Journal of Physiology (London) 428, 387-403.
Gandevia SC & Plassman BL (1988). Responses in human intercostal and truncal muscles to motor cortical
and spinal stimulation. Respiration Physiology 73, 325-338.
Gandevia SC & Rothwell JC (1987). Activation of the human diaphragm from the motor cortex. Journal of
Physiology (London) 384, 109-118.
Gertz I, Hedenstierna G, Hellers G & Wahren J (1977). Muscle metabolism in patients with chronic obstructive
lung disease and acute respiratory failure. Clinical Science and Molecular Medicine 52, 395-403.
Goldstone JC, Green M & Moxham J (1994). Maximum relaxation rate of the diaphragm during weaning from
mechanical ventilation. Thorax 49, 54-60.
Gribben HR, Gardiner IT, Heinz CJ, Gibson TJ & Pride NB (1983). The role of impaired inspiratory muscle
function in limiting ventilatory response to CO 2 in chronic airflow obstruction. Clinical Science 64,
487-495.
Gross D, Grassino A, Ross WRD & Macklem PT (1979). Electromyogram pattern of diaphragmatic fatigue.
Journal ofApplied Physiology 46, 1-7.
Hershenson MB, Kikuchi Y & Loring SH (1988). Relative strengths of the chest wall muscles. Journal of
Applied Physiology 65, 852-862.
Hershenson MB, Kikuchi Y, Tzelepis GE & McCool FD (1989). Preferential fatigue of the rib cage muscles
during inspiratory resistive loaded ventilation. Journal ofApplied Physiology 66, 750-754.
Heywood P, Moosavi S, Morrell MJ & Guz A(1993). CO 2 sensitivity and breathing at rest and during exercise
in human subjects with unilateral medullary lesions. Journal of Physiology (London) 459, 353P.
Jardim J, Farkas G, Prefaut C, Thomas D, Macklem PT & Roussos Ch (1981). The failing inspiratory muscles
under normoxic and hypoxic conditions. American Review of Respiratory Disease 124, 274-279.
Respiratory Muscle Fatigue 413

Johnson BD, Babcock MA, Suman OE & Dempsey JA (1993). Exercise-induced diaphragmatic fatigue in
healthy humans. Journal of Physiology (London) 460, 385-405.
Juan G, Calverly P, Talamo C, Schnader J & Roussos C (1984). Effect of carbon dioxide on diaphragmatic
function in the human being. New England Journal ofMedicine 310, 874-879.
Killian KJ & Jones NL (1988). Respiratory muscles and dyspnea. Clinical Chest Medicine 9,237-248.
Koulouris N, Moxham J, Barnes N, Gray B, Heaton R & Green M (1984). Effect of hypoxia on respiratory
and limb muscle endurance. Thorax 39,714.
Lee FS, Guenther AE & Melaney HE (1916). Some of the general physiological properties of diaphragm
muscle as compared with certain other mammalian muscles. American Journal of Physiology 40,
446-473.
Loverdige B, West P, Kryger MH & Anthonisen NR (1986). Alteration in breathing pattern with progression
of chronic obstructive pulmonary disease. American Review of Respiratory Diseases 134, 930-934.
Macefield G & Gandevia SC (1991). The cortical drive to human respiratory muscles in the awake state
assessed by premotor cerebral potentials. Journal of Physiology (London) 439, 545-558.
Mayock DE, Badura RJ, Watchko IF, Standaert TA & Woodrum DE (1987). Response to resistive loading in
the newborn piglet. Pediatric Research 21,121-125.
Mayock DE, Standdaert TA, Murphy TD & Woodrum DA (1991). Diaphragmatic force and substrate response
to resistive loaded breathing in the piglet. Journal ofApplied Physiology 70, 70-76.
McKenzie DK, Bigland-Ritchie B, Gorman RB & Gandevia SC (1992). Central and peripheral fatigue of
human diaphragm and limb muscles assessed by twitch interpolation. Journal ofPhYSiology (London)
454, 643-656.
McKenzie DK & Gandevia SC (1986). Strength and endurance of inspiratory, expiratory, and limb muscles
in asthma. American Review of Respiratory Diseases 134, 999-1004.
McKenzie DK & Gandevia SC (1987). Influence of muscle length on human inspiratory and limb muscle
endurance. Respiration Physiology 67, 171-182.
McKenzie DK & Gandevia SC (1991a). Recovery from fatigue of human diaphragm and limb muscles.
Respiration Physiology 84, 49-60.
McKenzie DK & Gandevia SC (199Ib). Skeletal muscle properties: Diaphragm and chest wall. In: Crystal
RG, West JB (eds.), The Lung. pp. 649-659. New York: Raven Press Ltd.
McKenzie DK, Gandevia SC, Gorman RB & Southon FCG (1994). Dynamic changes in the zone of apposition
and diaphragm length during maximal respiratory efforts. Thorax, 49, 634-638.
Munschauer FE, Mador MJ, Ahuja A and Jacobs L (1991). Selective paralysis of voluntary but not limbically
influenced automatic respiration. Archives of Neurology 48, 1190-1192.
Nava S & Bellemare F (1989). Cardiovascular failure and apnea in shock. Journal ofApplied Physiology 66,
184-189.
Newell SZ, McKenzie DK & Gandevia SC (1989). Inspiratory and skeletal muscle strength, endurance and
diaphragmatic activation in patients with chronic airflow limitation. Thorax 44,903-912.
NHLBI Workshop on Respiratory Muscle Fatigue: Report ofthe Respiratory Muscle Fatigue Workshop Group
(1990). American Review of Respiratory Disease 142, 474-480.
Nickerson BG & Keens TG (1981). Measuring ventilatory muscle endurance in humans as sustainable
inspiratory pressure. Journal ofApplied Physiology 52, 768-772.
Orem J & Netick A (1986). Behavioural control of breathing in the cat. Brain Research 366, 238-253.
Road J, West NH & Van Vliet BN (1987). Ventilatory effects of stimulation of phrenic afferents. Journal of
Applied Physiology 63, 1063-1069.
Rochester DF (1991a). Effects of COPD on the respiratory muscles. In: Cherniack NS (ed.), Chronic
Obstructive Pulmonary Disease, pp. 134-157, Philadelphia: WB Saunders Co.
Rochester DF (1991b). Respiratory muscle weakness, pattern of breathing, and C02 retention in chronic
obstructive pulmonary disease. American Review ofRespiratory Disease 143, 901-903.
Roussos Ch, Fixley M, Gross D & Macklem PT (1979). Fatigue of inspiratory muscles and their synergistic
behaviour. Journal ofApplied Physiology 46, 897-904.
Roussos Ch & Macklem PT (1977). Diaphragmatic fatigue in man. Journal ofApplied Physiology 43, 189-197.
Sadoul N, Bazzy AR, Akabas SR & Haddad GG. (1985) Ventilatory response to fatiguing and nonfatiguing
resistive loads in awake sheep. Journal ofApplied Physiology 59, 969-978.
Saunders NA & Sullivan CE (eds.) (1994). Sleep and Breathing. 2nd Ed. New York: Marcel Dekker.
Scardella AT, Parisi RA, Phair DK, Santiago TV & Edelman NH (1986). The role of endogenous opioids in
the ventilatory response to acute flow-resistive loads. American Review ofRespiratory Disease 133,
26-31.
Shea SA, Andres LP, Shannon DC & Banzett RB (1993). Ventilatory responses to exercise in humans lacking
ventilatory chemosensitivity. Journal ofPhysiology (London) 468, 623-640.
414 D. K. McKenzie and F. Bellemare

Sieck GC & Fournier M (1989). Diaphragm motor unit recruitment during ventilatory and nonventilatory
behaviours. Journal ofApplied Physiology 66, 2539-2545.
Tobin Ml, Press V & Guenther SM (1986). The pattern of breathing during successful and unsuccessful trials
of weaning from mechanical ventilation. American Review of Respiratory Disease 134, 111l-1118.
Vianna LG, Koulouris N, Lanigan C & Moxham 1 (1990). Effect of acute hypercapnia on limb muscle
contractility in humans. Journal ofApplied Physiology 69,1486-1493.
Watchko IF, Mayock DE, Standaert TA & Woodrum DE (1988). Ventilatory failure during loaded breathing:
the role of central neural drive. Journal ofApplied Physiology 65, 249-255.
Westerblad H, Lee lA, Lannergren 1 & Allen DG (1991). Cellular mechanisms of fatigue in skeletal muscle.
American Journal of Physiology (Cell Physiology) 261, C 195-C209.
 31

FATIGUE OF JAW MUSCLES AND SPEECH

MECHANISMS

T. S. Miles and M. A. Nordstrom

Department of Physiology
University of Adelaide
Adelaide SA 5005, Australia

ABSTRACT

Histochemical studies show that the distribution of fiber types in human jaw muscles is
different from that in various limb muscles, no doubt representing different functional
demands as well as a different embryological derivation. Jaw-closing muscles appear more
resistant to fatigue than limb muscles with intermittent maximal contractions. Endurance of
continuous isometric biting is limited by pain. Masseter motor unit fatigability in sub-maxi-
mal contractions is similar to the limb muscles. There are few physiological data for the
jaw-opening muscles. The distribution of fiber types in human speech muscles is consistent
with the high speeds of contraction that must be used in phonation. Although clinical
syndromes of fatigue of speech muscles are recognized, there is little direct information on
the fatigability of the muscle fibers themselves.

INTRODUCTION

The current world record for weightlifting in the snatch event, in which a barbell is
lifted above the shoulders in one movement, is 216 kg (about 2120 N). However, it is not
necessary to be an elite athlete to exert forces of this magnitude. In 1681, Borelli attached
weights to a cord which passed over the mandibular teeth and reported that his subjects were
able to lift about 200 kg (Rowlett, 1932-33). More recently, Gibbs and colleagues (1986)
reported that the highest isometric, bilateral bite force recorded with an improved transducer
was 4400 N, although the average value was 725 N. Another recent study has reported
maximal values for unilateral bites of 847 Nand 597 N for young adult males and females
respectively (Waltimo & K6n6nen, 1993).
These observations serve to illustrate the power of the jaw-closing muscles. The
muscles themselves are well adapted to produce the high forces that are required for breaking
down tough foods (although these are lamentably rare in modem diets). The six muscles that
act to elevate the jaw have a higher mechanical advantage than most limb muscles because
they are short and act directly across the joint without the interposition of long, flexible
tendons between the muscle and the bones. These muscles are phenotypically different from

415
416 T. S. Miles and M. A. Nordstrom

those in the limbs. These differences may reflect not only their unique functions, but also
their different embryological origin. In comparison to the limb and trunk muscles which are
derived from myotones of the somites, the masticatory muscles arise from somitomeres in
the first branchial arch (Noden, 1983). The masseter in particular expresses myosin isoforms
that are observed only during ontogeny in trunk and limb muscles (Soussi-Yanicostas et aI.,
1990), including the cardiac-specific heavy chain that is found only in muscles (including
the heart) that arise from the cranial part of the embryo (Bredman et aI., 1990).
In marked contrast to the powerful jaw-closing muscles, the jaw-opening muscles
are long and thin, and do not exert much force. They usually operate against minimal external
load when opening the jaw, and are rarely tonically active.
The concept of fatigue of the masticatory muscles may seem strange to those of us
accustomed to a modem diet in which the food is chewed a few times, then swallowed.
Indeed, it is likely that jaw muscle fatigue is rare in these circumstances. However, the
histochemical appearance of the jaw-closing muscles suggests that they evolved to deal with
diets that required considerable force. Waugh (1937) reported that Alaskan Eskimos could
exert bite forces as high as 158 kg (about 1550 N). Their masticatory prowess may reflect
their diet at that time, in which the practice of chewing whale skins to remove the blubber
gave their jaw-closing muscles intensive training. Masticatory fatigue in those eating a
modem diet is probably limited mainly to individuals who grind their teeth for protracted
periods, usually while sleeping. (The syndrome of night grinding, known as bruxism, is
important clinically because it is associated with severe pain in the head and neck).
This review will also touch briefly on fatigue in the muscles involved in speech. This
will be limited to the little that is known about the laryngeal muscles as the respiratory
muscles are dealt with elsewhere (Dempsey & Babcock, Chapter 29; McKenzie & Bellamare,
Chapter 30.

FATIGUE OF THE JAW MUSCLES

Histochemical Profile of Human Masticatory Muscles

The muscle fiber composition of the jaw muscles provides useful information on the
likely fatigue resistance of the jaw muscles (Table 1). The fiber types of the human jaw
muscles have been classified histochemically using biopsy (Ringquist, 1971; Ringquist,
1973; Serratrice et aI., 1976; Ringquist et aI., 1982) and autopsy techniques (Serratrice et
aI., 1976; Vignon et aI., 1980; Eriksson et aI., 1981; Eriksson et aI., 1982; Eriksson &
Thornell,1983).
The main fiber type in the jaw closers and lateral pterygoid is type I, with type lIB
the next most numerous. Very few type IIA fibers are found in the jaw muscles. The
jaw-opening digastric muscle has a more even distribution of the three main fiber types, and
its fiber type composition and appearance is more like the limb muscles. The histological
appearance of the jaw-closing muscles (masseter, medial pterygoid, temporalis) and lateral
pterygoid differs from that normally seen in the limb muscles. These differences are relevant
for fatigue resistance of the muscles, and are summarized below:
1. The diameter of type I fibers is larger than that of type II fibers in nearly all sites
in the muscle (Ringquist, 1973; Serratrice et aI., 1976; Vignon et aI., 1980;
Eriksson & Thornell, 1983). This is in contrast to the normal situation in the limb
muscles where type II fibers are either larger, or of similar diameter to type I fibers
(Dubowitz & Brooke, 1973). The total cross-sectional area of the jaw-closing
Fatigue of Jaw Muscles and Speech Mechanisms 417

Table 1. Relative proportion of histochemical fiber types in

masticatory muscles
Histochemical fiber type proportion
Muscle 1 1M lIA lIB lIe
Masseter* 63 6 2 27 3
Medial Pterygoid* 54 8 0 36 2
Temporalis* 53 5 0 39 2
Lateral pterygoidt 70 14 0 11 5
Digastric§ 34 0 27 38 0
Values represent percent of fibers of each histochemical type in
the jaw muscles.
*Eriksson and Thornell (1983).
tEriksson and colleagues (1981).
§Eriksson and colleagues (1982).

muscles and lateral pterygoid occupied by type I fibers (up to 90%; Eriksson &
Thornell, 1983) is therefore larger than suggested by the fiber type proportions.
2. Diameters of both the type I and II fibers are smaller than in other skeletal muscles
(Ringquist, 1971; Eriksson & Thornell, 1983; Vignon et aI., 1980). Small fiber
diameters may enhance fatigue resistance by facilitating diffusion of metabolites
and energy substrates in and out of muscle cells (Edstrom & Grimby, 1986).
3. The jaw-closing muscles contain a large proportion of fibers with intermediate
staining for myosin-ATPase (MATPase) (Vignon et aI., 1980; Ringquist et aI.,
1982; Eriksson & Thornell, 1983). Fibers of this type (1M and IIC) are rare or
non-existent in normal adult limb muscles (Dubowitz & Brooke, 1973), but have
been seen with extreme endurance training of muscles (Edstrom & Grimby, 1986).
4. It appears that the jaw-closing muscles and lateral pterygoid in general do not
contain type IIA fibers (Eriksson et aI., 1981; Ringquist et ai. 1982, Eriksson &
Thornell, 1983), although these may be found in significant proportions in some
individuals. Type IIA fibers correspond to motor units of the FR type.
5. The masticatory muscles do not have the mosaic pattern of fiber types normally
found in limb muscles, but rather have large groups of densely packed fibers of
the same histochemical type (Eriksson & Thornell, 1983). Such an appearance in
a limb muscle would be considered pathological (Dubowitz & Brooke, 1973).
In summary, the histological appearance of the masticatory muscles is sufficiently
different from the normal appearance oflimb muscles that care should be taken in inferring
the physiological properties of the muscles from this criterion alone. The high proportion of
type I fibers, and the small diameters of the fibers suggest that masticatory muscles should
be reasonably resistant to fatigue.

Physiological Data from Animal Experiments

In contrast to the limb muscles, the physiological properties of jaw muscles and their
motor units have not been well characterized in animal experiments. For example, there have
been no studies in which electrical stimulation has been used to determine the mechanical
properties of jaw-muscle motor units in animals. This is presumably because of the difficulty
in gaining access to the motor nerves in these muscles without damaging the muscles
themselves. Motor roots are not accessible for fiber-splitting as they are for limb muscles. It is
feasible to activate motoneurons in the trigeminal motor nucleus by intracellular current
418 T. S. Miles and M. A. Nordstrom

injection through a microelectrode, but this technique has not been used to investigate the
mechanical properties of motor units in the masticatory system. Consequently, the physiologi-
cal properties of motor units in the jaw muscles of animals have mostly been inferred from
whole-muscle studies. Even here, the literature is sparse. The consensus of the few animal
studies in the literature is that the jaw muscles are fast-twitch muscles. Direct stimulation of the
muscle has been used in most animal studies because of the difficulty of access to the motor
nerves. Tamari and colleagues (1973) reported that the twitch time-to-peak tension (TTP) of
the cat masseter was faster than temporalis (18 vs. 34 ms). Very fast twitch TTPs have also been
reported in the masseter of the rat (14 ms; Nordstrom & Yemm, 1974) and jaw-closers in the
possum (16-18 ms; Thexton & Hiiemae 1975). The most extensive study is that of Taylor and
colleagues. (1973), who combined an investigation of the mechanical properties of masseter
and temporalis muscle strips and a histochemical analysis. The muscle strips were very
fast-contracting, with a twitch TTP of 11-13 ms. A fatigue test was applied to the muscle strips
(train of pulses at 55-90 Hz, of 330 ms duration, repeated once every second). With this
stimulation pattern there was a rapid loss of tension in the first minute to about 25% of the initial
level, and then a gradual decline over the next 2-3 min. Although this test paradigm was chosen
to be compatible with a widely used fatigue test (see Burke, 1981) it is possible that with these
high rates of stimulation fatigue resulted primarily from failure of activation of the muscle
fibers. It is not possible to assess this as EMG was not recorded. The fatigue profile was deemed
to be consistent with the relative proportions of fiber types determined histochemically. The
predominant fiber type was large in size and stained strongly for MATPase, and weakly for the
oxidative enzyme succinate dehydrogenase (SDH). The second fiber type was of intermediate
size, stained strongly for MATPase and SDH. These fiber types were assumed to correspond to
motor unit types FF and FR, respectively, and in the cat masseter they comprised 90% of the
muscle cross-sectional area. The remaining fibers had low MATPase activity, and stained
strongly for SDH, corresponding to type S motor units.
Even if extensive data on jaw muscle fatigue were available from animals, its
relevance to fatigue ofhumanjaw muscles is questionable. There are significant differences
in the fiber type composition of the human and animal jaw muscles. Human jaw-closing
muscles are composed predominantly of type I fibers, whereas in the cat and rat the
jaw-closers are almost exclusively type II fibers. Humanjaw muscles appear to lack the type
IIA fibers (corresponding to the physiological type FR units). The masseter in the rhesus
monkey does appear to have a similar histochemical profile to humans (Maxwell et aI., 1979)
and might therefore be a valuable animal model for physiological data.

Physiological Data from Whole-Muscle Studies in Humans

Fatigue can be defined as a reduction of force-producing capacity of the neuromus-
cular system with prolonged activity (Asmussen, 1979). It is often useful to distinguish
between central and peripheral fatigue, particularly when attempting to quantify fatigue.
Central fatigue is characterized by a reduction of electrical activation of the muscle,
accompanied by (but not necessarily causing) force loss. Peripheral (or contractile) fatigue
is characterized by a reduction in force for a given level of muscle excitation. When
measuring fatigue it is desirable to assess, and if possible control the level of excitation of
the muscle in order to differentiate central from peripheral causes of force loss. For most
practical purposes, it is desirable to test fatigue under conditions of constant excitation (i.e.,
in the absence of central fatigue), so that fatigue may be objectively measured in terms of
loss of force. This ideal experimental design is possible in animal experiments, but is very
difficult to achieve in human experiments. Unfortunately, in many studies of human jaw
muscle fatigue, objective quantification offatigue (deficit in force producing capacity) has
been hampered by loose operational definitions ofjatigue. The approaches to fatigue testing
Fatigue of Jaw Muscles and Speech Mechanisms 419

in the human jaw muscles can be grouped into three broad categories, which are considered
below. Of these, only the use of the maximal voluntary contractile force (MVC) approaches
the ideal of an objective measure of fatigue.
1. The endurance time, or the time elapsed until the subject is unwilling to continue
a required task, has been used as an index offatigue in both maximal (Christensen,
1979; Palla & Ash, 1981; Christensen & Mohamed, 1983; Christensen et aI., 1985)
and submaximal biting (Naeije, 1984; Clark et aI., 1984; Clark & Carter, 1985;
Clark & Adler, 1987; Maton et aI., 1992). Endurance has been confused with
resistance to fatigue in many of these studies but while they may be related, they
are not necessarily the same. Endurance is the ability to withstand prolonged
strain, and is limited by pain and motivational factors as well as contractile fatigue.
Endurance times are shorter in the jaws than limb muscles for isometric contrac-
tions at comparable relative forces (Maton et aI., 1992), but this may indicate that
isometric contractions ofjaw muscles are more painful rather than more fatigable
than the limbs. Clark and colleagues have examined this issue in the jaw muscles
in well-designed experiments. During sustained biting at force levels varying from
25-100% of maximal they demonstrated that the subject's endurance was limited
by pain (Clark & Carter, 1985). The ability to perform intermittent MVCs at
intervals during sustained submaximal biting was unaffected, indicating that
fatigue or inability to produce the target force was not the limiting factor. Pain in
a contracting muscle is believed to be related to an increased concentration of
metabolic by-products of contraction such as H+ and K+ as a consequence of
contraction-induced occlusion of blood flow through the muscle (Mense, 1977).
2. The shift in the power spectrum of the surface EMG signal towards greater
low-frequency energy content with muscle fatigue (Lindstrom et aI., 1970; Mills,
1982) has been used in numerous attempts to quantify jaw muscle fatigue in
humans (Naeije & Zorn, 1981; Palla & Ash, 1981; Lindstrom & Hellsing, 1983;
van Boxtel et aI., 1983; Naeije, 1984; Kroon et aI., 1986; Maton et aI., 1992). All
of these studies show the familiar increase in low frequency content in the EMG
signal during a variety offatiguing tasks yet, to be meaningful, any index of fatigue
must consider force, and studies in which force has not been measured (Palla &
Ash, 1981; Naeije & Zorn, 1981; van Boxtel et aI., 1983; Naeije, 1984; Kroon et
aI., 1986) are difficult to interpret in terms of force-producing capacity. Clark and
colleagues (1988) showed that increased low-frequency power in the masseter and
temporalis EMG power spectrum during prolonged isometric biting was not
accompanied by a reduction in MVC force. The change in power spectrum of the
surface EMG signal appears to be closely related to metabolic changes in the active
muscle and concomitant reductions in muscle fiber conduction velocity (Lind-
strom et aI., 1970; Eberstein & Beattie, 1985), but the relationship between the
EMG signal and force-producing capacity during fatigue is complex (Mills, 1982;
Sandercock et aI., 1985).
3. The ability to produce maximal force during repeated contractions appears to give
a reasonably objective measure of fatigue. In a comparison of the ability to
maintain repeated maximal contractions of short duration, van Steenberghe and
colleagues (1978) found that the jaw-closing muscles were not affected under
conditions in which the muscles producing hand-grip, and arm-flexion forces were
significantly weakened. The ability to preserve maximal force-producing capacity
during and following prolonged submaximal contractions at various proportions
of maximal force also appears to be greater in the jaw-closing muscles than other
muscles studied (Clark et aI., 1984; Clark & Carter, 1985).
420 T. S. Miles and M. A. Nordstrom

Fatigue of the jaw-opening muscles does not seem to have been studied in humans.
Jaw-opening muscles normally do not contract against a load, are very weak in relation to
jaw-closers, and are used intermittently. Fatigue in jaw-opening muscles is not likely to be
a problem in everyday life. Endurance tests of jaw-retruders (which include anterior
digastric) show a similar pattern as the jaw-closers (Clark & Adler, 1987).
In summary, quantitative information on the fatigability of the human jaw muscles
from whole-muscle studies of voluntary contractions is limited. There is evidence from
maximal biting force experiments that the jaw-closing muscles are more resistant to fatigue
of maximal contractions than non-masticatory muscles (van Steenberghe et aI., 1978; Clark
et aI., 1984; Clark & Carter, 1985). The paucity of relevant data from animal studies, and
the unusual histochemical appearance of the human jaw muscles, further complicate the
assessment of fatigability of the human jaw muscles in submaximal contractions. For these
reasons, it is useful to examine single motor units in the jaw muscles to estimate fatigability
during submaximal efforts.

Physiological Data from Single Motor Units in Humans

In contrast to the limb muscles, there are no animal data on the physiological
properties of individual motor units in the jaw muscles. The only information on this topic
comes from human studies employing the spike-triggered averaging (STA) technique
(Goldberg & Derfler, 1977; Yemm, 1977; Nordstrom & Miles, 1990; McMillan et aI., 1990).
With this technique, a needle or fine-wire electrode is inserted into the muscle to record the
activity of a single motor unit. Subjects maintain a low discharge rate of the motor unit with
the aid of feedback. The partially fused twitch force of a single motor unit is extracted from
the total force signal using ensemble averaging (see Thomas, Chapter 10). Yemm (1977)
reported a continuous range of contraction times for masseter units from 24-91 ms, with no
evidence for separate popUlations of fast and slow motor units. Goldberg and Derfler (1977)
also found a continuous distribution of contractile speeds in the masseter motor unit twitches,
but reported that they were distributed over a narrower range (38-69 ms). A similar range
(25-67 ms) was reported by McMillan and colleagues (1990).
Motor unit fatigability in human jaw muscles has been assessed only in the masseter
muscle (Nordstrom & Miles, 1990). Most (95%) were fast-twitch (mean twitch TTP ± SD
34 ± 10 ms) and showed a broad range of fatigability. Although most units were recruited at
less than 20% MVC, only 2 of37 (5%) were classified as type S units on the basis of their
twitch properties. Fig. I shows examples of twitch fatigue in 3 masseter motor units during
a 15-min isometric bite. The averaged motor unit twitches in the first and fifteenth minute
of the contraction are shown on the left. The time-course of the change in twitch amplitude
is shown immediately to the right for each motor unit. The units display a range offatigability,
with unit A being fatigue resistant, B fatiguing gradually over the fifteen minutes, and C
fatiguing rapidly in the ftrst few minutes without subsequent recovery. Three indices of
fatigue were calculated. 1) FI6 =100 x (the ratio of twitch amplitude in the 6th and 1st min).
This index facilitated comparison with other human motor unit fatigue studies (Table 2); 2)
FI I5a = 100 x (the ratio oftwitch amplitude in the 15th and 1st min); and 3) FII5b = 100 x
(the ratio of twitch amplitude in the 5th 3-min epoch and the first 3-min epoch).
There were non-signiftcant correlations of fatigability with both twitch amplitude
and contractile speed in the predominantly fast-twitch masseter motor units (Nordstrom &
Miles, 1990). There are several anomalies between the physiological data and the available
histochemical data for the human masseter. Despite the histochemical evidence that a
substantial proportion of the masseter is comprised of type I ftbers (Eriksson & Thornell,
1983), we found very few physiological type S units in the masseter. This is not unprece-
dented for physiological fast-twitch muscle (Goslow et aI., 1977). The human masseter
Fatigue of Jaw Muscles and Speech Mechanisms 421

Z
N o~--~----~--~
I
CJ

.:s IS

Figure 1. Twitch fatigue in 3 human mas-

seter motor units. Averaged motor unit
twitches calculated during the 1st and 15th
min of contraction for 3 different motor units
are shown on left in A-C. Uppermost trace in
each pair, twitch calculated during the 1st
min; arrow, time of occurrence of trigger.
Vertical calibration bar, 200 mN, horizontal
bar, 50 ms. Immediately to right of each pair
of twitches, mean twitch amplitude measured
in each I-min epoch for corresponding unit
is plotted over 15 min. Note different vertical
scale in each. Error bars, SE (adapted from
Nordstrom & Miles, 1990).

generally lacks the type IIA fibers which are believed to correspond to the physiological type
FR motor units (Eriksson & Thornell, 1983). However, we found a substantial popUlation
of fast-twitch, fatigue-resistant units (type FR) in masseter using STA. The most likely
explanation for the conflicting physiological and histological evidence is that at least some
type FR units in human masseter have muscle fibers which stain histochemically as type I.
This suggests that the correlation between the physiological properties of motor units and
the staining of their fibers for MATPase is not as strong as is presently believed. Unique
isoforms of myosin are present in the jaw-closing muscles of cats and other carnivores (Hoh
et ai., 1993). The cat jaw-closers contain superfast fibers, which contract isometrically at
about twice the speed as limb fast fibers, apparently because of their superfast myosin.
Although these fibers are present in most primates, they are believed to be absent in man
(reviewed in Hoh et ai., 1993). Although a relationship between histochemical fiber type and
motor unit fatigue resistance has been demonstrated in the cat hindlimb (Burke, 1981), there
is evidence that fatigue resistance correlates with the activity of oxidative enzymes, not
MATPase activity per se (Kugelberg & Lindegren, 1979; Hamm et ai. 1988). It seems
necessary to assess metabolic enzyme activities in the human jaw muscles to make reliable
estimates of fatigue tolerance of the tissues.
How does the fatigability of human masseter motor units compare with motor units
in other human muscles? A summary is presented in Table 2. The first 3 columns show the
data for the 3 fatigue indices used for masseter units by Nordstrom and Miles (1990). In a
study of human FDI motor units, Young and Mayer (1981) used intramuscular microstimu-
lation to activate 41 motor units. The fatiguing paradigm was 30 Hz pulse trains for 330 ms,
repeated at 1 S-l for 2 min, i.e., about 1200 stimuli. This test was designed to be similar to
the standard fatigue test which has been used to categorize motor unit popUlations of a
number of animal muscles (see Burke, 1981); however this test does not seem appropriate
for human muscle. Stimulation for 2 min is insufficient to induce contractile fatigue in human
FDI, as 83% of the units were potentiated by the test (FI > 100). Similar results were found
in human thenar muscles using intermittent 40 Hz stimulation of motor axons for 2 min
422 T. S. Miles and M. A. Nordstrom

Table 2. Distribution of fatigability among motor units in several human muscles

HumanFDI
Human masseter
Stephens and Young and HumanMG Human thenar
FI6 FIls. FI ISb Usherwood* Mayert Garnet et al. § Thomas et al.§§
FI> 75 56 34 19 62 90 55 72
FI25-75 28 38 59 24 7 39 28
FI < 25 16 28 22 14 3 6 0
Values represent percent of units studied with fatigue indices (FI) falling within a particular
range. See text for explanation ofFI 6, FIlS., FI ISb ' FDI, first dorsal interosseus; MG, medial
gastrocnemius.
*Stephens and Usherwood (1977).
tYoung and Mayer (1981).
§Garnett and colleagues (1978).
§§Thomas and coworkers (1991).

(Thomas et aI., 1991). Two minutes of activity is also insufficient for fatigue in the masseter,
as many units were potentiated during the first 2-4 minutes of continuous 10Hz activation.
It seems that human motor units are less fatigable than motor units studied in animals such
as cat and rat.
Stephens and U sherwood (1977) used STA to assess twitch fatigue of 22 motor units
in human FDI after 5 minutes of continuous activity at 10Hz. A pressure cuff was used to
prevent recovery. This fatigue index is directly comparable with our FI6 in masseter,
particularly as in both studies units were tested at comparable initial total force levels. In the
masseter study, 75% of tested units had an activation threshold below 10% of maximal force,
while in FDI approximately 63% of the tested units had an activation threshold below 10%
of maximal. The distribution of FIs after 5 minutes of activity was similar in the ischaemic
FDI and masseter units.
The only other fatigue data for human motor units is in the MG muscle (Garnett et
aI., 1978). Intramuscular micro stimulation was used to fatigue 18 units by activation at 10-20
Hz for 0.5 s, repeated every 1 or 2 s. The FI was the ratio of the initial twitch or tetanic
tension to that after 3000 stimuli, which is comparable with the FI6 in the masseter study.
Although the fatiguing paradigm activated a single unit in isolation (in contrast to the
situation with voluntary activation), the distribution of motor unit fatigability in this muscle
was also similar to the masseter.
This comparison suggests that the pattern of contractile fatigue in masseter motor
units is similar to that seen in FDI and MG after about 3000 activations. Although different
factors undoubtedly influence fatigability in maximal and low-force isometric contractions,
the present findings suggest that the superiority of the masseter in repeated high-force
contractions (van Steenberghe et aI., 1978; Clark et aI., 1984; Clark & Carter, 1985) is not
due to an inherent advantage in aerobic fatigue-resistance of its low- and moderate-threshold
motor units. Therefore, other explanations, such as superior oxygen delivery become more
likely (van Steenberghe et aI., 1978).
The motor unit organization in the masseter should be considered from a functional
viewpoint. Unlike the limb muscles, the masseter has minimal tonic postural activity. At rest,
the jaw is supported against gravity mostly by passive tension, with perhaps some activity
in temporalis but not in masseter (Yemm, 1976). The paucity of physiological type S units
in the masseter is therefore not surprising. The masseter is involved in a wide variety of
activities which include mastication, speech, swallowing, and facial expression. These
diverse tasks, and the rapid, intermittent activation of motor units necessary for many of
Fatigue of Jaw Muscles and Speech Mechanisms 423

these functions may produce quite different functional demands on the muscles than the more
stereotyped functions of the limb muscles.

FATIGUE OF THE MUSCLES OF PHONATION

The primary role of the laryngeal muscles is to protect the airways against the
invasion of foreign objects. In some species, notably humans, they have evolved a
secondary role in sound production. Although any anesthetist will relate experiences of
laryngospasm, in which the laryngeal muscles continue to contract until death is imminent,
there is comparatively little direct information about fatigue properties of the laryngeal
muscles.
Human voice muscles contain a higher proportion of oxidative fibers than sub-pri-
mate species. Histochemically, the posterior cricoaretynoid muscle in humans consists
mostly of type I fibers, while the thyroaretynoid has about 42% type I fibers and 58% type
II fibers (see Cooper & Rice, 1990). Claassen and Werner (1992) reported that type I, IIA
and lIB fibers were present in the human thyroarytenoid and posterior cricoarytenoid
muscles. The percentage of type lIB fibers was low in the thyroarytenoid and posterior
cricoarytenoid muscles and in the other laryngeal muscles. The type IIA fiber content of the
arytenoid muscle was significantly higher than in the posterior cricoarytenoid muscle.
Cooper and Rice (1990) stimulated canine vocal fold muscles with trains of electrical
stimuli and observed that they are resistant to fatigue, but it is not clear how this would relate
to the work done by these muscles in phonation when the contractions occur in brief,
high-speed bursts. Robin and coworkers (1992) examined the strength and endurance of
human tongue muscles in subjects who had acquired high skill levels with their tongues. The
maximal strength of the tongue muscles was found to be similar in skilled and control
subjects but, when pressing with their tongues on air-filled bulb, the skilled group could
maintain 50% of their maximal pressure for longer than the controls, indicating an increased
endurance time.
For obvious technical reasons, there is little direct information on the fatigability of
speech muscles in humans. However, one interesting clinical syndrome which is believed to
represent a discrete vocal fatigue syndrome has been described in professional voice users
(Koufman & Blalock, 1988). This syndrome is characterized by muscle tension in the neck,
poor control of the breath stream and an abnormally low-pitched speaking voice in both men
and women. The evocative expression Bogart-Bacal/ syndrome has been coined to describe
the characteristically husky voice in this disorder.
In summary, it is apparent that motor units in the human masticatory muscles possess
combinations of motor unit physiological, histochemical, and anatomical properties which
differ from the relationships widely accepted for animal limb muscles. One can only
speculate that their unique features reflect their complex functional requirements. The fatigue
properties of the speech muscles await further elucidation.

ACKNOWLEDGMENTS

The authors' laboratories are supported by the National Health & Medical Research
Council of Australia. Attendance of T.S.M at the 1994 Bigland-Ritchie conference was
supported, in part, by a grant from the Muscular Dystrophy Association (USA) and the
University of Arizona Regents' Professor funds of Douglas Stuart.
424 T. S. Miles and M. A. Nordstrom

REFERENCES
Asmussen E (1979). Muscle fatigue. Medicine Science and Sports 11, 313-32l.
Bredman n, Weijs WA, Moorman AMF (1990). Expression of "cardiac specific" myosin heavy chain in rabbit
cranial muscles. In Marechal G, Carraro U (eds.), Muscles and Motility, pp. 329-335. Andover,
Hampshire, Intercept.
Burke RE (1981). Motor units: anatomy, physiology, and functional organization. In: Brookhart JM,
Mountcastle VB (sec. eds.), Brooks VB (vol. ed.), Handbook of Physiology, sec. 1, vol. II. pt 1, The
Nervous System: Motor Control., pp. 345-422. Bethesda, MD: American Physiological Society.
Christensen LV (1979). Some subjective-experimental parameters in experimental tooth clenching in man.
Journal of Oral Rehabilitation 6, 119-136.
Christensen LV & Mohamed SE (1983). The possible activity of large and small jaw muscle units in
experimental tooth clenching in man. Journal of Oral Rehabilitation 10,519-525.
Christensen LV, Mohamed SE & Rugh JD (1985). Isometric endurance of the human masseter muscle during
consecutive bouts of tooth clenching. Journal of Oral Rehabilitation 12, 509-514.
Claassen H & Werner JA (1992). Fiber differentiation of the human laryngeal muscles using the inhibition
reactivation myofibrillar ATPase technique. Anatomy and Embryology 186, 341-346.
Clark GT & Adler RC (1987). Retrusive endurance, fatigue and recovery of human jaw muscles at various
isometric force levels. Archives of oral Biology 32, 61-65.
Clark GT, Beemsterboer PL & Jacobsen R (1984). The effect of sustained submaximal clenching on maximum
bite force in myofascial pain dysfunction patients. Journal of Oral Rehabilitation 11, 387-391.
Clark GT & Carter MC (1985). Electromyographic study of human jaw-closing muscle endurance, fatigue and
recovery at various isometric force levels. Archives of oral Biology 30,563-569.
Clark GT, Carter, MC & Beemsterboer PL (1988). Analysis of myoelectric signals in human jaw closing
muscles at various isometric force levels. Archives of oral Biology 33,833-837.
Cooper DS & Rice DH (1990). Fatigue resistance of canine vocal fold muscle. Annals of Otolaryngology
Rhinology and Laryngology 99, 228-233.
Dubowitz V & Brooke MH (1973). Muscle Biopsy- A Modern Approach. London: Saunders.
Eberstein A & Beattie B (1985). Simultaneous measurement of muscle conduction velocity and EMG power
spectrum changes during fatigue. Muscle & Nerve 8, 768-773.
Edstrom L & Grimby L (1986). Effect of exercise on the motor unit. Muscle & Nerve 9, 104-126.
Eriksson P-O, Eriksson A, Ringquist M & Thornell L-E (1981). Special histochemical muscle-fibre charac-
teristics of the human lateral pterygoid muscle. Archives of oral Biology 26, 495-507.
Eriksson P-O, Eriksson A, Ringquist M & Thornell L-E (1982). Histochemical fibre composition of the human
digastric muscle. Archives of oral Biology 27, 207-215.
Eriksson P-O & Thornell L-E (1983). Histochemical and morphological muscle-fibre characteristics of the
human masseter, the medial pterygoid and temporal muscles. Archives of oral Biology 28, 781-795.
Garnett RAF, O'Donovan MJ, Stephens JA & Taylor A (1978). Motor unit organisation of human medial
gastrocnemius. Journal of Physiology (London) 287, 33-43.
Gibbs CH, Mahan PE, Mauderli A, Lundeen HC & Walsh EK (1986). Limits of human bite strength. Journal
of Prosthetic Dentistry 56, 226-229.
Goldberg LJ & Derfler B (1977). Relationship among recruitment order, spike amplitude, and twitch tensions
of single motor units in human masseter muscle. Journal of Neurophysiology 40, 879-890.
Goslow GE, Cameron WE & Stuart DG (1977). The fast twitch motor units of cat ankle flexors. 1. Tripartite
classification on basis of fatigability. Brain Research 134, 35-46.
Hamm TS, Nemeth PM, Solanki L, Gordon DA, Reinking RM & Stuart DG (1988). Association between
biochemical and physiological properties in single motor units. Muscle & Nerve 11, 245-254.
Hoh JFY, Hughes S, Kang LDH, Rughani A & Qin H (1993). The biology of cat jaw-closing muscle cells.
Journal of Computer-Assisted Microscopy 5, 65-70.
Koufman JA & Blalock PD (1988). Vocal fatigue and dysphonia in the professional voice user. Laryngoscope
98, 493-498.
Kroon GW Naeije M & Hansson TL (1986). Electromyographic power-spectrum changes during repeated
fatiguing contractions of the human masseter muscle. Archives of oral Biology 31, 603-608.
Kugelberg E & Lindergren B (1979). Transmission and contraction fatigue of rat motor units in relation to
succinate dehydrogenase activity of motor unit fibres. Journal ofPhysiology (London) 288, 285-300.
Lindstrom L & Hellsing G (1983). Masseter muscle fatigue objectively quantified by analysis of myoelectric
signals. Archives of oral Biology 28, 297-301.
Lindstrom L, Magnusson R & Petersen I (1970). Muscular fatigue and action potential conduction velocity
changes studied with frequency analysis ofEMG signals. Electromyography 4,341-353.
Fatigue of Jaw Muscles and Speech Mechanisms 425

Maton B, Rendell J, Gentil M & Gay T (1992). Masticatory muscle fatigue: endurance times and spectral
changes in the electromyogram during the production of sustained bite forces. Archives of oral
Biology 37,521-529.
Maxwell LC, Carlson DS, McNamara JA & Faulkner JA (1979). Histochemical characteristics of the masseter
and temporalis muscles of the rhesus monkey (Macaca mulatta). Anatomical Record 193, 389-403.
McMillan A, Sasaki K & Hannam AG (1990). The estimation of motor unit twitch tensions in the human
masseter muscle by spike-triggered averaging. Muscle & Nerve 13, 697-703.
Mense S (1977). Muscular nociceptors. Journal of Physiology (Paris) 73, 233-240.
Mills KR (1982). Power spectral analysis of electromyogram and compound muscle action potential during
muscle fatigue and recovery. Journal of Physiology (London) 326, 401-409.
Naeije M (1984). Correlation between surface electromyograms and the susceptibility to fatigue of the human
masseter muscle. Archives of oral Biology 29, 865-870.
Naeije M & Zorn H (1981). Changes in the power spectrum of the surface electromyogram of the human
masseter muscle due to local muscle fatigue. Archives of oral Biology 26, 409-412.
Noden DM (1983). The embryonic origins of avian cephalic and cervical muscles and associated connective
tissues. American Journal ofAnatomy 168, 257-276.
Nordstrom MA & Miles TS (1990). Fatigue of single motor units in human masseter. Journal of Applied
Physiology 68, 26-34.
Nordstrom SH & Yemm R (1974). The relationship between jaw position and isometric active tension produced
by direct stimulation of the rat masseter muscle. Archives of oral Biology 19, 353-359.
Palla S & Ash Jr MM (1981). Power spectral analysis of the surface electromyogram of human jaw muscles
during fatigue. Archives of oral Biology 26.547-553.
Ringquist M (1971). Histochemical fibre types and fibre sizes in human masticatory muscles. Scandinavian
Journal of Dental Research 79, 366-368.
Ringquist M (1973). Fibre sizes of human masseter muscle in relation to bite force. Journal of Neurological
Sciences 19, 297-305.
Ringquist M, Ringquist I, Eriksson P-O & Thornell L-E (1982). Histochemical fibre type profile in the human
masseter muscle. Journal of Neurological Sciences 53, 273-282.
Robin DA, Goel A, Somodi LB & Luschei ES (1992). Tongue strength and endurance: relation to highly skilled
movements. Journal of Speech and Hearing Research 35, 1239-1245.
Rowlett AE (1932-33). The gnathodynamometer and its use in dentistry. Proceedings of the Royal Society for
Medicine 26, 463-471.
Sandercock TG, Faulkner JA, Albers JW & Abbrecht PH (1985). Single motor unit and fiber action potentials
during fatigue. Journal of Applied Physiology 58, 1073-1079
Serratrice G, Pellisier JF, Vignon C & Baret J (1976). The histochemical profile of the human masseter. An
autopsy and biopsy study. Journal of Neurological Sciences 30, 189-200.
Soussi-Yanicostas N, Barbet JP, Laurent-WinterC, Barton P & Butler-Browne GS (1990). Transition of myosin
isozymes during development of human masseter muscle. Persistence of developmental isoforrns
during postnatal stage. Development 108, 239-249.
Stephens JA & Usherwood TP (1977). The mechanical properties of human motor units with special reference
to their fatiguability and recruitment threshold. Brain Research 125, 91-97.
Tamari JW, Tomey GT, Ibrahim MZM, Baraka A, Jabbur SJ & Bahuth N (1973). Correlative study of the
physiologic and morphologic characteristics of the temporal and masseter muscles of the cat. Journal
of Dental Research 52, 538-543.
Taylor A, Cody FWJ & Bosley MA (1973). Histochemical and mechanical properties of the jaw muscles of
the cat. Experimental Neurology 38, 99-109.
Thexton AJ & Hiiemae KM (1975). The twitch tension characteristics of opossum jaw musculature. Archives
of oral Biology 20, 743-748
Thomas CK, Johansson RS, Bigland-Ritchie B (1991). Attempts to physiologically classify human thenar
motor units. Journal ofNeurophysiology 65, 1501-1508.
Van Boxtel A, Goudswaard P, van der Molen GM & van den Bosch WEJ (1983). Changes in electromyogram
power spectra of facial and jaw-elevator muscles during fatigue. Journal of Applied Physiology 54,
51-58.
Van Steenberghe D, De Vries JH & Hollander AP (1978). Resistance of jaw-closing muscles to fatigue during
repetitive maximal voluntry voluntary clenching efforts in man. Archives oforal Biology 23, 697-70 I.
Vignon C, Pellisier JF & Serratrice G (1980). Further histochemical studies on the masticatory muscles.
Journal of Neurological Sciences 45, 157-176.
Waltimo A & Kononen M (1993). A novel bite force recorder and maximal isometric bite force values for
healthy young adults. Scandinavian Journal ofDental Research 101, 171-175.
426 T. S. Miles and M. A. Nordstrom

Waugh LM (1937). Dental observations among Eskimo. VII. Survey of mouth conditions, nutritional study
and gnathodynamometer data, in most primitive and populous native villagers in Alaska. Journal of
Dental Research 16, 355-356.
Yemm R (1976). The role of tissue elasticity in the control of mandibular resting posture. In: Anderson DJ,
Matthews B (eds.), Mastication, pp. 81-89. Bristol: John Wright and Sons.
Yemm R (1977). The orderly recruitment of motor units of the masseter and temporal muscles during voluntary
isometric contraction in man. Journal of Physiology (London) 265, 163-174.
Young JL & Mayer RF (1981). Physiological properties and classification of single motor units activated by
intramuscular microstimulation in the first dorsal interosseous muscle in man. In: Desmedt JE (ed.),
Progress in Clinical Neurophysiology, vol 9, Motor Unit Types, Recruitment and Plasticity in Health
and Disease, pp. 17-25. Basel: Karger.
SECTION IX
Fatigue of Adapted Systems: Overuse, Underuse and
Pathophysiology

The first chapters in this section deal with the adaptability of the neuromuscular
system to variations in usage and to the unavoidable effects of occasional damage, repair
and aging. The final two chapters consider how pathological changes in muscle function
should be approached in individual patients.
Muscles appear well adapted to the tasks they perform, in terms of endurance,
strength and speed. However, the mechanisms that underlie this apparent adaptation and the
limits to which it can be extended are less well understood. In Chapter 32, Gordon reviews
the extensive evidence that the neuromuscular system exhibits remarkable plasticity such
that it can respond dramatically to altered muscle use. Some limits to the various models of
altered use are discussed. Overall, the forces exerted by muscles seem to preferentially affect
strength, while the total daily amount of muscle activity sets the endurance capability.
Systemic effects produced by the altered use are also likely to contribute to the performance
capabilities of individual muscles. Review of this field emphasizes that motoneurons do not
exert total trophic control over their muscle fibers. Rather the motoneuron modulates the
properties of its muscle fibers within a limited (adaptive) range, which is partly a function
of muscle architecture and mechanical factors, and partly a function of intrinsic factors preset
during development (i.e., muscle genotype). Material in this chapter is relevant to a variety
of pathological circumstances, as well as the development of functional electrical stimulation
in rehabilitation, and understanding the effects of long-term space flight.
Eccentric exercise (i.e., lengthening contractions) produces familiar deleterious
short-term effects that are manifested as delayed muscle soreness and increased fatigue. This
syndrome is reviewed in Chapter 33 (Clarkson & Newham). Eccentric exercise is metabo-
lically attractive (as it uses less oxygen) and biomechanically useful (as it stores elastic
energy, which can be used in the stretch-shortening cycle), but it is especially likely to
damage muscle. This effect may partly be mediated by an uneven distribution of sarcomere
strengths, such that the weakest sarcomeres pop during muscle lengthening, although other
factors are likely to be involved. The exact biochemical triggering mechanisms within the
active muscle cell that are responsible for the damage are not yet clear. However, the damaged
cells release a cocktail of substances that induce reflex effects through the activation of
small-diameter muscle afferents.
The decline in muscle strength and power accompanying aging is highlighted in
Chapter 34 (Faulkner & Brooks). These decreased capabilities are due to a reduced number
and size of muscle fibers pursuant to a loss of motoneurons. Although inevitable, the
428 Fatigue of Adapted Systems: Overuse, Underuse and Pathophysiology

physiological consequences can be partly offset by strength and endurance training. Intra-
cellular processes are likely to mediate the decline in performance because the maximal force
achieved by high calcium concentrations in permeabilized fibers of old animals is the same
as that in younger animals. Large, fatigable motor units, with their extensive territories, are
likely to be preferentially affected by aging. Contraction-induced damage may underlie the
denervation of fast fibers with consequent reinnervation by slow fibers.
In the final two chapters of this section, some principles for understanding the patient
with fatigue are presented. Edwards and colleagues (Chapter 35) describe the major
advances in understanding human muscle fatigue from an historical perspective. These range
from the technical advances which made new measurements possible (e.g., myographs,
needle muscle biopsy, 3 1p NMR), new conceptual advances, and insights that resulted from
understanding pathophysiology, such as the deficits attributable to single biochemical
mechanisms. The focus is extended beyond the 1980 symposium in London to then propose
a new way of categorizing the muscle fatigue found in patients with neuromuscular and other
disorders. In Chapter 36 (McComas and colleagues), the clinical approach to the patient
with fatigue is highlighted with some upper and lower motoneuron diseases discussed,
including mUltiple sclerosis and poliomyelitis. There is a need to make comparatively simple
assessments of muscle performance more often in the clinic.
 32

FATIGUE IN ADAPTED SYSTEMS

Overuse and Underuse Paradigms

T. Gordon

Department of Pharmacology
Division of Neurosciences
University of Alberta
Edmonton, Canada, T6G 2S2

ABSTRACT

Alterations in structural and biochemical properties of muscles that underlie physiological

parameters of contractile force, speed and fatigability are described under conditions of 1)
overuse: imposed electrical stimulation, natural exercise and functional overload; 2) rein-
nervation of denervated muscles; and 3) underusage: conditions of restricted use after spinal
cord injury, weightlessness, immobilization and drug-induced neuromuscular blockade.
These conditions demonstrate the remarkable plasticity of muscle fibers with obvious
implications in health and disease. They also identify that the amount of neuromuscular
activity and loading of muscle contractions are major factors determining susceptibility to
fatigue and muscle strength, respectively.

INTRODUCTION

The capacity of some skeletal muscles but not others to contract with little fatigue
has often been related to their function. In the male frog, for example, the slow tonic muscles
in the forearm develop slow contractures that can be maintained without fatigue for many
hours (Kuffler & Vaughn-Williams, 1953). These muscles are clearly adapted for their
function of clasping the female during mating. In the bird, the slow-tonic anterior latissimus
dorsi muscle can maintain sufficient force to successfully hold the wings close to the body
when the bird is not in flight. In mammals, the slow-twitch soleus muscle is ideally suited
to maintain tension at the low frequencies of motoneuron discharge during posture. In
contrast, fast-twitch muscles which develop much larger forces at higher frequencies of
discharge are far more susceptible to fatigue but they are well suited to the development of
high forces during movement. In all species, recruitment of the fast-twitch muscle fibers in
the hindlimb can generate large forces for short periods required to accelerate the body during
walking, running and jumping.

429
430 T. Gordon

Differences in the susceptibility of muscle to fatigue were initially associated with

obvious differences in the color of red and white skeletal muscles (Kuhne, 1863; Ranvier,
1874). These in tum, were recognized to indicate differences in capillary density and
myoglobin content (Ranvier, 1874; reviewed by Burke & Edgerton, 1975; Burke, 1981;
Saltin & Gollnick, 1983; Vrbova et aI., 1994). Red and white muscles also differ in their
rates of contraction. Red muscles were, therefore, denoted as red slow muscles and distin-
guished from white fast-twitch muscles (Denny-Brown, 1929). Slow contraction speeds
combined with low fatigability are recognized as properties ideally suited for the continuous
low force, high endurance requirements of postural slow-twitch muscles as opposed to the
intermittent high force, low endurance characteristics of the fast-twitch muscles (Vrbova et
aI., 1994).
The matching of muscle properties to their characteristic activation patterns begged
the question of whether the patterns of muscle activation are actually responsible for the
expression of the slow or fast muscle phenotype. The natural extension of this question is
how far muscle properties can adapt to altered usage in health and disease. These questions
have been addressed directly or indirectly in many hundreds of studies of mature muscle
phenotype, development, capacity for change during adult life, and for recovery after injury.
Recent reviews of these studies include those of Roy and colleagues (1991), Henriksson
(1992), Pette and Vrbova (1992); Gordon and Pattullo (1993) and Vrbova and colleagues
(1994). Dramatic alterations in structural and biochemical properties that underlie the
capacity of the muscle to develop force (muscle strength) over time (contractile speed) and
to sustain muscle force (muscle endurance) have been demonstrated after the nerve supply
(self- and cross-reinnervations) or the amount of neuromuscular activity (overuse and
underuse paradigms) is altered. These experiments point to the remarkable plasticity of
muscle fibers. Because motor control depends on muscle power, speed and endurance, the
potential for manipulating muscle properties for required function was immediately recog-
nized. Many applications have been extensively explored including sports training, rehabili-
tation of muscles after trauma, injury and disease, paralysis after lower and upper
motoneuron lesions, limb immobilization, and space flight. To provide insight into this topic,
the extent of muscle plasticity is considered in this chapter with emphasis on muscle
endurance. The matching properties of muscles and their motoneurons are briefly reviewed
prior to considering adaptability of muscle endurance under conditions of I) overuse:
imposed electrical stimulation of muscle, natural exercise and functional overload of
muscles; 2) reinnervation of denervated muscles; and 3) under-usage: conditions of restricted
use after spinal cord injury, weightlessness, immobilization and drug-induced neuromuscular
blockade.

METABOLIC PROFILE OF SLOW AND FAST SKELETAL

MUSCLE IN NORMAL FUNCTION

High resistance to fatigue of slow-twitch muscles is associated with dense capillari-

zation, sustained blood flow through active muscles and generation of ATP via carbohydrate
and fat oxidation. Anaerobic glycolysis which generates only 3 moles of ATP per mole of
glycogen plays a minor role in contrast to aerobic metabolism via the mitochondrial pathway
which generates 39 moles of ATP per mole of glycogen (Fox et aI., 1988). In fast-twitch
muscles, relatively poor blood flow during muscle contraction is associated with more
reliance on anaerobic pathways and lower resistance to fatigue. Thus, reciprocal distribution
of oxidative and glycolytic enzymes appears to be well adapted to the high and low endurance
capabilities of slow- and fast-twitch muscles respectively (Vrbova et aI., 1994).
Fatigue in Adapted Systems 431

A. Fatigue Test

FF FI FR

B. Glycogen Depletion

FF

FI

t
FR

2 m in

Figure 1. Fatigability of fast motor units in rat tibialis anterior muscle. A, decline in isometric force during a
2-min period of 300 ms duration tetanic stimulation at 40 Hz repeated every min. Force declines to less than
25% of initial force in FF units, between 25% and 75% in FI units and to more than 75% in FR units. B,
repetitive tetanization with short bursts of 100 Hz stimulation (50 ms tetanus) repeated at progressively faster
rates leads to fatigue in fast units. The pattern of fatigue was different for the 3 fast unit types with force
declining most rapidly in the FF units. The repetition rate was increased at each arrow at the time point when
force had declined to a steady level. The arrows from left to right are 1 Hz, 1.25 Hz, 1.7 Hz, 2.5 Hz and 0.5
Hz. The latter frequency was used to permit recovery of force. This tetanization protocol was repeated 6 or 7
times for each unit and is associated with complete glycogen depletion of the stimulated muscle fibers.

Presence of oxidative and glycolytic enzymes and their specific activities also reflect
differences in substrate availability. Thus slow-twitch muscle fibers have large quantities of
stored triglycerides which provide energy via 13-oxidation and ATP generation via the Krebs
cycle in contrast to the large glycogen stores in fast-twitch muscle fibers which rely on
anaerobic glycolysis (Pette & Staron, 1990; for additional discussion see Kushmerick,
Chapter 4).
Yet, oxidative and glycolytic enzymes are not always reciprocally related as both
oxidative and glycolytic enzyme activities are high in some muscle fibers. These fibers were
identified histochemically and designated as fast oxidative glycolytic fibers (FOG) on the
basis of their acid labile myosin ATPase and high enzyme activities for oxidative and
glycolytic enzymes (Peter et al., 1972). These were distinguished from I) the fast glycolytic
(FG) fibers which contained acid labile myosin ATPase, associated with fast isoforms of
myosin heavy chains, and have high glycolytic but not oxidative enzyme activity (FG); and
2) the slow oxidative (SO) fibers which contained acid stable myosin ATPase and are highly
reactive for oxidative but not glycolytic enzyme activity. Although an earlier classification
of fibers into acid stable Type I and acid labile type IIA and lIB fibers is based only on pH
sensitivity of mATPase (Brooks & Kaiser, 1970), there is generally good correspondence
432 T. Gordon

between the two systems (type I: SO; type IIA: FOG; type lIB: FG; Gordon & Pattullo,
1993).
By isolating single motor units and recording twitch and tetanic contractions to
determine contractile force, speed and endurance, Burke and his colleagues (1973) classified
motor units into slow fatigue resistant (S), fast fatigue resistant (FR), fast fatigue intermediate
(Fint) and fast fatigable (FF). Following repetitive tetanization to deplete muscle unit fibers
of glycogen for later recognition and muscle fiber typing, they showed that the Sand FR
units contained SO and FOG fibers, respectively (see also Totosy de Zepetnek et ai., 1992b).
Fint and FF units contained FG fibers. Thus at the single motor unit level, muscle endurance
is correlated with metabolic capacity. Both Sand FR units have high oxidative capacity
which is associated with high resistance to fatigue during tetanic stimulation. The FF motor
units which have low oxidative capacity and demonstrate high glycolytic enzyme activity
are readily fatigable.
Motor units are normally recruited from the smallest and slowest contracting S units
to the fastest and most forceful FF units (Henneman & Mendell, 1981). As a result, the FF
units are recruited infrequently. When they are recruited, as for example in maximal
voluntary contractions or during short rapid movements such as jumping, blood flow can be
occluded. Although anaerobic glycolysis can generate 3 ATP per mole of glycogen, less is
generated due to the accumulation of lactic acid which inhibits the rate limiting enzyme
phosphofructokinase (PFK; Triveldi & Danforth, 1966). Therefore, brief intermittent acti-
vation of the most fatigable units is the most energy efficient method of recruiting the
strongest and most fatigable motor units.
Glycogen stores are highest in FG fibers (Saltin & Gollnick, 1983). As a result,
intermittent contractions in which there is a high work-to-rest ratio but in which lactic acid
accumulation is minimized is the most effective method of utilizing glycogen stores. This
principle is readily tested for single motor units in which stimulation regimes designed to
elicit brief intermittent, high force tetani at progressively higher stimulation rates of 1 to 2.5
Hz were very effective in depleting all unit types of glycogen (Totosy de Zepetnek et aI.,
1992b; their Fig. 1). The more fatigue resistant units, however, required considerably longer
periods of stimulation for fatigue and corresponding depletion of glycogen.
Skeletal muscles differ considerably in the proportion of the different types of motor
units and their constituent muscle fiber types (Ariano et ai., 1973; reviewed by Saltin &
Gollnick, 1983). Generally, most muscles are mixed but contain either a majority of slow or
fast muscle fibers depending on their function. Many muscles contain motor units that
demonstrate a continuous distribution of fatigability and corresponding oxidative and
glycolytic potential, for example the rat tibialis anterior (TA; Kugelberg & Lindegren, 1979;
Totosy de Zepetnek et aI., 1992b).

METABOLIC AND FUNCTIONAL ADAPTATION TO INCREASED

USAGE DURING CHRONIC ELECTRICAL STIMULATION

Fast-to-Slow Conversion
In fast-twitch muscles, synchronous low-frequency stimulation of all muscle fibers
for 3-24 hrs per day leads to a reproducible change in the ratio of oxidative to glycolytic
enzyme activities and an associated increase in their resistance to fatigue (reviewed by Pette
& Vrbova, 1992). Within days of daily stimulation, mRNA levels and the activities of
enzymes of the citric acid cycle, fatty acid ~-oxidation and the respiratory chain increase
significantly (Pette et aI., 1973; Henriksson et aI., 1986; Williams et aI., 1987; Hood et aI.,
Fatigue in Adapted Systems 433

Alk mATPase NADH a-GP

Control

56 days
stim

76 days
stim

Figure 2. Fast-to-slow conversion of chronically stimulated cat medial gastrocnemius muscle. Daily low-fre-
quency (20 Hz) stimulation in a 50% duty cycle (2.5 son, 2.5 s off) led to increased expression ofNADH and
downregulation of a-glycerophosphate. The characteristic mosaic of muscle fibers in the normal muscle was
replaced by a homogeneous fiber population in which oxidative enzyme (NADH) activity was high and
glycolytic enzyme (a-OP) activity is low. This change was associated with a conversion of myosin heavy chain
composition of the muscle fibers to the slow form, which accounts for the loss of all alkali-stable mATPase.

1989). These enzymes include succinic dehydrogenase and NADH, which increase concur-
rently with a decline in glycolytic enzymes such as a-glycerophosphate (Fig. 2). As shown
in Fig. 3, the increase in oxidative- and decline in glycolytic enzyme activities follow an
exponential time course. The time constants depend on the total amount of neuromuscular
activity, being significantly longer when muscles were stimulated for 12 hrs than for 24 hrs
(Pette & Vrbova, 1992). Consistent with the normal pattern in which enzyme activities are
jointly regulated (Pette & Luh, 1962; Pette & Hofer, 1980), the conversion of glycolytic to
oxidative potential is accompanied by an increase in the number of mitochondria per muscle
fiber (Williams et aI., 1986). The most rapid change is an increased activity of hexokinase,
a glycolytic enzyme associated with uptake and oxidation of blood glucose (Bass et aI.,
1969). This is the only enzyme of the glycolytic pathway to be elevated by chronic
stimulation (Pette et aI., 1973). The transient increase presumably reflects the early response
to high activity with generation of ATP via glycolysis and metabolism of pyruvate via the
citric acid cycle rather than to lactate.
Associated with the metabolic adaptation, there is an early and dramatic increase in
capillary density surrounding the stimulated muscle fibers (Brown et aI., 1976; Hudlicka et
al., 1977; Hudlicka, 1984) and an increase in myoglobin content (Pette et aI., 1973). Thus,
the high metabolic demand of the active muscles results in an increased oxygen uptake and
metabolism to generate ATP for contraction. A significant decline in muscle fiber diameter
434 T. Gordon

A B
70 Citrate Synthase 1500 Glyceraldehyde-3P dehydrogenase
Gi'

':"\,
~ 60 i
:::J
E 50
!?!
2-
~_2......... Diap~ragm
40
Heart I
~
> 30
g 20
Diaphragm I
600 -
~ §
_f"---_-OD

CD
E Soleus • 300
~ 10 Soleus. !leart
c:
W
o I' o I'
0 20 40 60 80 100 120 o 20 40 60 eo 100 120

Stimulation time (days)

Figure 3. Time course of changes in the activity levels of two representative enzymes of aerobic (citrate
synthase) and anaerobic (glyceraldehyde phosphate dehydrogenase) energy metabolism in rabbit tibialis
anterior muscle as induced by chronic low-frequency stimulation (10 Hz, 12 hr daily), For comparison,
activities of the two enzymes were also detennined in the diaphragm, soleus, and cardiac muscles (means ±
SE, n= 3-5; modified from Pette & Vrbova, 1992),

is also consistent with more efficient oxygen delivery to the mitochondria for oxidative
metabolism (Pette et aI., 1976; Donselaar et aI., 1987),
Concurrent with reduced muscle fiber size, muscle force declines reciprocally with
muscle endurance within the first month of stimulation (Fig, 4), As reviewed elsewhere,
contractile speed and underlying changes in regulatory and contractile proteins occur with
a slower time course. The extent of fast-to-slow conversion is species-dependent, possibly
involving differences in thyroid status and/or adaptive ranges of fast muscles in different
species (Pette & Vrbova, 1992; Gordon & Pattullo, 1993; Vrbova et aI., 1994). Except in the
rat where conversion of phenotype is more limited, slowing of contraction is associated with
slower release of Ca2 + from the sarcoplasmic reticulum via ryanodine/Ca2+ release channels
(Ohlendieck et aI., 1991), expression of slow isoforms of the regulatory proteins troponin
and tropomyosin (Roy et aI., 1979; Hartner et aI., 1989), and reduced rate of cross-bridge
cycling as slow forms of myosin heavy and light chain isoforms are incorporated into the
thick filaments (Sreter et aI., 1973; Pette et aI., 1976; Sweeney et aI., 1988). In addition, the
active state is prolonged as a result of reduced sarcoplasmic reticulum Ca2+-ATPase activity.
This, in turn, is associated with expression of the slow isoform of the enzyme and the
regulatory protein, phospholamban and the downregulation of Ca2+ binding proteins, calse-
questrin in the sarcoplasmic reticulum and parvalbumin in the cytoplasm (Leberer et aI.,
1986; 1989). Reduced t-tubular volume and enhanced Na+/K+ pumping (Eisenberg &
Salmons, 1981) may be responsible for preventing K+ accumulation and thereby ensuring
transmission of muscle action potentials (Kernell et aI., 1987b). Incorporation of increased
numbers of myonuclei in stimulated muscles is consistent with the higher numbers in slow
as compared to fast muscle fibers (Russell et aI., 1992).
Thus a comprehensive documentation of changes in metabolism, protein expression
and function has shown that adult skeletal muscle adapts to the mechanical signals imposed
by electrical stimulation. This involves switching specific genes for conversion of muscle
fibers to slow phenotype. Different genes are up-regulated or downregulated resulting in the
metabolic adaptation with the result that muscles become more fatigue-resistant.
Despite the many studies that document the phenotypic changes in stimulated
muscles, the question of how the changes come about has been largely neglected. The trigger
for change may be linked to the muscle fatigue that occurs during repetitive low-frequency
Fatigue in Adapted Systems 435

A
1.0
• • • •
x
OJ
"0
5
O.B 0
0
A()
- 0
• •
0.6 0
OJ
U
c re 0
<tl
=:;
0.4
~" •
IiI-
0
"0
C "
• ..•
0.2 .Ii c.
W
0.0
0 ,0 20 30 40 50 60 70 BO 90 ,00

BO B
en
g
Figure 4. Fast-to-slow conversion of chronic al- OJ
E 60
Iy stimulated cat medial gastrocnemius muscle. i=
Daily low-frequency (20 Hz) stimulation in a c
0
50% duty cycle (2.5 son, 2.5 s off) led to: A, U 40
increased resistance to fatigue; endurance in- ~
C
dex: tetanic force recorded 2 min after tetanic 0
()
stimulation at 13 pulses at 25 Hz every s for 2 20
0 ,0 20 30 40 50 60 70 BO 90 100
min relative to tetanic force at 0 min; B, slowing
1.0 C
of the contraction time; time to peak twitch c
0
force; and C, fall in contractile tension; isomet- '(ij
O.B
C
ric tension developed in response to 100 Hz OJ
I-
stimulation normalized to tension developed by "0
0.6 •
the muscles prior to onset of daily electrical
Q)
.t:' 0.4
",*
~
A •
. •
Cii
stimulation. The data for each of 8 muscles are E to. "
plotted with separate symbols (from Gordon & 0 0.2
z
Mao, 1994; reprinted from Physical Therapy
0.0
with the permission of the American Physical 0 10 20 30 40 50 60 70 BO 90 100
Therapy Association). Days of Stimulation

stimulation (Fig. 5A). The low levels of force generated by the unfused tetanic contractions
do not occlude the blood supply (Sj0gaard, 1987). However, the relatively sparse capillary
network does not deliver sufficient oxygen to the working muscle fibers even though cardiac
output and blood pressure are reflexly elevated to deliver more blood to the muscle per unit
time (Hudlicka, 1984). The oxygen deficit and poor oxidative capacity of the fast-twitch
muscle fibers cannot sustain ATP production, and the relatively little ATP generated via
anaerobic glycolysis is associated with glycogen depletion and rapid accumulation oflactate,
free ADP and AMP (Green et aI., 1992). The reliance on creatine phosphate as the immediate
source of ATP in fast-twitch muscles results in a rapid depletion (Westerblad et aI., 1991;
Nagesser et aI., 1992). In addition, the acidic pH inhibits PFK which, in tum, reduces the
generation of ATP via anaerobic glycolysis (Trivedi & Danforth, 1966; Sahlin, 1978). The
resulting metabolite accumulation of glucose-l ,6, phosphate and fructose-2,6,biphosphate
within the first day of stimulation has an inhibitory effect on glycolysis and later may
stimulate glycogenolysis for replenishment of glycogen.
As reviewed earlier, generation of inorganic phosphate and decline in pH following
lactate accumulation in the working muscle may account for the immediate decline in force
(Edman, Chapter 1; Stephenson et aI., Chapter 2; Allen et aI., Chapter 3). Phosphate and
hydrogen ions reduce the Ca2+ sensitivity ofthe contractile filaments and the force generated
by the cross-bridges. As fatigue progresses, a fall in ATP may reduce Ca2+ release via
ryanodine release channels (because ATP is required for opening of the channels, Smith et
aI., 1985), and increase levels of cytoplasmic Mg2+ (Lamb & Stephenson, 1991) and
extracellular K+. The latter arises partly as a result of increased opening of ATP-sensitive K+
channels at low [ATP]i levels (Castle & Haylett, 1987) and despite increased Na+/K+
436 T. Gordon

A. Before Stimulation

.,
U
1.0

0
LL
a.9

".,
0 .8
N
a.7
'0
E
a
0.6

Z 0.5
0 2 J 4 5 6 8 9 10
Time (min)

B. After Stimulation
rrrr(r(r(rr((rrrr(,(rrrr

J~~uJvw ~~~~ J~ ~ ~~~ ~v

.,
U
1.0

0
LL
0 .9

".,
N
0.8
0.7
a
E 0 .6
0
•Time (min)
Z 0.5
0 5 6 8 9 10

Figure S. Fatigue of the cat medial gastrocnemius muscle during repetitive trains oflow-frequency stimulation
and long-term consequences. A, 2.5 strains of20 Hz repeated every 2.5 s in a 50% duty cycle was associated
with a 50% fall in tetanic force. This fall was presumably due to the fatigue of the most fatigable FF units,
which normally contribute 60% of the total force of the maximally stimulated muscle (Gordon & Mao, 1994).
S , after subjecting the muscle to this stimulation pattern for 7 weeks, performance improved and the chronically
stimulated muscle experienced less than a 20% decline in tetanic force in the same 10 min period. The majority
of the motor units were classified as fatigue resistant, consistent with the presence of highly oxidative muscle
fibers as shown in Fig. 2.

pumping (Hicks & McComas, 1989). Progressive failure of the Ca2+ ATPase to sequester
Ca2+ leads to a progressive increase in free intracellular Ca 2+ which in tum may further inhibit
Ca 2+ release (Allen et aI., 1992). The high levels of free intracellular Ca2+ tend to continue
during the first weeks of stimulation possibly as a result of the reduced Ca 2+-ATPase activity
(Leberer et aI., 1987).
Both the reduced energy supply and the increase in intracellular Ca2+ have been
suggested to trigger change in gene expression (Saltin & Gollnick, 1983; Pette & Vrbova,
1992). It has been speculated that the [ATP]/[ADP][P,] ratio is sensed by the genome of the
nucleus and mitochondria to upregulate the expression of enzymes of the citric acid cycle
and oxidative phosphorylation. Within the first 15 min of continuous 24 hr daily stimulation,
the ratio declined almost to zero and did not recover in the 50 days of study (Pette & Vrbova,
1992). Thus, chronic ischemia, appears to be a potent stimulus to increase oxidative potential.
Fatigue in Adapted Systems 437

In addition, elevated substrate concentrations may be important for inducing gene expression
for oxidative enzymes (Saltin & Gollnick, 1983).
A five-fold increase in free [Ca2+] in the resting muscle within the first 2 weeks of
stimulation has been postulated to trigger conversion of sarcoplasmic reticular and myofi-
brillar proteins to the slow type (Sreter et aI., 1987). The mechanical loading of actively
contracting muscles may also influence gene expression (Goldspink et aI., 1992). The signal
transduction mechanisms that mediate the changes in gene expression, however, are not yet
understood.
Ischemia and increased K+, H+, and Ca2+ ionic concentrations could also be involved
in protein catabolism and, in tum, reduced muscle fiber volume. Oxidative stress in
stimulated muscles and associated damage to muscle fibers and connective tissue may lead
to invasion of blood borne cells and release of growth factors. These, in tum, are likely to
help increase the number of myonuclei and the growth of capillaries (angiogenesis;
Thompson et aI., 1988; Grounds, 1991; Russell et aI., 1992). The satellite cells, which
normally lie dormant under the muscle fiber basement membrane, are the most likely source
of the increased number of nuclei found in the stimulated muscle fibers (Russell et aI., 1992).
Proliferation of endothelial cells is essential for the marked growth of capillaries in the
stimulated muscles (Hudlicka, 1984).
The growth factors include TGF-131 and platelet derived growth factor (PDGF) from
platelets and basic fibroblast growth factor (bFGF) from macrophages. bFGF may also be
released from endothelial cells and from binding sites on muscle extracellular matrix proteins
in response to plasmin and heparinases released from lymphocytes and macrophages and
plasmin from the endothelium (Klagsbrun, 1989; Grounds, 1991). TGF-13 and PDGF, much
like helper cells of the immune system, recruit cells such as macrophages into the site of
injury and stimulate their release of bFGF, other angiogenic factors, and interleukin-l.
Importantly, macrophages grown under conditions of low oxygen tension have been shown
to secrete angiogenic factors (Knighton et aI., 1983; Gimbrone, 1984). Oxidative stress may
also trigger important changes in the active nerve terminals. Even low amounts of stimulation
of fast motor units in the crayfish increase the mitochondrial content and synaptic efficiency
of the terminals (Nguyen & Atwood, 1994).

Slow-to-Fast Adaptation
The reverse conversion from slow-to-fast muscle phenotype has been more difficult
to demonstrate. Salmons & Vrbova argued that low-frequency stimulation promoted slow
phenotype and high intermittent stimulation more typically seen in fast flexor muscles
promoted fast phenotype (Vrbova, 1963; Salmons & Vrbova, 1969; Salmons & Sreter, 1976).
In experiments in which reflexly elicited contractions of the soleus muscle were reduced by
tenotomy and central drive eliminated by spinal cord transection, Vrbova (1963) demon-
strated that slow contractions were maintained only if the muscles were stimulated with
10-20 Hz stimulation. If the muscles were stimulated intermittently with 40-100 Hz, the
muscle contractions were fast. Even though tenotomy is associated with considerable muscle
necrosis (McMinn & Vrbova, 1964), the results were convincing in their demonstration of
the effects oflow-frequency stimulation in promoting slow phenotype.
The effects of high-frequency intermittent stimulation are controversial. Intermittent
stimulation of denervated slow- and fast-twitch muscles with high-frequency tetanic bursts
increased the rate of contraction in the muscles (Gorza et aI., 1988; Gundersen et aI., 1988).
However, slow-twitch soleus muscles retained their high fatigue resistance and did not
express the type lIB fast myosin heavy chain isoform typical of the FF motor units
(Gundersen et aI., 1988; Ausoni et aI., 1990). Thus, although conversion did occur, conver-
sion was limited to fiber type I -> type IIA.
438 T. Gordon

In innervated muscles, the effects of intermittent stimulation were quite different.

Three laboratories observed, to their surprise, that high-frequency stimulation of innervated
fast muscles resulted in a fast-to-slow conversion rather similar to that observed after
low-frequency stimulation (Hudlicka et ai., 1982; Sreter et ai., 1982; Eerbeek et ai., 1984).
It appeared that the important factor controlling muscle speed and endurance was the amount
of daily activity (total number of impulses per day) rather than the pattern of activity as
originally suggested. A progressive increase in daily activity in paralysed flexor muscles in
the cat from as little as 0.5% to 50% was associated with a progressive conversion of muscle
contractile properties from FF -> FI -> FR -> S (Kernell et ai., 1987a,b; Kernell & Eerbeek,
1989). In contrast, maximal force appeared to depend on both pattern and amount of chronic
stimulation: the fall in muscle force that accompanies an increase in muscle endurance could
be reversed by interposing occasional tetanic bursts at high stimulation frequencies (Kernell
et ai., 1987b). These results suggested that occasional high force contractions are required
to maintain muscle strength and that this effect is independent of the effects ofiow-frequency
stimulation on muscle endurance. Alternatively, the decline in muscle force during low-fre-
quency stimulation can be reversed by preventing muscle shortening during contraction
(Fig. 6).
In summary, oxidative capacity and muscle endurance increase with the duration of
neuromuscular activity. Muscle strength and endurance are reciprocally regulated with high
neuromuscular activities favoring high endurance but low force in small muscle fibers that
have high mitochondrial content and high oxidative enzyme capacity. Muscle strength is
promoted under conditions of relatively little neuromuscular activity but depends on the
intermittent development of high forces or resisted contractions against a load.

EXERCISE TRAINING AND MUSCLE ENDURANCE, STRENGTH

AND SPEED

In their excellent review of biochemical adaptations to endurance exercise, Holloszy

and Booth (1976) stated that the nature of the exercise stimulus determines which of two
distinct adaptive responses are induced in skeletal muscle. Increase in muscle strength but
not endurance is exemplified by muscle adaptation in weight lifters and body builders.
Increase in muscle endurance but not strength is found in the muscles of competitive
long-distance runners and cross-country skiers, cyclists and swimmers. Because the training
in the former primarily involves isometric or loaded muscle contractions as opposed to
dynamic contractions in the latter, these are sometimes referred to as static vs. dynamic forms
of exercise (Longhurst & Stebbins, 1992). However, since many strengthening activities also
involve dynamic exercise, the terms must be used with caution.

Endurance or Dynamic Training

There are many parallels in the extent and time course of muscle adaptation to
imposed low-frequency electrical stimulation and the adaptive changes during endurance
exercise (reviewed by Henrikkson, 1992). Typical endurance training regimes are 2-3 months
with bouts of 30-60 min of exercise corresponding to 70-80% of maximal oxygen uptake,
3-5 times per week in humans or strenuous running programs in rats (up to 2 hrs per day).
These lead to a progressive increase in muscle endurance, which is associated with increased
capacity of the trained muscles to generate ATP from the oxidative metabolism of carbohy-
drates and long chain fatty acid substrates (Holloszy & Booth, 1976; Saltin & Gollnick,
1983). These adaptations occur in all types of muscle fibers (Terjung, 1976). Increased
Fatigue in Adapted Systems 439

A. Extensor
25
B. Flexor
50

40 20
........
Z
15
30
c
0
(f)
20 10
c
OJ
f- 5
10

0
0 free spring isom
free spring isom (short)
(short)

0.5 1.0

X 0.4 0.8
OJ
"0
c 0.3 0.6
OJ
::J
OJ 0.2 0.4
..-
0
LL 0.1 0.2

0.0 0.0

Figure 6. Effects of muscle loading on reduced tetanic tension and fatigability of paralysed cat (A) medial
gastrocnemius extensor muscles and (B) common peroneal innervated flexor muscles in response to low-fre-
quency (20 Hz) daily electrical stimulation in a 50% duty cycle for 3 hrs/day (5% total daily activity). Paralysis
was induced by unilateral deafferentation and hemisection at T12. Isometric tension was recorded at regular
intervals at optimal length by non-invasive coupling of the cat foot to an external-held force transducer, under
Halothane anesthesia (chronic recording methods described by Davis et aI., 1978). Fatigue index was
calculated as the ratio of the tetanic tension at the end and the beginning of a 2-min period of repetitive trains
of stimuli (13 pulses at 40 Hz every s). Since recordings were made in the paralysed muscles before and after
daily stimulation, each animal served as its own control. The open bars are the mean (± S.E.) of5-15 recorded
values before the onset of stimulation. The closed bars are the average of 10-15 values recorded more than 60
days after stimulation. Under conditions in which muscle contractions were unopposed, tetanic tension
declined by 20% (A) associated with an increase in endurance (a 3-fold increase above prestimulation values).
Spring loading of the contractions had little effect on muscle strength but was associated with increased fatigue
resistance, particularly in the extensor muscles. Preventing muscle shortening by fixing the ankle in the same
boot used to spring load the contractions largely eliminated the fall in muscle force without severely
compromising the endurance response of the muscles (Tyreman, Mao & Gordon, unpublished observations).

oxidative capacity arises as a result of increased oxygen uptake, myoglobin content,

hexokinase activity, mitochondria number, oxidative enzyme content, and increased capil-
larization with little change in glycolytic enzyme activities (Lamb et aI., 1969; Baldwin et
aI., 1973; Saltin & Gollnick, 1983). There is, in addition, an increased capacity for glycogen
and triglyceride synthesis in trained muscles (Morgan et aI., 1969; Taylor et aI., 1972). The
small changes in glycolytic enzyme activities in athletes (e.g., Baldwin et aI., 1973) were
attributed to the normally low levels in these individuals who tend to have high percentages
of type I slow fibers in contrast to the high type II fiber content in sprinters (Gollnick et aI.,
1972; Mero et aI., 1981; 1992). Another explanation is that exercise-induced decreases of
glycolytic enzyme activity in type IIA fibers in muscles containing all three muscle fiber
types were obscured by the increases in type I fibers and the small changes in glycolytic
enzyme levels in the type lIB fibers (Holloszy & Booth, 1976). In rats, the question of
whether the treadmill running and swimming programs were sufficiently severe to stress the
440 T. Gordon

A B
Percentage of fast-twitch fibers Percentage of fast-twitch fibers
o 3l 40 00 00 100 o 3l 40 00 00 100

~~~~i.i~iw;m~~'ii~~~:
$hor·Pun.,..
DI,cu, Throwefs
!W."ttfLIf,.r,

um,."." MEN WOMEN

-Rang_
lDO.",.,1I $'11',. ........... ::.":ct·.rd t--tStandlrd
[Link]
Oevia1ion

[Link],8OQm

C~,,"

100 00 00 40 3l 0
1:;. Hoclley I - - + - - t
Percentage of slow-twitch fibers
A'pj"'S/l,.r~

[Link]. RUM'"

100 00 00 40 3l 0
Percentage of slow-twitch fibers

Figure 7. Distribution of fast-twitch (FT) and slow-twitch (ST) fibers in (A) vastus lateralis muscles of
different groups of male athletes and (B) in medial gastrocnemius of female athletes. Although there is some
degree of variation, on the average, endurance athletes tend to have greater percentages of ST fibers, whereas
non-endurance athletes have greater percentages of FT fibers than their nonathletic counterparts (modified
from Fox et ai., 1993).

glycolytic potential (Saltin & Gollnick, 1983). The exception is the increase in hexokinase
which, as described above, is the only glycolytic enzyme to be significantly elevated with
increased activity (Barnard & Peter, 1969).
Increased oxidative capacity is associated with conversion of type lIB to IIA muscle
fibers in human muscles but relatively little conversion from type II to type I (Anderson &
Henriksson, 1977; Jansson & Kaijser, 1977; Schantz et ai., 1983). The trend for endurance
athletes to have a high percentage of type I muscle fibers in contrast to weight-lifters and
sprinters (Fig. 7) was attributed to genetic factors (Komi et ai., 1977; Edstrom & Grimby,
1986; Salmons & Henriksson, 1986; Bouchard et ai., 1992; Mero et ai., 1992), based very
much on the early findings of Gollnick and colleagues (1973) that there was no interconver-
sion of fiber types in trained muscles. More recently, however, the observation of increasing
numbers of fast muscle fibers during detraining suggests that there may be some intercon-
version (Henriksson, 1992), consistent with histochemical findings of more intermediate
fiber types (Schantz et ai., 1983). The absence of adaptation in fast-twitch white muscles in
rats after endurance training was explained by their lack of recruitment during exercise
regimes including marathon running (Holloszy & Booth, 1976; Costill et ai., 1977). More
extreme and prolonged training, however, led to significant fiber conversions in the same
direction as found with chronic electrical stimulation; namely type IIB-> IIA-> I (Green et
ai., 1984). Thus, failure to detect more type I fibers following endurance training may be
attributed to the failure of exercise programs to recruit the larger fast motor units for sufficient
time and intensity to result in extensive muscle fiber conversion.
Fatigue in Adapted Systems 441

The earliest adaptation to muscle exercise is the reflex increase in blood flow to the
working muscle that is mediated by increased cardiac output. During endurance or dynamic
exercise, muscle blood flow increases linearly with exercise intensity with little change in
local vascular resistance (Hudlicka, 1984; Sj0gaard, 1987). The limiting factor to oxygen
supply to the muscle is the transit time for the blood through the capillary network. As the
capillary density increases with training (reviewed by Henriksson, 1992), this time is
prolonged. As a result, oxygen, energy substrates, and metabolites are better exchanged
across the capillary walls. However, an upper limit in cardiac output ultimately is a major
factor that limits muscle performance particularly for whole muscle exercise (see Lindstedt
& Hopeller, Chapter 28). Consequently, cardiac and respiratory training are a major focus
in endurance training (e.g., Anholm et aI., 1989; Boutellier et aI., 1992). The vasodilatation
in the working muscles constitutes a volume load to the heart. This leads to a significant
ventricular hypertrophy in trained endurance athletes, which is an essential component of
their cardiovascular training (Longhurst et aI., 1980b). The volume load is considerably
greater than the afterload due to increased blood pressure during static or isometric exercise
(Longhurst & Stebbins, 1992).
The competitive athlete is, therefore, well adapted to use oxidative substrates to
generate ATP during exercise. A resting bradycardia and lower absolute heart rate and trained
capacity to increase cardiac output and ventilation ensures delivery of oxygen and substrates
to the working muscles (Longhurst et aI., 1980a; Brooks & Fahey, 1985). During the exercise,
carbohydrate and fat are metabolized for energy with exhaustion being experienced only
when glycogen is significantly depleted (Hagerman, 1992). During competitive marathon
running which requires amongst the highest total energy expenditure, muscle glycogen was
depleted to extremely low levels (Costill, 1986). Short, high-intensity exercise such as
sprinting produces smaller augmentation of the oxidative pathways (Saubert et aI., 1973).

Static or Isometric Exercise

Static or high resistance exercise training is the typical form of exercise for weight
lifters and wrestlers. The duration of exercise is short but, most importantly, the muscles
perform relatively little external work in comparison with the dynamic exercise oflong-dis-
tance runners, for example. The muscles consume energy and liberate it as heat (Fenn, 1923)
and there is little increase in oxygen consumption during exercise (Longhurst & Mitchell,
1983; Reid et aI., 1987). During the exercise, blood flow to the active muscle is progressively
impeded as intramuscular pressure exceeds arterial and capillary blood pressure. With forces
that exceed 30% maximal voluntary contraction, intramuscular blood vessels may be
completely occluded (Humphreys & Lind, 1963) and the muscles are highly adapted for
anaerobic glycolysis, containing high glycolytic enzyme activities (Saltin & Gollnick, 1983).
There is a tendency for these athletes to have a higher type II1type I fiber ratio but there is
relatively little change in endurance with strictly isometric exercise training (Salmons &
Henriksson, 1986; their Fig. 7). There is also little metabolic adaptation to this type of
exercise; there is a small but significant increase in resting muscle creatine, CP, ATP, and
glycogen (MacDougall et aI., 1977). However, a disproportionate increase in myofibrillar
proteins tends to dilute the metabolic changes (Gollnick et aI., 1972). In rats, high intensity
but brief exercise similarly results in little metabolic adaptation but significant fiber hyper-
trophy (Hoppeler, 1988).
The primary effect of training and the most important goal of resistance training is
to increase the force generated by the muscle. Some of the increase in muscle strength and
corresponding muscle bulk is undoubtedly due to increased muscle fiber cross-sectional area
(hypertrophy) and/or increased force per unit area (specific force; Gollnick et aI., 1972;
MacDougall et aI., 1977; 1982; Saltin & Gollnick 1983). Evidence for an increase in the
 442 T. Gordon

number of muscle fibers (hyperplasia; reviewed by Antonio & Gonyea, 1993) has been
criticized on methodological considerations of the difficulties in obtaining representative
and accurate muscle fiber counts (see Saltin & Gollnick, 1983; Edstrom & Grimby, 1986;
Antonio & Gonyea, 1993). Hypertrophy, however, cannot fully account for increased muscle
bulk or increased force. Firstly, a dramatic increase in the hydrophillic connective tissue,
which increases water content, accounts for a substantial component of the increased muscle
bulk (Jablecki et aI., 1973; Saltin & Gollnick, 1983). Secondly, comparison of evoked and
voluntary muscle contractions has shown that the increase in volitional muscle strength is
due, in part, to an enhanced capacity of the trained individual to recruit additional motor
units and increase the firing rates of recruited units (Gollnick et aI., 1974; Jones et aI., 1989;
Narici et aI., 1989; reviewed by Edstrom & Grimby, 1986; Enoka, 1988). Thus, neural as
well as muscle plasticity contribute to increased strength after weight training.

ADAPTATION TO EXPERIMENTAL MUSCLE OVERLOAD

The adaptive changes in the functionally overloaded muscles following experimental

removal of muscle synergists have been compared to those subjected to endurance training
(reviewed by Gardiner, 1991). In an early study, Schiaffino and Bormioli (1973) noted a
dramatic shift in type lIB to IIA fiber type shift in remaining ankle flexor muscles when the TA
muscle was extirpated in the neonatal rat. Clearly the overuse had induced an adaptive
metabolic response. Yet, in the adult animal the adaptation is usually small by comparison,
especially in physiological flexors (Olha et aI., 1988; Frischknecht & Vrbova, 1991; Rosenblatt
& Parry, 1993), unless muscle activity is increased by additional activation (Riedy et aI., 1985;
Roy et aI., 1992) or the functional overload is maximized by removal of all synergists but the
experimental muscle (Chalmers et aI., 1992). In adult extensor muscles, there is a small but
significant increase in the proportion ofS motor units and corresponding SO fibers, particularly
in the deep portion of the muscles (Walsh et aI., 1978; Tsika et aI., 1987; Olha et aI., 1988; Crerar
et aI., 1989). Interestingly, all motor units show compensatory hypertrophy but conversion of
fast-to-slow motor units is limited. The limited conversion was attributed to the minimal
recruitment of the larger FI and FF units in accordance with the good correlation of motor unit
endurance with total amount of neuromuscular activity (Kernell & Eerbeek, 1989). This is the
same argument used for the limited fast-to-slow adaptation after endurance training, as
discussed above. In contrast, when all but one muscle is left in a muscle compartment, the
adaptation is more marked. Nonetheless, a disproportionate increase in fiber volume as
compared with oxidative enzyme activity dilutes the influence of overload on the total enzyme
activity (Chalmers et aI., 1992), as described for static exercise below.
The more generalized muscle fiber hypertrophy, however, is consistent with the
chronic stimulation experiments which showed that it is the type rather than the amount of
muscular activity that is important in determining muscle fiber size and motor unit force.
Increasing mechanical loading of contracting muscle is a powerful stimulus for muscle fiber
enlargement (Riedy et aI., 1985). Even passive muscle responds to loading since compensa-
tory hypertrophy following synergist ablation even occurs in denervated muscle (Gutmann
et aI., 1971). With few exceptions, increased muscle mass and strength are due to hypertrophy
rather than hyperplasia (James, 1976; Gollnick et aI., 1981; but see Antonio & Gonyea, 1993)
and requires incorporation of satellite cells (Rosenblatt & Parry, 1993). Recruitment of
satellite cells as the source of new nuclei maintains the normal ratio of myonuclei to cell
volume (Schmalbruch & Hellhamer, 1977; Russell et aI., 1992). Studies of muscle regen-
eration after muscle trauma indicate that many of the growth factors implicated in the
revascularization of damaged muscles are also important for the proliferation, differentiation
and fusion of muscle cells from satellite cells (Grounds, 1991). Many of the same muscle
Fatigue in Adapted Systems 443

regulatory factors are expressed as during the commitment and maturation of developing
muscle fibers. For example, the expression of a muscle-specific regulatory factor qmfl is
increased several fold when avian muscles are stretched (Russell et aI., 1992). Expression
of skeletal muscle specific genes such as MyDI and myogenin have been used to identify
muscle precursor cells in regenerating adult muscle after direct muscle damage (Grounds,
1991). Using gamma irradiation to prevent satellite cell incorporation and compensatory
hypertrophy, recent experiments show that the functional overload can affect a fiber-type
conversion in the absence of compensatory hypertrophy (Rosenblatt & Parry, 1993). Thus,
the progressive conversion in the direction ofFF-> FI -> FR -> S with increased neuromus-
cular activity in the overload model provides additional evidence for the direct association
between neuromuscular activity, metabolic enzyme activities, and metabolic adaptation.

MUSCLE ADAPTATION TO SELF- AND CROSS-REINNERVATION

OF DENERVATED MUSCLES

The dramatic results of the cross-reinnervation experiments of Buller and colleagues

(1960) that the fast flexor digitorum longus muscles became slowly contracting and the slow
soleus muscles contracted more rapidly when their nerves were cross-united provided
irrefutable evidence for muscle plasticity in the adult. Findings that these changes could be
reversed if the original patterns of stimulation were superimposed provided strong evidence
that the neural effects were mediated by neuromuscular activity (Salmons & Sreter, 1976).
Changes in contractile speed are associated with corresponding changes in myosin light
and heavy chain isoforms, although cross-reinnervated motor units may continue to express
their original myosin isoform in some oftheir muscle fibers (Sreter et aI., 1976; Gauthier et aI.,
1983; Fu et aI., 1992). Also extensively documented are changes in sarcoplasmic reticular and
regulatory proteins that support the conversion of fiber types in cross-reinnervated muscles
(e.g., Margreth et aI., 1973; Amphlett et aI., 1975; Sreter et aI., 1976).
Increase in the number ofS motor units and SO fibers in cross-reinnervated fast-twitch
muscles is associated with increased oxidative and decreased glycolytic enzyme activity,
accounting for the increased resistance of the muscles to fatigue (e.g., Dum et aI., 1985;
Gillespie et aI., 1986; 1987; Gordon et aI., 1988). The glycolytic enzyme activity increases in
many fibers in cross-reinnervated soleus muscles in association with expression of fast
contractile and regulatory proteins. However, the muscle fibers retain their high oxidative
capacity with the result that all motor units classified as F units are FR units and the muscles
retain their high resistance to fatigue (Chan et aI., 1982; Gillespie et aI., 1986; 1987).
Even after self-reinnervation of skeletal muscles, there is extensive mismatch of
nerves and muscle fibers because reinnervating nerves do not selectively reinnervate their
original muscle fibers (Kugelberg et aI., 1970; Totosy de Zepetnek et aI., 1992a). Newly
reinnervated motor units, therefore, include muscle fibers of different types which formerly
belonged to different motor units. Interestingly, many of the self-reinnervated motor units
are FI (Gordon, 1987: Gordon et aI., 1988) and variance in fiber size within reinnervated
motor units is high (Totosy de Zepetnek et aI., 1992b). These findings suggest that reinner-
vated muscle fibers retain some of their original phenotype. Normally the variance in
oxidative and glycolytic enzyme activities between unit muscle fibers is very low (Nemeth
et aI., 1986; their Fig. 8). After self- or cross-reinnervation, however, the variance for
representative enzymes is increased several fold (Sesodia et aI., 1993; their Fig. 8). This
heterogeneity could explain the trend for motor units to be fatigue intermediate. It is unlikely
that the heterogeneity can be explained by incomplete activation of all motor unit fibers
444 T. Gordon

Adenylokinose Lactate dehydrogenase Malate dehydrogenase

2. o MU fibers
~ Non-MU fibers
E 20
<

l ,
o

1S r-
-
,-
.
'0

o
"
10

5 -----I- - ,....--- -'- '-- -

'--- '0., • endy Koy Den 4 Wendy - Kay

Figure 8. Coefficient of variation of the enzyme activity of oxidative and glycolytic enzymes in muscle fibers
in reinnervated tibialis anterior muscles of the rat after self-reinnervation by common peroneal nerve (Den 4)
or cross-reinnervation by the posterior tibial nerve (Wendy, Kay). At least 4 months after nerve repair, a single
motor unit was isolated for glycogen depletion (methods described by Totosy de Zepetnek et aI., 1992a) and
the enzyme activities of oxidative (malate dehydrogenase) and glycolytic enzymes (lactate dehydrogenase,
adenylokinase) analyzed using microdissection and microanalysis of 8-20 muscle fibers (methods described
by Nemeth et aI., 1986). The coefficient of variation expressed as a percentage of the normal variation in
age-matched control muscles was obviously higher in muscle fibers of the same type randomly sampled outside
of the depleted unit. Within motor units, the variability between fibers in enzyme activities were greatly
increased above normal and sometimes as high as between fibers of the same type outside of the motor unit
(unpublished data of Sesodia, Fu, Nemeth and Gordon).

because the heterogeneity was seen in long-term reinnervated motor units in which
neuromuscular transmission was stable.
One explanation for the apparent retention of former phenotype is that adult muscle
fibers can adapt within a limited range set by their developmental origins (reviewed by
Gordon & Pattullo, 1993). Another consideration is that the architecture of reinnervated
muscles is not altered, irrespective of the source of innervation. As a result, the length and
load on the muscles are not necessarily altered after reinnervation and may contribute to the
retention of original muscle phenotype. Experimental support for the second suggestion is
provided by the findings that almost complete slow-to-fast conversion occurred in cross-re-
innervated soleus muscles when the common peroneal nerve (CP) was the source of fast
nerves but conversion was not complete when the fast nerves were synergistic flexor hallucis
(FHL) or digitorum longus (FDL) muscle nerves (Dhoot & Vrbova, 1991). In the former
case (CP), complete denervation of antagonist muscles removed the normal stretch on the
soleus muscle. As a result, the cross-reinnervated soleus muscles experienced an unloaded
contraction which possibly accounted for the transition to fast phenotype. In the latter case
(FHL or FDL), denervation of one physiological extensor did not change the loading
conditions of the soleus substantially, and the loaded soleus muscle retained a slow pheno-
type, despite cross-reinnervation.
Thus, metabolic and phenotypic adaptations in reinnervated muscles reveal a com-
bined effect of neuromuscular activity and muscle loading which together determine the
conversion of muscle properties by the innervating nerves. The important contribution of
muscle loading to the expression of slow phenotype and metabolic adaptation is further
revealed in underuse paradigms as described below.

METABOLIC ADAPTATION TO MUSCLE DISUSE

The effects of underusage paradigms are quite variable in natural and laboratory
conditions, between muscles, species, procedures and different laboratories. Nevertheless,
Fatigue in Adapted Systems 445

there is a general trend for muscles to express fast phenotype with respect to contractile and
regulatory proteins and to upregulate enzymes of anaerobic glycolysis (reviewed by Gordon
& Pattullo, 1993). However, in no case has complete conversion taken place. Many muscle
properties are unchanged, in particular, susceptibility to fatigue in postural muscles which
normally cross a single joint. These muscles do not fatigue readily and maintain their high
oxidative capacity even though glycolytic enzyme activity increases. The adaptations have
been recently reviewed (Roy et aI., 1991; Gordon & Pattullo, 1993) and are, therefore, only
briefly described here.

Spinal Cord Transection

Fast-twitch but not slow-twitch postural muscles become more fatigable after hem i-
and complete spinal cord transections. There is a coordinated increase in glycolytic and
mATPase enzyme activity associated with a limited fiber type I to II conversion (Gordon &
Pattullo, 1993). In the cat fast-twitch MG muscle, for example, in which there are normally
only 25% FF units, there is a significant increase in susceptibility to fatigue (Fig. 9). This in
turn coincides with increased numbers of FF units, a corresponding increase in type IIB
fibers (Gordon & Mao, 1994; see also Mayer et aI., 1984; Munson et aI., 1986) and increased
glycolytic enzyme activity relative to oxidative enzyme activity (Hoffman et aI., 1990; Jiang
et ai., 1990b).
In slow-twitch and postural muscles, retention of the high oxidative potential despite
increased mATPase and glycolytic enzyme activity explains the high fatigue resistance of
the paralyzed muscles (Hoffman et ai., 1990; Jiang et ai., 1990a,b) and the limited conversion
of S to FR motor units and associated type I to IIA fiber type conversion (Cope et ai., 1986).
This adaptive pattern resembles that of chronically stimulated denervated muscles in the rat
where intermittent electrical stimulation caused a type I to IIA conversion of fiber types. The
same pattern is also seen in other disuse models, such as hindlimb suspension, spaceflight
and immobilization.

Reduced Load Bearing: Hindlimb Suspension and Space Flight

In hindlimbs that are chronically unloaded by tethering the tail to a swivel support
or by supporting the body weight with a sling, fast-twitch muscles become slightly more
fatigable after an initial delay (Winiarski et ai., 1987; Herbert et aI., 1988; Templeton et
aI., 1988; Pierotti et aI., 1990). The small increase was associated with a modest type I
to type II fiber conversion, an increase in the ratio of glycolytic to oxidative enzyme
activities (Roy et aI., 1987; Jiang et ai., 1992) and expression of the fast isoform of the
sarcoplasmic reticulum Ca 2+-pump (Schulte et aI., 1993). As above, soleus muscles
retained their high fatigue resistance, high oxidative capacity (Ohira et ai., 1992) and
showed a modest increase in contractile speed (Corley et ai., 1984; Templeton et aI.,
1984; Fitts et aI., 1986).
Although muscle endurance has not been measured under the non-weight bearing
conditions of space-flight, the direction of fiber-type conversion is the same (Roy et
aI., 1991). However, the metabolic changes are considerably less after short duration
space flight (14 days) than hindlimb suspension (Jiang et aI., 1992) particularly in
the monkey (Bodine-Fowler et aI., 1992). In the latter experiments, the little or no
impact of spaceflight on muscle metabolism or muscle fiber size was attributed to the
work load of the muscles in the constrained sitting position of the animal (Bodine-Fowler
et aI., 1992).
446 T. Gordon

A. Co ntro l

B. Paralyzed

10NL 30 sec

Figure 9. Fatigue of cat medial gastrocnemius muscle before and 93 days after paralysis by hemisection at
TI2 and unilateral deafferentation from L2 to S2. The MG nerve was stimulated maximally via an implanted
cuff electrode. 13 pulses delivered at a rate of 40 Hz were repeated every second for 2 min and the ankle
extensor tension was recorded by non-invasive coupling the foot to an external force transducer, under
Halothane anesthesia (methods described in detail by Davis et aI., 1978).

Limb Immobilization
There is partial type I to type II conversion in slow- and fast-twitch hindlimb muscles
after limb immobilization. However, there is little or no change in susceptibility to fatigue
despite a decline in the number of mitochondria and oxidative enzyme activity (Rifenberick
et al., 1973; Krieger et al., 1980; Mayer et al., 1981: Witzmann et al., 1983). Many studies
have noted a small increase in contraction speed in slow-twitch as well as fast-twitch muscles
(reviewed by Gordon & Pattullo, 1993). The most dramatic and most well studied effects of
limb immobilization relate to the atrophy of the immobilized muscles, particularly in a
shortened position.

Pharmacological Paralysis

Application of tetrodotoxin (TTX) via cuff electrodes on sciatic or tibial nerves

provides an effective block of evoked neuromuscular activity and has been explored over
periods of2-6 weeks as a model of hindlimb disuse. Increase in contractile speed associated
with type I to II fiber conversion and expression of fast myosin heavy chains was not
accompanied by any significant change in muscle fatigability (Spector, 1985a,b; St-Pierre
& Gardiner, 1985, Gardiner et al., 1992). Decline in oxidative potential was noted but the
change was small (Gardiner et al., 1992).
Fatigue in Adapted Systems 447

How Much Can Adaptive Changes Be Attributed to Loss of

Neuromuscular Activity in Disuse Models
The direction of motor unit and fiber type conversion (from S to FF and type I to lIB,
respectively) is consistent with the scheme of Kernell and coworkers (1987a,b; Kernell &
Eerbeek, 1989) in which progressively less daily activity shifts properties progressively from
S -> FR-> FI -> FF. Summed EMG activity after complete spinal transection in cats and
kittens (Alaimo et aI., 1984; Lovely et aI., 1990) and humans (Stein et aI., 1992) may be
reduced to less than 10% of normal but depends on the extent of spasticity. Perhaps the
remaining neuromuscular activity is sufficient to maintain slow phenotype in some fibers
and limit progressive conversion to FF units. This does not appear to be so since many type
I fibers remain in soleus, gastrocnemius and TA muscles that have been completely silenced
by spinal isolation (Graham et aI., 1992; Pierotti et aI., 1991). Furthermore, the adaptive
response of different muscles to the same loss of neuromuscular activity varies considerably
(e.g., Roy & Acosta, 1986) and the adaptive response of anyone muscle to different degrees
of paralysis may be surprisingly similar (Jiang et aI., 1990a,b; Jiang et aI., 1991; Graham et
aI., 1992).
In most disuse models, with the exception of complete pharmacological blockade,
remaining neuromuscular activity may be considerable. After hindlimb suspension for
example, neuromuscular activity is reduced only temporarily, returning to normal within a
week (Alford et aI., 1989). Daily integrated EMG activity in immobilized soleus muscles
may be as low as 5-15% but as high as 50-80% of normal (Fournier et aI., 1983). In
immobilized fast-twitch extensor and flexor muscles, neuromuscular activity approaches
normal levels (Mayer et aI., 1981; Hnik et aI., 1985).
Similar to the adaptive responses in overuse paradigms and reinnervated muscles,
muscle loading is likely to make important contributions to the adaptive responses of muscles
in underuse paradigms. Muscle paralysis is usually associated with muscle unloading. This
is particularly striking in postural muscles that are normally weight bearing and which show
the most pronounced fiber type conversion and atrophic changes after paralysis (e.g., Lieber
et aI., 1988; reviewed by Gordon & Pattullo, 1993). Spastic contractions in spinal-cord-in-
jured patients may exacerbate the atrophy because unloaded muscles, particularly single joint
postural muscles are known to be damaged when muscle shortening is unopposed (Vrbova,
1963; McMinn & Vrbova, 1967; Tabary et aI., 1972; Baker & Matsumoto, 1988). Loss of
weight bearing is the major consequence of hindlimb suspension, which is the basis for this
experimental model being used to simulate space-flight conditions with unresisted muscle
contractions.
Expression of fast or slow phenotype appears to strongly influenced by muscle
length, whether the muscle is active or not. Immobilization of soleus muscle in a shortened
position, for example, initiates expression of fast myosin heavy chains (Goldspink et aI.,
1992), consistent with fiber type I to II conversion seen in immobilized muscles. Immobi-
lization of the fast-twitch TA in a lengthened position repressed the normal expression of
fast phenotype (Goldspink et aI., 1992) suggesting that muscle stretch is an important
determinant of slow phenotype and that unloading is an important contributing factor to type
I to II conversion of immobilized muscles. Stretch and chronic stimulation enhances the
expression of slow myosin heavy chains in these muscles indicating that loading of chroni-
cally stimulated muscles is likely to be an important contributing factor in their fast-to-slow
conversion.
Changes in blood flow and bone density associated with muscle paralysis are also
factors that may be involved in the adaptive changes of muscle and may also be linked
to the loss of weight-bearing (see Gordon & Mao, 1994). Finally, generalized effects of
448 T. Gordon

stress must be considered in the adaptive responses of disused muscles (reviewed by

Gordon & Pattullo, 1993). In animal models, considerable stress may be provoked by
the experimental manipulations of spinal cord surgery, pinning of joints and the animal
handling. In human patients, the immobilization and muscle paralysis after spinal cord
and peripheral nerve, bone, tendon or ligament injuries must be considered in the context
of the pain and stress associated with the injuries. Responses to pain and stress, which
include raised levels of glucocorticoids, adrenal hypertrophy and gastric ulcers, have
been recorded in animal models in association with muscle atrophy (Thomason & Booth,
1990). Glucocorticoids released in response to stress and binding to receptors induced
in immobilized muscles may contribute to the increase in contractile speed (Dubois &
Almon, 1980; Gardiner et aI., 1980).
In summary, changes in muscle phenotype and disuse atrophy have often been
attributed to reduced muscle use but review of the many studies using disuse models indicate
that changes in muscle length and/or weight-bearing are likely to be important contributing
factors. Fiber type I to type II conversions with remaining heterogeneity of muscle fibers
and motor units occur in all these models which share in common the unloading ofthe muscle
contractions even when they differ considerably in their total amount of daily activity.
Hormonal factors and changes in blood flow associated with the traumatic manipUlations
may also contribute to the phenotypic expression of muscles, including their susceptibility
to fatigue.

CONCLUSIONS

Adaptive responses of skeletal muscles to overuse, reinnervation and underuse

paradigms may be striking with conversion of muscle speed, endurance and strength.
Although dramatic conversion of muscle properties can be demonstrated under conditions
of chronic electrical stimulation, many studies identify the limits of the muscle adaptability
and the role of cardiovascular and central plasticity. Careful analysis of the adaptive
responses of muscles to novel innervation and to underuse paradigms within the context
of alterations in muscle activity, length and loading reveal the importance of the load on
the passive and active muscles in expression of fast and slow muscle phenotype. Finally,
systemic responses to altered neuromuscular activity must be recognized as significant
contributing factors in the adaptive responses of skeletal muscles in over- and underuse
paradigms. Overall, there is strong evidence to support the ideas that I) the amount of
neuromuscular activity strongly determines muscle endurance and the underlying meta-
bolic adaptations; and 2) the loading against which a muscle contracts determines muscle
strength.

ACKNOWLEDGMENTS

The support of the Medical Research Council and Muscular Dystrophy Association
of Canada is gratefully acknowledged. The author would also like to thank many of her
colleagues who participated in the studies referred to in the chapter, particularly Drs. Jian
Mao and Susan Fu whose unpublished data is discussed, and Neil Tyreman for his tireless
work in the laboratory, and assistance in preparation of the manuscript. The author's
attendance at the 1994 Bigland-Ritchie conference was supported, in part, by the University
of Miami.
Fatigue in Adapted Systems 449

REFERENCES
Alaimo MA, Smith AIL, Roy RR & Edgerton VR (1984). EMG activity of slow and fast ankle extensors
following spinal cord transection. Journal ofApplied Physiology 56, 1608-1613.
Alford EK, Roy RR, Hodgson JA & Edgerton VR (1989). Electromyography of rat soleus, medial gastroc-
nemius, and tibialis anterior during hind-limb suspension. Experimental Neurology 96,635-649.
Allen DG, Westerblad H, Lee JA & Lannergren J (1992). Role of excitation-contraction coupling in muscle
fatigue. Sports Medicine 13, 116-126.
Amphlett GW, Perry SV, Syska H, Brown M & Vrbova G (1975). Cross innervation and the regulatory protein
system of rabbit soleus muscle. Nature 257,602-604.
Andersen P & Henriksson J (1977). Training induced changes in the subgroups of human type II skeletal muscle
fibers. Acta Physiology Scandinavica 99, 123-125.
Anholm JD, Stray-Gundersen J, Ramanathan M & Johnson RL Jr (1989). Sustained maximal ventilation after
endurance exercise in athletes. Journal ofApplied Physiology 67, 1759-1763.
Antonio J & Gonyea WJ (1993). Skeletal muscle fiber hyperplasia. Medicine and Science in Sports and
Exercise 25, 1333-1345.
Ariano MA, Armstrong RB & Edgerton VR (1973). Hindlimb muscle fiber populations of five mammals.
Journal of Histochemistry and Cytochemistry 21,51-55, 1973.
Ausoni S, Gorza L, Schiaffino S, Gundersen K & Lomo T (1990). expression of myosin heavy chain isoforms
in stimulated fast and slow rat muscles. Journal ofNeuroscience 10, 153-160.
Baker JH & Matsumoto DE (1988). Adaptation of skeletal muscle to immobilization in a shortened position.
Muscle & Nerve 11,231-244.
Baldwin KM, Winder WW, Terjung RL & Holloszy J (1973). Glycolytic enzymes in different types of skeletal
muscle: adaptation to exercise. American Journal of Physiology 225, 962-966.
Barnard RJ & Peter JB (1969). Effect of training and exhaustion on hexokinase activity of skeletal muscle.
Journal ofApplied Physiology 27,691-695.
Bass A, Brdiczka D, Eyer P, Hofer S & Pette D (1969). Metabolic differentiation of distinct muscle types at
the level of enzymatic organization. European Journal of Biochemistry 10, 198-206.
Bodine-Fowler SC, Roy RR Rudolph W, Haque N, Kozlovskaya IB & Edgerton VR (1992). Spaceflight and
growth effects on muscle fibers in the rhesus monkey. Journal ofApplied Physiology 73, 82S-89S.
Bouchard C, Dionne FT, Simoneau J-A & Boulay MR (1992). Genetics of aerobic and anaerobic performances.
Exercise and Sport Sciences Review 20, 27-58.
Boutellier U, Buchel R, Kundert A & Spengler C (1992). The respiratory system as an exercise limiting factor
in normal trained subjects. European Journal of Applied Physiology and Occupational Physiology
65,347-353.
Brooks GA & Fahey TA (1985). Exercise Physiology: Human Bioenergetics and its Applications. New York:
Macmillian.
Brooks MH & Kaiser KK (1970). Three "myosin adenosine trophosphatase" systems: the nature oftheir pH
lability and sulfhydryl dependence. Journal of Histochemistry and Cytochemistry 18, 670-672.
Brown MD, Cotter MA, Hudlicka 0 & Vrbova G (1976). The effect of different patterns of muscle activity
on capillary density, mechanical properties and structure of slow and fast rabbit muscles. Pflugers
Archiv 361,241-250.
Buller AJ, Eccles JC & Eccles RM (1960). Interactions between motoneurones and muscles in respect ofthe
characteristic speeds of their responses. Journal ofPhysiology (London) 150, 399-416.
Burke RE (1981). Motor units: anatomy, physiology, and functional organization. In: Brookhart JM,
Mountcastle VB (sec. eds.), Brooks VB (vol. ed.), Handbook of Physiology, sec. 1, vol. II, pt 1, The
Nervous System: Motor Control, pp. 345-422. Bethesda, MD: American Physiological Society.
Burke RE & Edgerton VR (1975). Motor unit properties and selective involvement in movement. Exercise
and Sport Sciences Reviews 3, 31-81.
Burke RE, Levine DN, Tsairis P & Zajac FE (1973). Physiological types and histochemical profiles in motor
units of the cat gastrocnemius. Journal ofPhysiology (London) 234, 723-748.
Castle NA & Haylett DG (1987). Effect of channel blockers on potassium efflux from metabolically exhausted
frog skeletal muscle. Journal ofPhysiology (London) 383, 31-43.
Chalmers GR, Roy RR & Edgerton VR (1992). Variation and limitations in fiber enzymatic and size responses
in hypertrophied muscle. Journal ofApplied PhYSiology 73, 631-641.
Chan M, Edgerton VR, Goslow GE Jr, Kurata H, Rasmussen S & Spector SA (1982). Histochemical and
physiological properties of cat motor units after self- and cross-reinnervation. Journal of Physiology
(London) 332, 343-361.
450 T. Gordon

Cope TC, Bodine SC, Fournier M & Edgerton VR (1986). Soleus motor units in chronic spinal transected cats:
physiological and morphological alterations. Journal ofNeurophysiology 55, 1202-1220.
Corley K, Kowalchuk B & McComas AJ (1984). Contrasting effects of suspension on the hindlimb muscles
in the hamster. Experimental Neurology 85, 30-40.
Costill DL (1986). Inside Running: Basics of Sports Physiology. Indianapolis, IN: Bechmark Press, Incorpo-
rated.
Costill DL, Coyle EF, Dalsky G, Evans W, Fink W & Hoopes D (1977). Effects of elevated plasma FFA and
insulin on muscle glycogen usage during exercise. Journal ofApplied Physiology 43,695-699.
Crerar MM, Hamilton NC, Blank S, Urdea MS & lanuzzo CD (1989). The genes for j3-myosin heavy chain
and glycogen phosphorylase are discoordinately regulated during compensatory growth of the
plantaris muscle in the adult rat. Molecular Cellular Biochemistry 86, 115-123.
Davis LA, Gordon T, Hoffer JA, Ihamandas J & Stein RB (1978). Compound action potentials recorded from
mammalian peripheral nerves following ligation and resuturing. Journal ofPhysiology (London) 285,
543-559.
Denny-Brown, D (1929). On the nature of postural reflexes. Proceedings of Royal Society (Biology) 104,
252-301.
Dhoot GK & Vrbova G (1991). Conversion of cat and rabbit soleus muscle fibres by alien nerves. Muscle &
Motility 2, 125-130.
Donselaar Y, Eerbeek 0, Kernell D & Verhey BA (1987). Fibre sizes and histochemical staining characteristics
in normal and chronically stimulated fast muscle of cat. Journal ofPhysiology (London) 382, 237-254.
Dubois DC & Almon RR (1980). Disuse atrophy of skeletal muscle is associated with an increase in number
of glucocorticoid receptors. Endocrinology 107: 1649-1651.
Dum RP, O'Donovan MJ, Toop J, Tsairis P, Pinter MJ & Burke RE (1985). Cross-reinnervated motor units in
cat muscle. II. Soleus muscle reinnervated by flexor digitorum longus motoneurons. Journal of
Neurophysiology 54,837-851.
Edstrom L & Grimby L (1986). Effect of exercise on the motor unit. Muscle & Nerve 9, 104-126.
Eerbeek 0, Kernell D & Verhey BA (1984). Effects of fast and slow patterns of tonic long-term stimulation
on contractile properties of fast muscles in cat. Journal ofPhysiology (London) 352, 73-90.
Eisenberg BR & Salmons S (1981). The reorganization of subcellular structure in muscle undergoing
fast-to-slow type transformation. A streological study. Cell Tissue Research 220, 449-471.
Enoka RM (1988). Muscle strength and its development: new perspectives. Sports Medicine 6, 148-168.
Fenn WO (1923). A quantitative comparison between the energy liberated and the work performed by the
isolated sartorius muscle of the frog. Journal of Physiology (London) 58, 175-203.
Fitts RH, Metzger DA, Riley DA & Unsworth BR (1986). Models of disuse: a comparison of hindlimb
suspension and immobilization. Journal ofApplied Physiology 60, 1946-1953.
Fournier M, Roy RR, Perham H, Simard CP & Edgerton VR (1983). Is immobilisation a model of disuse?
Experimental Neurology 80, 147-156.
Fox EL, Bowers RW & Foss ML (1988). The Physiological Basis of Physical Education and Athletics. New
York: WB Saunders.
Fox EL, Bowers RW & Foss ML (1993). The Physiological Basis for Exercise and Sport, 5th Ed. Dubuque,
IA: Wm. C. Brown Communications, Inc.
Frischknecht R & Vrbova G (1991). Adaptation of rat extensor digitorum longus to overload and increased
activity. Pjliigers Archiv 419,319-326.
Fu S, Gordon T, Parry DJ & Tyreman N (1992). Immunohistochemical analysis of fiber types within
physiologically typed motor units of rat tibialis anterior muscle after long-term cross-reinnervation.
Society for Neuroscience Abstracts 18, 1557.
Gardiner PG (1991). Effects of exercise training on components of the motor unit. Canadian Journal ofSports
Sciences 16,271-288.
Gardiner PG, Favron M & Corriveau P (1992). Histochemical and contractile responses of rat medial
gastrocnemius to two weeks of complete disuse. Canadian Journal ofPhysiology and Pharmacology
70, 1075-81.
Gardiner PG, Montanaro DS & Edgerton VR (1980). Effects ofglucocortocoid treatment and food restriction
on rat hindlimb muscles. Pjliigers Archiv 385, 147-153.
Gauthier GF, Burke RE, Lowey S & Hobbs AW (1983). Myosin isozymes in normal and cross-reinnervated
cat skeletal muscle fibers. Journal of Cell Biology 97, 756-771.
Gillespie MJ, Gordon T & Murphy PA(1986). Reinnervation of the lateral gastrocnemius and soleus muscles
in the rat by their common nerve. Journal ofPhYSiology (London) 372, 485-500.
Fatigue in Adapted Systems 451

Gillespie MJ, Gordon T & Murphy PA (1987). Motor units and histochemistry in the rat lateral gastrocnemius
and soleus muscles: evidence for dissociation of physiological and histochemical properties after
reinnervation. Journal ofNeurophysiology 57,921-937.
Gimbrone MA (1984). Macrophages, neovascularisation and the growth of vascular cells. In: Jaffe EA (ed),
Biology of Endothelial Cel/s, pp. 97-107. Boston: Martinus NihoffPublications.
Goldspink G, Scutt A, Loughna PT, Wells DJ, Jaenicke T & Gerlach GF (1992). Gene expression in skeletal
muscle in response to stretch and force generation. American Journal ofPhysiology 262, R356-R363.
Gollnick PD, Armstrong RB, Saltin B, Saubert CW, Sembrowich WL & Shepherd RE (1973). Effect of training
on enzyme activity and fiber composition of human skeletal muscle. Journal ofApplied Physiology
34,107-111.
Gollnick PD, Armstrong RB, Saubert IV CW, Piehl K & Saltin B (1972). Enzyme activity and fiber composition
in skeletal muscle of untrained and trained men. Journal ofApplied Physiology 33,312-319.
Gollnick PD, Piehl K & Saltin B (1974). Selective glycogen depletion pattern in human skeletal muscle fibres
after exercise of varying intensity and at varying pedaling rates. Journal ofPhysiology (London) 216,
1502-1509.
Gollnick PD, Timson BF, Moore RL & Riedy M (1981). Muscular enlargement and the number of fibers in
the skeletal muscles of rats. Journal of Applied Physiology: Respiratory Environment Exercise
Physiology 50, 936-943.
Gordon T (1987). Muscle plasticity as demonstrated during sprouting and reinnervation. American Zoology
27, 1055-1066.
Gordon T & Mao J (1994). Muscle atrophy and procedures for training for spinal cord injury. Physical Therapy
74,50-60.
Gordon T & Pattullo MC (1993). Plasticity of muscle fiber and motor unit types. Exercise and Sports Sciences
Reviews 21, 331-362.
Gordon T, Thomas CK, Stein RB & Erdebil S (1988). Comparison of physiological and histochemical
properties of motor units after cross-reinnervation of antagonistic muscles in the cat hindlimb. Journal
ofNeurophysiology 60,365-378.
Gorza L, Gundersen K, Lomo T, Schiaffino S & Westgaard, RH (1988). Siow-to-fast transformation of
denervated soleus muscles by chronic high-frequency stimulation in the rat. Journal of PhYSiology
(London) 402, 627-749.
Green HJ, Dusterhoft S, Dux L & Pette D (1992). Metabolite patterns related to exhaustion, recovery, and
transformation of chronically stimulated rabbit fast-twitch muscle. Pflugers Archiv 420, 359-366.
Green HJ, Klug GA, Reichman H, Seedorf D, Wiehrer W & Pette D (1984). Exercise-induced fibre type
transitions with regard to myosin, parvalbumin, and sarcoplasmic reticulum in muscles of the rat.
Pflugers Archiv 400, 432-438.
Graham SC, Roy RR, Navarro B, Jaing B, Pierotti D, Bodine-Fowler S & Edgerton VR (1992). Enzyme and
size profiles n chronically inactive cat soleus muscle fibers. Muscle & Nerve 15, 27-36.
Grounds MD (1991). Towards understanding skeletal muscle regeneration. Pathology Research Practice 187,
1-22.
Gundersen K, Leberer E, Lomo T, Pette D & Staron RS (1988). Fibre types, calcium-sequestering proteins
and metabolic enzymes in denervated and chronically stimulated muscles of the rat. Journal of
Physiology (London) 398, 177-189.
Gutmann E, Schiaffino S & Hanzlikova V (1971). Mechanism of compensatory hypertrophy in skeletal muscle
of the rat. Experimental Neurology 31, 451-464.
Hagerman FC (1992). Energy metabolism and fuel utilization. Medical Sciences Sports Exercise 24, S309-
S314.
Hartner K -T, Kirschbaum BJ & Pette D (1989). The multiplicity of troponin T isoforms. Norm"al rabbit muscles
and effects of chronic stimulation. European Journal ofBiochemistry 179, 31-38.
Henneman E & Mendell LM (1981). Functional organization of the motoneurone pool and its inputs. In:
Brookhart JM, Mountcastle VB (sec. eds.), Brooks VB (vol. ed.), Handbook of Physiology, sec. 1.
vol. II, pt. 1, The Nervous System: Motor Control, pp. 423-508, Bethesda: Ameriocan Physiological
Society.
Henriksson J (1992). Effects of physical training on the metabolism of skeletal muscle. Diabetes Care 15,
1701-1711.
Henriksson J, Chi MM-Y, Hintz CS, Young DA, Salmons S & Lowry OH (1986). Chronic stimulation of
mammalian muscle: changes in enzymes of six metabolic pathways. American Journal ofPhysiology
251, C614-C632.
Herbert ME, Roy RR & Edgerton VR (1988). Influence of one-week hindlimb suspension and intermittent
high load exercise on rat muscles. Experimental Neurology 102, 190-198.
452 T. Gordon

Hicks A & McComas AJ (1989). Increased sodium pump activity following repetitive stimulation of rat soleus
muscles. Journal of Physiology (London) 414, 337-349.
Hnik P, Vejsada R, Goldspink DF, Kasicki S & Krekule I (1985). Quantitative evaluation of electromyogram
activity in rat extensor and flexor muscles immobilized at different lengths. Experimental Neurology
88,515-528.
Hoffmann SJ, Roy RR, Bianco CE & Edgerton VR (1990). Enzyme profiles of single muscle fibers in the
absence of normal neuromuscular activity. Journal ofApplied Physiology 69, 1150-1158.
Holloszy JO & Booth FW (1976). Biochemical adaptations to endurance exercise in muscle. Annual Reviews
ofPhysiology 38,273-291.
Hood DA, Zak R & Pette D (1989). Chronic stimulation of rat skeletal muscle induces coordinate increases
in mitochondrial and nuclear mRNAs of cytochrome c oxidase subunits. European Journal of
Biochemistry 179, 275-280.
Hoppeler H (1988). Exercise-induced structural changes of skeletal muscle. lSI Atlas Science Biochemistry 1,
247-255.
°
Hudlicka (1984). Development of microcirculation: capillary growth and adaptation. In: Renkin EM, Michel
CC, Geiger SR (eds), Handbook of Physiology, sec. 2, The Cardiovascular System, pp 165-216.
Baltimore, MD: Williams and Wilkins.
Hudlicka 0, Brown M, Cotter M, Smith M & Vrbova G (1977). The effect oflong-term stimulation of fast
muscles on their blood flow, metabolism and ability to withstand fatigue. Pfliigers Archiv 369,
141-149.
Hudlicka 0, Tyler KR, Srihari T, Heilig A & Pette D (1982). The effect of different patterns of long-term
stimulation on contractile properties and myosin light chains in rabbit fast muscles. Pfliigers Archiv
393, 164-170.
Humphreys PW & Lind AR (1963). The blood flow through active and inactive muscles ofthe forearm during
sustained hand-grip contractions. Journal of Physiology (London) 166, 12-131.
Jablecki CK, Heuser JE & Kaufman S (1973). Auto-radiographic localisation of new RNA synthesis in
hypertrophying skeletal muscle. Journal of Cellular Biology 57,743-759.
James NT (1976). Compensatory hypertrophy in the extensor digitorum longus muscle of the mouse. Journal
ofAnatomy 122, 121-131.
Jansson E & Kaijser L (1977). Muscle adaptation to extreme endurance training in man. Acta PhYSiology
Scandinavica 100, 315-325.
Jiang B, Ohira Y, Roy RR, Nguyen Q, Ilyina-Kakueva EI, Oganov V & Edgerton VR (1992). Adaptation of
fibers in fast-twitch muscles of rats to spaceflight and hindlimb suspension. Journal of Applied
Physiology Supplement 73, 58S-65S.
Jiang B, Roy RR & Edgerton VR (1990a). Expression ofa fast fiber enzyme profile in the cat soleus after
spinalization. Muscle & Nerve 13, 1037-1049.
Jiang B, Roy RR & Edgerton VR (1990b). Enzymatic plasticity of medial gastrocnemius fibers in the adult
chronic spinal cat. American Journal of Physiology 259, C507-C514.
Jiang B, Roy RR, Navarro C, Nguyen Q, Pierotti D & Edgerton VR (1991). Enzymatic responses of cat medial
gastrocnemius fibers to chronic inactivity. Journal ofApplied PhYSiology 70, 231-239.
Jones DA, Rutherford OM & Parker DF (1989). Physiological changes in skeletal muscle as a result of strength

°
training. Quarterly Journal Experimental Physiology 74, 233-256.
Kernell D, Donselaar Y & Eerbeek (1987b). Effects of physiological amounts of high- and low-rate chronic
stimulation on fast-twitch muscle of the cat hindlimb. II. Endurance-related properties. Journal of

°
Neurophysiology 58, 614-627.
Kernell D & Eerbeek (1989). Physiological effects of different patterns of chronic stimulation on muscle
properties. In: Rose FC, Jones R (eds), Neuromuscular Stimulation, pp 193-200. New York: Demos.
Kernell D, Eerbeek 0, Verhey BA & Donselaar Y (1987a). Effects of physiological amounts of high- and
low-rate chronic stimulation on fast-twitch muscle of the cat hindlimb. I. Speed and force-related
properties. Journal of Neurophysiology 58, 598-613.
Klagsbrun M (1989). The fibroblast growth factor family: structural and biological properties. Progress in
Growth Factor Research 1, 207-235.
Komi PV, Viitasalo JHT, Havu M, Thorstensson A, Sjodin B & Karsson J (1977). Skeletal muscle fibres and
muscle enzyme activities in monozygous and dizygous twins of both sexes. Acta PhYSiology
Scandinavica 100, 385-392.
Knighton DR, Hunt TK, Scheunesnstuhl H, Halliday BJ, Werb Z & Banda MJ (1983). Oxygen tension regulates
the expression of angiogenic factor by macrophages. Science 221, 1283-1285.
Krieger DA, Tate CA, McMillin-Wood J & Booth FW (1980). Populations of rat skeletal muscle mitochondria
after exercise and immobilization. Journal ofApplied Physiology 48, 23-28.
Fatigue in Adapted Systems 453

Kumer SW & Vaughn-Williams EM (1953). Properties of the "slow" skeletal muscle fibres ofthe frog. Journal
of Physiology (London) 121,318-340.
Kugelberg E, Edstrom L & Abruzzese M (1970). Mapping of motor units in experimentally reinnervated rat
muscle. Journal of Neurology, Neurosurgery and Psychiatry 33, 310-329.
Kugelberg E & Lindegren B (1979). Transmission and contraction fatigue of rat motor units in relation to
succinate dehydrogenase activity of motor unit fibres. Journal ofPhysiology (London) 288, 285-300.
Kuhne W (1863). Uber die Endignung der Nerven in den Muskeln. Vischow Archives 27, 508-533.
Lamb DR, Peter JB, Jeffress RN & Wallace HA (1969). Glycogen, hexokinase, and glycogen synthetase
adaptations to exercise. American Journal of Physiology 217, 1628-1632.
Lamb GD & Stephenson DG (1991). Effect ofMg2+ on the control ofCa 2+ release in skeletal muscle fibres of
the toad. Journal of Physiology (London) 434, 507-528.
Leberer E, Hartner K-T, Brandl CJ, Fujii J, Tada M, MacLennan DH & Pette D (1989). Slow\cardiac
sarcoplasmic reticulum Ca-ATPase and phospholamban mRNAs are expressed in chronically stimu-
lated rabbit fast-twitch muscle. European Journal of Biochemistry 185, 51-54.
Leberer E, Hartner K-T & Pette D (1987). Reversible inhibition of sarcoplasmic reticulum Ca-ATPase by
altered neuromuscular activity in rabbit fast-twitch muscle. European Journal of Biochemistry 62,
255-561.
Leberer E, Seedorf U & Pette D (1986). Neural control of gene expression in skeletal muscle Ca-sequestering
proteins in developing and chronically stimulated rabbit skeletal muscles. Biochemical Journal 239,
295-300.
Lieber RL, Friden JO, Hargens, AR, Danzig LA & Gershuni DH (1988). Differential response of the dog
quadriceps muscle to external skeletal fixation of the knee. Muscle & Nerve It 193-201.
Longhurst JC, Kelly AR, Gonyea WJ & Mitchell JH (1980a). Echocardiographic left ventricular mass in
distance runners and weight lifters. Journal of Applied Physiology 48, 154-162.
Longhurst JC, Kelly AR, Gonyea WJ & Mitchell JH (1980b). Cardiovascular responses to static exercise in
distance runners and weight lifters. Journal ofApplied Physiology 49,676-683.
Longhurst JC & Mitchell JH (1983). Does endurance training benefit the cardiovascular system? Journal of
Cardiovascular Medicine 8, 227-236.
Longhurst JC & Stebbins CL (1992). The isometric athlete. Cardiology Clinics 10, 281-294.
Lovely RG, Gregor RJ, Roy RR & Edgerton VR (1990). Weight-bearing hindlimb stepping in treadmill
exercised adult spinal cats. Brain Research 514, 206-218.
MacDougall JD, Sale DG, Elder DG, Elder GB & Sutton JR (1982). Muscle ultrastructural characteristics of
elite power lifters and bodybuilders. European Journal ofApplied Physiology 48, 117-126.
MacDougall JD, Ward GR, Sale DG & Sutton JR (1977). Biochemical adaptation of human skeletal muscle
to heavy resistance training and immobilization. Journal of Applied Physiology: Respiratory Envi-
ronment Exercise Physiology 43, 700-703.
Margreth A, Salviati G & Carraro U (1973). Neural control of the activity of the calcium transport system in
sarcoplasmic reticulum of rat skeletal muscle. Nature 241, 285-286.
Mayer RF, Burke RE, Toop J, Hodgson JA, Kanda K & Walmsley B (1981). The effect of long-term
immobilization on the motor unit population of the cat medial gastrocnemius muscle. Neuroscience
6,725-739.
Mayer RF, Burke RE, Toop J, Walmsley & Hodgson JA (1984). The effect of spinal cord transection on motor
units in cat medial gastrocnemius muscles. Muscle & Nerve 7, 23-31.
McMinn RMH & Vrbova G (1964). The effect of tenotomy on the structure of fast and slow muscle in the
rabbit. Quarterly Journal of Experimental Physiology 49,424-429.
McMinn RMH & Vrbova G (1967). Motoneurone activity as a cause of degeneration in the soleus muscle of
the rabbit. Quarterly Journal of Experimental Physiology 52, 411-415, 1967.
Mero A, Komi PV & Gregor RJ (1992). Biomechanics of sprint running. A review. Sports Medicine 13,
376-392.
Mero A, Luhtanen P, Viitasalo JT & Komi PV (1981). Relationships between the maximal running velocity,
muscle fiber characteristics, force production and force relaxation of sprinters. Scandinavian Journal
of Sports Sciences 3, 16-22.
Morgan TE, Short FA & Cobb LA (1969). Effect oflong-term exercise on skeletal muscle lipid composition.
American Journal of Physiology 216, 82-86.
Munson JB, Foerhing RC, Loften SA, Zengel JE & Sypert GW (1986) Plasticity of medial gastrocnemius
motor units following cordotomy in the cat. Journal of Neurophysiology 55, 619-634.
Nagesser AS, van der Laarse WJ & Elzinga G (1992). Metabolic changes with fatigue in different types of
single muscle fibres of Xenopus Laevis. Journal of Physiology (London) 448, 511-523.
454 T. Gordon

Narici MV, Roi GS, Landoni L, Minetti AE & Cerretelli P (1989). Changes in force, cross-sectional area and
neural activation during strength training and detraining of the human quadriceps. European Journal
ofApplied Physiology 59,310-319.
Nemeth PM, Solanki L, Gordon DA, Hamm TM, Reinking RM & Stuart DG (1986). Uniformity of metabolic
enzymes within individual motor units. Journal of Neuroscience 6, 892-898.
Nguyen PV & Atwood HL (1994). Altered impulse activity modifies synaptic physiology and mitochondrial
rhodamine-123 florescence in crayfish phasic motor neurons. Journal of Neurophysiology 72,
2944-2955.
Ohira Y, Jiang B, Roy RR, Oganov V, I1yina-Kakueva E, Marini JF & Edgerton VR (1992). Rat soleus muscle
fiber responses to 14 days of spaceflight and hindlimb suspension. Journal of Applied Physiology
Supplement 73, 51 S-57S.
Ohlendieck K, Briggs KN, Lee KF, Wechsler AW & Campbell KP (1991). Analysis of excitation-contraction-
coupling components in chronically stimulated canine skeletal muscle. European Journal ofBiochem-
istry 02, 739-747.
Olha AE, Jasmin BJ, Michel RN & Gardiner PF (1988). Physiological responses of rat plantaris motor units
to overload induced by surgical removal of its synergists. Journal ofNeurophysiology 60, 2138-2151.
Peter JB, Barnard RJ, Edgerton VR, Gillespie CA & Stempel KE (1972). Metabolic profiles of three fibre
types of skeletal muscle in guinea pigs and rabbits. Biochemistry 11, 2627-2633.
Pette D & Hofer HW (1980). The constant proportion enzyme group concept in the selection of reference
enzymes in metabolism. Trends in Enzyme Histochemistry and Cytochemistry: Excerpta Medica
(Ciba Foundation Symposium) 73, 231-244.
Pette D & Luh W (1962). Constant-proportion groups of multilocated enzymes. Biochemistry Biophysics
Research Communications 8, 283-287.
Pette D, Muller W, Leisner E & Vrbova G (1976). Time dependent effects on contractile properties, fibre
population, myosin light chains and enzymes of energy metabolism in intermittently and continuously
stimulated fast twitch muscle of the rabbit. Pjlugers Archiv 364, 103-112.
Pette D, Smith ME, Staudte HW & Vrbova G (1973). Effects of long-term electrical stimulation on some
contractile and metabolic characteristics of fast rabbit muscles. P/lugers Archiv 338,257-272.
Pette D & Staron RS (1990). Cellular and molecular diversities of mammalian skeletal muscle fibres. Reviews
of Physiology. Biochemistry and Pharmacology 116, 1-76.
Pette D & Vrbova G (1992). Adaptation of mammalian skeletal muscle to chronic electrical stimulation.
Reviews of Physiology. Biochemistry and Pharmacology 120, 115-202.
Pierotti DJ. Roy RR, Bodine-Fowler SC, Hodgson JA & Edgerton VR (1991). Mechanical and morphological
properties of chronically inactive cat tibialis anterior motor units. Journal of Physiology (London)
444, 175-192.
Pierotti DJ, Roy RR, Flores V & Edgerton VR (1990). Influence of 7 days of hindlimb suspension and
intermittent weight support on rat muscle mechanical properties. Aviation Space Environmental
Medicine 61, 205-210.
Ranvier L (1874). De quelques faits relatifs a I'histologie et a la physiologie des muscles stries. Archives
Physiology and Normal Pathology 6, 1-15.
Reid CM, Yeater RA & Ullrich H (1987). Weight training and strength, cardiorespiratory functioning and body
composition of men. British Journal of Sports Medicine 21, 40-44.
Riedy MR, Moore RL & Gollnick PD (1985). Adaptive response of hypertrophied skeletal muscle to endurance
training. Journal of Applied PhYSiology 59, 127-131.
Rifenberick DH, Gamble J & Max SR (1973). Response of mitochondrial enzymes to decreased muscular
activity. American Journal of Physiology 225, 1295-1299.
Rosenblatt JD & Parry DJ (1993). Adaptation of rat extensor digitorum longus muscle to gamma irradiation
and overload. Pjlugers Archiv 423, 255-264.
Roy RR & Acosta L Jr (1986). Fiber type and fiber size changes in selected thigh muscles six months after
low thoracic spinal cord transection in adult cats: Exercise effects. Experimental Neurology 92,
675-685.
Roy RR, Baldwin KM & Edgerton VR (1991). The plasticity of skeletal muscle: effects of neuromuscular
activity. Exercise Sports Sciences Reviews 19, 269-312.
Roy RR, Bello MA, Bouissou P & Edgerton VR (1987). Size and metabolic properties of fibers in rat fast-twitch
muscles after hindlimb suspension. Journal of Applied Physiology 62, 2348-2357.
Roy RR, Hodgson JA, Chalmers GR, Buxton W & Edgerton VR (1992). Responsiveness of the cat plantaris
to functional overload. In: Sato Y, Portmans J, Hashimoto I, Oshida Y (eds), Integration of Medical
and Sports Sciences, Medical Sport Sciences 37, pp. 43-51. Basel: Karger.
Fatigue in Adapted Systems 455

Roy RK, Mabuchi K, Sarkar S, Mis C & Sreter FA (1979). Changes in tropomyosin subunit pattern in chronic
electrically stimulated rabbit fast muscles. Biochemistry Biophysics Research Communications 89,
181-187.
Russell B, Dix DJ, Haller DL & Jacobs-EI J (1992). Repair of injured skeletal muscle: A molecular approach.
Medicine and Science in Sports and Exercise 24, 189-196.
Sahlin K (1978). Intracellular pH and energy metabolism in skeletal muscle in man. Acta Physiology
Scandinavica Supplement 455 1-56.
Salmons S & Henriksson J (1986). The adaptive response of skeletal muscle to increased use. Muscle & Nerve
4,94-105.
Salmons S & Sreter FA (1976). Significance of impulse activity in the transformation of skeletal muscle type.
Nature 263, 30-34.
Salmons S & Vrbova G (1969). The influence of activity on some contractile characteristics of mammalian
fast and slow muscles. Journal of Physiology (London) 201, 535-549.
Saltin B & Gollnick PD (1983). Skeletal muscle adaptability: significance for metabolism and performance.
In: Geiger SR, Adrian RH (eds.), Peachey LD (sec. ed.), Handbook of Physiology, sec. 10, Skeletal
Muscle. Specialization, Adaptation, and Disease, pp. 555-631. Bethesda, MD: American Physiologi-
cal Society.
Saubert CW, Armstrong IV RB, Shepherd RE & Gollnick PD (1973). Anaerobic enzyme adaptations to sprint
training in rats. Pfliigers Archiv 341, 305-312.
Schantz PG, Henriksson J & Jansson E (1983). Adaptation of human skeletal muscle to endurance training of
long duration. Clinical Physiology 3, 141-151.
Schiaffino S & Bormioli SP «(1973). Adaptive changes in developing rat skeletal muscle in response to
functional overload. Experimental Neurology 40, 126-137.
Schmalbruch H & Hellhamer U (1977). The nuclei in adult rat muscles with special reference to satellite cells.
Anatomical Records 189, 169-176.
Schulte LM, Navarro J & Kandarian SC (1993). Regulation of sarcoplasmic reticulum calcium pump gene
expression by hindlimb unweighting. American Journal ofPhysiology 264 (Cellular Physiology 33),
C 1308-C 1315.
Sesodia S, Gordon T & Nemeth P (1993). Increased metabolic enzyme heterogeneity of muscle unit fibers
after reinnervation. Society for Neuroscience Abstracts 19, 152.
Sjegaard G (1987). Muscle fatigue. Medical Sport Sciences 26, 98-109.
Smith JS, Coronado R & Meissner G (1985). Sarcoplasmic reticulum contains adenine nucleotide-activated
calcium channels. Nature 316,736-738.
Spector SA (1985a). Effects of elimination of activity on contractile and histochemical properties of rat soleus
muscle. Journal of Neuroscience 5, 2177-2188.
Spector SA (1985b). Trophic effects on the contractile and histochemical properties of rat soleus muscle.
Journal of Neuroscience 5, 2189-2196.
Sreter FA, Gergely J, Salmons S & Romanul F (1973). Synthesis by fast muscle of myosin light chains
characteristic of slow muscle in response to long term stimulation. Nature New Biology 241, 17-19.
Sreter FA, Lopez JR, Alamo L, Mabuchi K & Gergely J (1987). Changes in intracellular ionized Ca
concentration associated with muscle fiber type transformation. American Journal ofPhysiology 253,
C296-300.
Sreter FA, Luff AR & Gergely J (1976). Effect of cross-reinnervation on physiological parameters and on
properties of myosin and sarcoplasmic reticulum of fast and slow muscles of the rabbit. Journal of
General Physiology 66,811-821.
Sreter FA, Pinter K, Jolesz F & Mabuchi K (1982). Fast to slow transformation offast muscles in response to
long-term phasic stimulation. Experimental Neurology 75, 95-102.
Stein RB, Gordon T, Jefferson J, Sharfenberger A, Yang JE, Totosy de Zepetnek & Belanger M (1992).
Optimum stimulation of paralysed muscle in spinal cord patients. Journal ofApplied Physiology 72,
1393-1400.
St-Pierre D & Gardiner P (1985). Effect of "disuse" on mammalian fast-twitch muscle: joint fixation compared
to neurally applied tetrodotoxin. Experimental Neurology 90, 635-651.
Sweeney HL, Kushmerick MJ, Mabuchi K, Sreter FA & Gergely J (1988). Myosin light chain and heavy chain
variations correlate with altered shortening velocity of isolated skeletal muscle fibers. Journal of
Biological Chemistry 263, 9034-9039.
Tabary JC, Tabary C, Tardieu C, Tardieu G & Goldspink G (1972). Physiological and structural changes in the
eat's soleus muscle due to immobilization at different lengths by plaster casts. Journal ofPhysiology
(London) 224, 231-244.
456 T. Gordon

Taylor AW, Thayer R & Rao S (1972). Human skeletal muscle glycogen synthase activities with exercise and
training. Canadian Journal of Physiology and Pharmacology 52, 119-122.
Templeton G, Padalino M, Manton J, Leconey T, Hagler H & Glasberg M (1984). The influence of rat
suspension-hypokinesia on the gastrocnemius muscle. Aviation Space Environmental Medicine 55,
381-386.
Templeton GH, Sweeney HL, Timson BF, Padalino M & Dudenhoeffer GA (1988). Changes in fiber
composition of soleus muscle during hindlimb suspension. Journal of Applied Physiology 65,
1191-1195.
Terjung RL (1976). Muscle fiber involvement during training of different intensities and durations. American
Journal of Physiology 230, 946-950.
Thomason DB & Booth FW (1990). Atrophy of the soleus muscle by hindlimb unweighting. Journal ofApplied
Physiology 68, 1-12.
Thompson JA, Anderson KD, Di Pietro JM, Swiebel JA, Zametta M, Anderson WF & Maciag T (1988.)
Site-directed neovessel formation in vivo. Science 241, 1349-1352.
Totosy de Zepetnek J, lung HV, Erdebil S & Gordon T (1 992a). Innervation ratio is an important determinant
of force in normal and reinnervated rat tibialis anterior muscle. Journal of Neurophysiology 67,
1385-1403.
Totosy de lepetnek JE, lung HV, Erdebil S & Gordon T (1992b). Motor unit categorization on the basis of
contractile and histochemical properties: a glycogen depletion analysis of normal and reinnervated
rat tibialis anterior muscle. Journal ofNeurophysiology 67,1404-1415.
Trivedi B & Danforth WH (1966). effect of pH on the kinetics of frog muscle phosphofructokinase. Journal
ofBiological Chemistry 241, 4110-4112.
Tsika RW, Herrick RE & Baldwin KM (1987). Time course of adaptation in rat skeletal muscle isomyosin
during compensatory growth and regression. Journal ofApplied Physiology 63, 2111-2121.
Vrbova G (1963). The effect of motoneurone activity on the speed of contraction of striated muscle. Journal
ofPhysiology (London) 169, 513-526.
Vrbova G, Gordon T & Jones R (1994). Nerve-Muscle Interaction. London: Chapman & Hall.
Walsh Iv, Burke RE, Rymer Wl & Tsairis P (1978). Effect of compensatory hypertrophy studied in individual
motor units in medial gastrocnemius muscle of the cat. Journal of Neurophysiology 41, 496-508.
Westerblad H, Lee JA, Lannergren J & Allen DG (1991). Cellular mechanisms offatigue in skeletal muscle.
American Journal of Physiology 261, C 195-C209.
Williams RS, Garcia-Moll M, Mellor J, Salmons S & Harlan W (1987). Adaptation of skeletal muscle to
increased contractile activity. Expression of nuclear genes encoding mitochondrial proteins. Journal
ofBiological Chemistry 262,2764-2767.
Williams RS, Salmons S, Newsholme EA, Kaufman RE & Mellor J (1986). Regulation of nuclear and
mitochondrial gene expression by contractile activity in skeletal muscle. Journal of Biological
Chemistry 261, 376-380.
Winiarski AM, Roy RR, Alford EK, Chiang PC & Edgerton VR (1987). Mechanical properties of rat skeletal
muscle after hind limb suspension. Experimental Neurology 96, 650-660.
Witzmann FA, Kim DH & Fitts RH (1983). Effect of hindlimb immobilization on the fatigability of skeletal
muscle. Journal ofApplied Physiology 54, 1242-1248.
 33

ASSOCIATIONS BETWEEN MUSCLE

SORENESS, DAMAGE, AND FATIGUE

P. M. Clarkson and D. J. Newham

1 Department of Exercise Science

University of Massachusetts
Amherst, Massachusetts 01003
2 Physiotherapy Group, Biomedical Sciences Division
King's College
London, England SE5 9RS

ABSTRACT

Eccentric exercise results in muscle soreness, structural damage, prolonged losses in strength
and range of motion, and neuromuscular dysfunction. Greater and longer lasting fatigue
occurs after eccentric compared with concentric and isometric exercise. Higher forces are
achieved during eccentric contractions with less ATP usage and greater increases in tempera-
ture. Although mechanisms involved in the damage and repair process are not well under-
stood, active strain during eccentric contractions is suggested to cause the initial damage
which increases over 2-3 days, followed by regeneration.

INTRODUCTION

Both recreational and serious athletes have experienced delayed onset muscle pain,
soreness, and stiffness following unaccustomed exercise or increased training workload.
Although research on exercise-induced muscle soreness dates back to 1902 (Hough), the
exact mechanism of soreness and pain, as well as the accompanying responses remain
unclear. These responses include a prolonged loss in range of motion and strength, increases
in muscle enzymes in the blood, swelling, and structural damage.
Several studies have shown that eccentric muscle contractions result in greater
soreness, fatigue, and damage than isometric or concentric contractions. This explains why
some activities such as hiking that incorporate a large degree of eccentric contractions result
in considerable soreness while others, like cycling, that are not biased toward eccentric
contractions produce little soreness. Exercise models developed to study muscle damage and
soreness are those where eccentric contractions predominate such as downhill running and
high-force exercise like lowering weights. The latter exercise results in the greater muscle
damage, fatigue, and soreness.

457
458 P. M. Clarkson and D. J. Newham

CHARACTERISTICS OF ECCENTRIC MUSCLE ACTIVITY

During active muscle lengthening, the mechanical and energetic behavior is very
different than during isometric or shortening contractions (see review by Woledge et aI., 1985).
Active muscles produce more force during stretch than when acting isometrically at the same
length, but the net energy produced and the chemical changes are smaller. This suggests that
the crossbridge itself is affected by stretch, which changes the way the muscle produces force.

Force-Velocity Relationship
The maximal forces generated during eccentric activity always exceed isometric
forces. As velocity increases, the force initially increases rapidly and then reaches a plateau
of nearly twice isometric force which is relatively independent of velocity. According to
Huxley's sliding filament theory (1957), this occurs because the extensible parts of the
myosin molecules forming the cross-bridges are stretched further than during isometric
activity. As the velocity of stretch increases, the number of attached cross-bridges at a given
muscle length is less, but those attached generate more force unless ripped apart by the extent
of the stretch. However, the energetic consequences of lengthening are not correctly
predicted by this theory in which the rate of crossbridge turnover, and thus ATP splitting,
should increase but this does not happen. While the stretch ofthe myosin molecules accounts
for some of the additional force generation during stretch, there is still an unexpectedly large
force because, for an unknown reason, there are more crossbridges attached. The mechanism
of ATP use during shortening contractions suggests that ATP is split to recharge the myosin
with energy after it has produced work during shortening. During stretch, however, the
myosin breaks from the actin in the backward position without doing work and does not
need recharging with energy. Thus, cyclic interactions between actin and myosin can occur
without ATP splitting (Woledge et aI., 1985).
Whether this process fully accounts for the lower ATP splitting during stretch is not
known. Furthermore, the energetics are different in slow and rapid stretches, where ATP
turnover decreases as the velocity of the stretch increases. Katz (1939) was the first to make
detailed studies of the force-velocity relationship during stretch and described a three phase
response using isotonic stretches of tetanized muscles. These were I) an instantaneous
lengthening of elasticity due to the series elastic component; 2) a relatively rapid stretch
followed by 3) slow lengthening of the fibers. During isovelocity stretches, the increased
force also varies with stretch velocity. At lower lengthening velocities it continues to
increase, becomes relatively constant at intermediate velocities, but peaks then declines at
higher velocities. While forces greater than isometric are generated during stretch, and are
affected by velocity, the force may also depend on mechanisms other than instantaneous
lengthening. The nature of these mechanisms and their behavior is uncertain. One factor may
be a disparity in sarcomere length along a stretched fiber, with simultaneous lengthening and
shortening in different segments.
The force-velocity relationship in intact humans differs from that of isolated muscle,
with relatively lower forces being generated during eccentric contractions (see Stauber,
1989). Biomechanical factors, such as joint angles, changes in the angle of pennation, and
the length of individual fibers, need to be taken into account in human studies. Also,
antagonist muscle activity appears to be greater in eccentric activity and this will affect the
force-velocity relationship (Gulch, 1994). There is evidence of activation failure (reduced
EMG activity) in some instances which is hypothesized to protect the musculoskeletal system
from additional damage (e.g., Westing et aI., 1991). However, other workers have recorded
similar EMG activity during maximal concentric and eccentric activity (e.g., Komi, 1973).
Associations between Muscle Soreness, Damage, and Fatigue 459

A logical deduction is that the high force generated by stretch in each active muscle
fiber causes mechanical trauma resulting in pain and damage. Warren and colleagues
(1993a), in isolated rat soleus muscle, showed that damage was initiated by mechanical
factors, with muscle tension being most important. However, it seems that damage is also a
function of active length changes (Lieber & Friden, 1993) and that, in humans, eccentric
activity performed at long muscles lengths is the most damaging (Jones et aI., 1989; Newham
et aI., 1988). These data indicate damage to the elastic cytoskeletal structures in series to the
contractile proteins (Jones et aI., 1989).

Energy Cost
During shortening contractions, work is done by the muscle, but during eccentric
contractions, work is done on the muscle by the external lengthening forces. This is
equivalent to the negative work done by the muscle. ATP splitting is reduced during eccentric
contractions. It is less than an isometric tetanus of the same duration, and is strongly velocity
dependent, being least at low velocities. ATP splitting increases at high stretch velocities,
but internal energy, that is the sum of heat production and work done on the muscle by
external forces (equivalent to the negative work done by it), does not. This supports the
hypothesis that energy is being stored in another form, possibly an endothermic process.
Bigland, Abbott and Ritchie (1952) studied oxygen consumption (assumed to be determined
by the number of active motor units and their firing frequency) during submaximal concen-
tric and eccentric activity of the same force. They demonstrated that oxygen intake was
substantially lower during eccentric contractions.
Elegant studies on the effects of force and speed on oxygen intake during eccentric
activity were performed by Abbott and Bigland (1953). When work rate was increased by
requiring greater force and lengthening velocity remained constant, the oxygen intake
increased rapidly. However, when work rate was increased by requiring greater speed,
oxygen intake remained constant. These results were explained by the fact that greater forces
can be generated at slower velocities, and at slower velocities more fibers are used. Abbott
and colleagues (1952), and others (Knuttgen et aI., 1971) have shown that oxygen intake
during maximal eccentric activity is much less than during concentric. Furthermore, during
eccentric contractions EMG was about half, and oxygen intake was one sixth, of that during
equivalent concentric (submaximal cycling) activity (Bigland-Ritchie & Woods, 1973).
Therefore, not only were fewer motor units recruited, but there was a 60% reduction in
oxygen and thus ATP consumed by each active fiber.

Muscle Temperature
Heat production is greater during lengthening, except at very low velocities, than
isometric contractions. However, the change in total internal energy is less during eccentric
contractions. The extra heat seems to be less than expected by the work done on the muscle.
The most likely explanation for this is that an extra endothermic process occurs during
stretch, such as the storage of energy in a stressed component of the muscle.
Intramuscular temperatures of up to 42°C have been recorded in human muscles
(Nadel et aI., 1972; Sargeant & Dolan, 1987). Heat liberation during eccentric contractions
is that produced by the metabolic processes, plus induced heat supplied by external energy.
It appears that muscle blood flow is proportional to metabolic rate and therefore inadequate
to dissipate the additional heat produced during eccentric exercise. The effects of these high
temperatures are not known, but have been speculated to: 1) increase the metabolic cost by
increasing the rate of crossbridge cycling; and 2) change mechanical properties to make the
muscle more susceptible to damage. Neither of these possibilities seems likely because
460 P. M. Clarkson and D. J. Newham

studies have shown that a progressive increase in oxygen intake during constant eccentric
exercise only occurs in individuals unfamiliar with that type of exercise (e.g., Bonde-Pe-
tersen et aI., 1973). Furthermore, similarly high muscle temperatures were found in trained
individuals (Knuttgen et aI, 1982) in whom pain and damage are substantially reduced or
eliminated (Newham et aI., 1987).
From these data, it is possible to deduce that the damage caused by eccentric activity
is not likely to be initiated by chemical processes in the active muscle. The high mechanical
tensions and strain, and perhaps temperature, are undoubtedly involved. However, their exact
role and the mechanisms to explain changes in contractile and non-contractile tissue are
unclear. In volitional exercise, the possibility for variability in all these factors is immense.
For example, motor unit recruitment, the length and changes in length probably alter from
one second to the next, and may be different in individual muscles and even within single
fibers (Faulkner et aI., 1993).

MANIFESTATIONS OF MUSCLE DAMAGE

Histological and Ultrastructural Changes

Muscle tissue can be structurally damaged by intense or long duration exercise,
particularly those where eccentric contractions predominate (see Friden & Lieber, 1992;
Faulkner et aI., 1993). Damage is present in the sarcolemma, T-tubules, myofibrils, and
cytoskeletal system.
After eccentric exercise, the damage ranges from mild, which is only detectable with
electron microscopy (EM), to more severe, which also can be seen with light microscopy
(LM). In the case of mild damage, immediately after exercise only single sarcomeres
throughout the muscle are affected and show evidence of Z-line streaming or bulging with
A-band and myofibrillar disarray. Over the next two to three days, the disruption becomes
more profound and extensive, with normal muscle architecture being lost from areas adjacent
to affected sarcomeres and fibers (Newham et aI., 1983b). This damage is repaired over
approximately ten days in humans.
The more severe form of damage has been reported repeatedly in human studies after
maximal eccentric exercise and is the type most often produced in animal studies. In human
studies the changes are minimal for the first couple of days, but then cytoskeletal and
myofibrillar damage, edema, invading mononuclear cells, muscle fiber atrophy and necrosis,
and finally regeneration are seen (Jones et aI., 1986). Full recovery and regeneration of fiber
size take about three weeks.
The progressive nature of structural damage after eccentric exercise is poorly understood.
The mechanism is thought to initially involve a mechanical insult followed by an accumulation of
intracellular calcium which triggers calcium mediated processes. Type II fibers are most affected
in humans and animals (Jones et al., 1986; Friden & Lieber, 1992; Faulkner et aI., 1993). Friden
and Lieber proposed that fast twitch fibers fatigue early in the exercise period and due to an
inability to regenerate ATP enter a state ofhigh stiffuess. Subsequent stretch mechanically disrupts
the fibers, causing the cytoskeletal and myofibrillar damage. Calcium chelators have been
reported to reduce structural changes (Duan et aI., 1990), but the buffering capacity may be
inadequate to prevent increases in cytosolic calcium (Lowe et aI., 1994).

Serum Levels of Muscle Proteins

When skeletal muscle is damaged, several intramuscular proteins leak out into the
blood. Although the most commonly measured is creatine kinase (CK), serum levels of
Associations between Muscle Soreness, Damage, and Fatigue 461

myoglobin, lactate dehydrogenase, aspartate aminotransferase, alanine aminotransferase and

myosin heavy chain fragments are also elevated following eccentric exercise (Nosaka et aI.,
1992; Mair et aI., 1992). Fatiguing concentric and isometric exercise result in relatively little
or no change in serum proteins, unless the exercise is prolonged.
After eccentric exercise there is a delay in the appearance of these proteins in the
blood (Clarkson et aI., 1992). Elevated levels may not be evident until 24-72 hours and reach
peak values 1 to 7 days later. Mair and colleagues (1992) reported that the time course of
increases in myosin heavy chain fragments and CK paralleled that of increases in MRI signal
intensity. It was suggested that the release of muscle specific proteins and their slow removal
from the extracellular area via the lymphatic system provide the colloid osmotic pressure to
produce edema.
The magnitude of increase in serum CK after eccentric exercise can be dramatic. For
example, peak values generally range from 1,000 to 10,000 lUlL after high force eccentric
exercise of the elbow flexors (Clarkson et aI., 1992). However, it is not uncommon to see
levels of25,000 lUlL, and in two cases levels of 100,000 lUlL were found. Such high levels
occasionally are found after exercises with a larger muscle mass, like those involved in
downhill walking and bench stepping (Newham et aI., 1983b; Jones et aI., 1986). There are
some subjects who show very little or no CK response. The explanation for this subject
variability is not known.
Although several studies have attempted to quantify the amount of muscle damage
based on serum CK levels, these studies failed to account for clearance rates. One recent
study assessed clearance at rest by injecting a CK solution into horses (Volfinger et aI., 1994).
However, it was assumed that clearance rates would be unaltered by exercise or by high
levels of CK in the blood. Nosaka and Clarkson (1993) found that eccentric exercise
performed when serum CK was elevated by prior eccentric exercise of another muscle group
resulted in almost no further increase in CK. It was suggested that elevated CK levels from
the first exercise accelerated clearance. Further studies of muscle protein release and
clearance are necessary to fully understand the serum protein response to eccentric exercise.

Soreness
Exercise-induced muscle soreness is different from other muscle pains such a cramps,
trauma, or ischemic pain (Miles & Clarkson, 1994). Usually a pain response is immediate,
while after eccentric exercise, pain and soreness do not appear for several hours and peak
24-48 hours post-exercise. Acute pain from a cramp or trauma is commonly described as a
sharp intense pain, while after eccentric exercise it is described as dull and aching. Muscle
cramps and pain due to trauma are characterized by ongoing pain, however the sensation of
muscle soreness is more commonly experienced when the muscle is active or being palpated.
Although the exact cause of the soreness is not known, inflammation and swelling
are considered prime factors (Smith, 1991). Initial studies found that the time course of
swelling and the development of soreness were not similar. Increases in limb circumference
peak about 5 days after exercise, approximately 3 days after peak soreness (Clarkson et aI.,
1992). However, circumference measures cannot differentiate between changes in different
compartments. More recent studies using MRI techniques show that fluid accumulates in
the muscle for 5 days after exercise and then moves into the subcutaneous area (Nosaka &
Clarkson, unpublished observation). Swelling within the muscle compartment could produce
pain and increase intramuscular pressures in low compliance compartments, as well as
sensitize pain receptors to other noxious stimuli.
Damaged cells release substances such as bradykinin, histamines, prostaglandins,
and potassium ions that may activate and sensitize pain receptors (Miles & Clarkson, 1994).
462 P. M. Clarkson and D. J. Newham

Slow release of these substances from the developing muscle damage may explain the
delayed sensation of soreness and pain.
Muscle soreness appears more intense at the musculotendinous junction. More
damage may occur at this site, and several studies of animals have documented that
mechanical stress can alter the integrity of the junction (Tidball, 1991). An increased number
of pain receptors in this region has also been suggested to account for the pain. However,
there are no data to substantiate this suggestion, rather pain receptors (type IVafferents)
appear to be located in the connective tissue throughout the muscle and not found in greater
concentration in any region (Miles & Clarkson, 1994).

Magnetic Resonance Imaging (MRI) and Magnetic Resonance

Spectroscopy (MRS)
These non-invasive techniques enable repeated studies on whole muscle, or a part of
it. Increases in MRI signal intensity and T2 relaxation times are thought to indicate edema
by detecting changes in intracellular water (Fried et aI., 1988). Shellock and colleagues
(1991) found that the T2 relaxation times increased for several days after eccentric but not
concentric exercise, and subclinical abnormalities persisted for 60-80 days in some subjects.
Muscle biopsies taken from abnormal areas observed on MRIs showed ultrastructural
damage (Nurenberg et aI., 1992).
Greater changes have been reported in the region of the myotendinous junction
(Shellock et at, 1991), which is usually identified as the most painful area. This study and
others (Fleckenstein et aI., 1988; Rodenburg et aI., 1994) found a large inter-subject
variability in the intensity of the response and in changes within individual muscles of a
functional group.
31p MRS studies have shown changes in the inorganic phosphate:phosphocreatine
ratio (Pi:PCr) with an unchanged muscle pH. This ratio appears normal immediately after
eccentric exercise (Aldridge et aI., 1986) but is elevated about an hour afterwards (McCully
et aI., 1988). The increase is maximal at about 24 hours after exercise and persists for up to
10 days. These changes clearly follow a different time course to pain. Moreover, the changes
in force loss after eccentric exercise appear unrelated to the altered metabolic state (Roden-
burg et aI., 1994).
An elevated Pi:PCr has also been reported in patients with various neuromuscular
diseases, many of which are not associated with pain or soreness. It has also been found after
high force concentric contractions (Newham & Cady, 1990) and so is not unique to eccentric
activity. McCully and colleagues (1988) suggested that it is caused by an increased resting
muscle metabolism, possibly reflecting repair processes such as protein synthesis. Combined
studies of MRI and MRS indicate that metabolic changes occur earlier than edema and are
more evenly distributed throughout the muscle (Rodenburg et aI., 1994).

FUNCTIONAL CONSEQUENCES OF MUSCLE DAMAGE

Changes in Force Generation

Eccentric exercise causes greater changes, including force fatigue, than comparable
contractions of either isometric or concentric activity (Jones et aI., 1989; Warren et aI.,
1993a). The resting membrane potential of isolated muscles fatigued by eccentric and
isometric contractions is reported to be similar, despite the greater force loss after the former
(Warren et aI., 1993c). Therefore, there is no evidence of excitation failure, and the ability
Associations between Muscle Soreness, Damage, and Fatigue 463

to conduct action potentials should be unimpaired. Human EMG studies have shown that
eccentric exercise causes the greatest and longest lasting changes in the total electrical
activity and mean power frequency (Kroon & Naejie, 1991).
Relatively few maximal eccentric contractions can reduce the isometric force of
humans by approximately 50% (e.g., Newham et aI., 1987). The time course of recovery of
isometric force is particularly slow, and is incomplete after two weeks in untrained SUbjects.
After exercise with a higher energy cost (either isometric or concentric), force recovery starts
almost simultaneously with the cessation of exercise and restoration of the circulation and
is usually recovered within minutes or hours.
The majority of studies have examined the effect of eccentric exercise on isometric
force generation. It is assumed these changes are comparable to those in eccentric and
concentric force, but there is no direct evidence that this is so. Most human studies have
measured maximal voluntary isometric forces and it is possible that force generation is
affected by the existence, or fear, of pain through either reflex or motivational mechanisms.
This does not seem to be the case, since superimposed electrical stimulation has shown full
voluntary activation during isometric contractions of painful human muscle (Rutherford et
aI., 1986; Newham et aI., 1987; cf. Gandevia & McKenzie, 1988). Furthermore, the changes
in human muscle are compatible with those of studies on electrically stimulated animal
muscle (Warren et aI., 1993c; Faulkner et aI., 1993).
A number of mechanical influences such as muscle force, length and velocity,
apparently playa part in the consequences of eccentric exercise. However their exact role,
and the relationship between them is not clear. Eccentric exercise performed at a long muscle
length causes greater pain in humans and fatigue in human and animal muscle than when
carried out a shorter lefigths (Jones et aI., 1989; Warren et aI., 1993a). Lieber and Friden
(1993) reported that the active strain during lengthening was the critical factor, rather than
force. It also appears that a critical number of high force contractions are required to cause
marked changes in force (Warren et aI., 1993b). These data strongly support the hypothesis
that non-contractile force-transmitting structures are damaged by eccentric exercise, in
addition to the contractile elements.
In addition to the reduction in maximal force generation, eccentric exercise also
affects contractile properties. This is demonstrated by a change in the force-frequency
relationship so that relatively lower forces are generated at low «20 Hz) frequencies
(Newham et ai., 1983a; Sargeant & Dolan, 1987). This is termed low-frequency fatigue
(LFF) and has been suggested to be the consequence of decreased calcium release by each
action potential or changes in the SR. Changes in both of these factors have been reported
after eccentric contractions as mentioned elsewhere in this chapter. However, LFF is not
unique to eccentric contractions and has been widely reported after both brief, high intensity
isometric contractions and also after prolonged, low intensity isometric activity (for reviews
see Kukulka, 1992; Enoka & Stuart, 1992). After eccentric contractions, the magnitude of
LFF appears approximately the same as the decrease in maximal isometric force generation,
that is it is greater than after isometric or concentric contractions. The recovery is rather
faster than that of maximal force generation, but still takes longer after eccentric contractions
and can take more than one week (Jones et aI., 1989). The functional significance ofLFF is
unknown, but force generation is impaired in the physiological range of motor unit discharge
rates for isometric and, presumably, eccentric activity as demonstrated by Bigland-Ritchie
and others (see Kukulka, 1992; Enoka & Stuart, 1992). Therefore, LFF would be expected
to decrease functional performance, particularly as the firing rate is decrease by fatigue, at
least in isometric fatigue.
While repeated eccentric exercise decreases and eventually eliminates pain and also
cellular structural changes, it appears to have less impact on the amount of force fatigue
caused by the exercise. However, the recovery rate is increased (Newham et aI., 1987;
464 P. M. Clarkson and D. J. Newham

Clarkson & Tremblay, 1988). It may be that repeated exercise has different effects on the
contractile and non- contractile components of the muscle and that these have different time
courses. However, both could be expected to affect force generation.
Work on the human elbow flexors show that the activation of individual muscles in
a functional group varies according to whether concentric or eccentric exercise is being
performed and also joint angle (Nakazawa et aI., 1993). Thus the interaction of synergistic
muscles is complex and variable. It cannot be assumed that changes in anyone muscle are
representative of all those in the same functional group.

Power Output

In functional terms, the ability to generate power and do work is more important than
isometric force generation, but this has been studied infrequently. McCully and Faulkner
(1985) reported a reduction in the maximal power output of mice after eccentric activity. In
human studies, peak power was reduced immediately after eccentric exercise (Beelan &
Sargeant, 1991), but only at pedaling speeds> 90 rpm. It is not known if the time course of
recovery for power generation is the same as for isometric force, nor are the exact effects of
velocity known. This area warrants further investigation.
The consequences of eccentric activity are not always to be viewed in terms of pain
and damage. Work output can be increased by approximately 40% by a brief stretch
immediately prior to shortening with no apparent increase in fatigue during short periods of
activity (de Haan et aI., 1990). This stretch-shortening cycle is frequently found during
normal activities and is a way in which the body utilizes the additional force that can be
produced by eccentric contractions.

Prolonged Loss in Range of Motion and Stiffness

When the elbow flexors are damaged by high force eccentric exercise, there is a loss
in the ability to fully flex the joint. The peak change occurs immediately after eccentric
exercise and gradually recovers in the next 5 - 7 days (Clarkson et aI., 1992). Full flexion
can be easily achieved passively.
The time course of change in the ability to flex the forearm coincides with the loss
in muscle strength. Certainly damage to the contractile units and cytoskeleton as described
previously would impact on the ability to generate tension. However, greatest damage occurs
in the days after exercise while strength is improving. Also, it is less clear how damage affects
the ability of muscles to fully shorten. Using modeling techniques, Morgan (1990) proposed
that during eccentric exercise some sarcomeres become lengthened to the point that they are
stretched until there is no overlap between the thick and thin filaments. Furthermore, upon
relaxation the overstretched sarcomeres cannot fully return to the original position, there
would be less overlap of the filaments, and less crossbridges will be able to form.
After eccentric exercise of the elbow flexors, there is an inability to fully extend the
joint, producing a sensation of stiffness. When the arm is hanging in relaxed position, the
elbow angle becomes approximately 10-20 degrees more acute immediately after exercise
and for the next 48-72 hours. Thereafter there is a gradual improvement with full recovery
about 10 days after exercise.
Jones and colleagues (1987) examined the force necessary to extend the elbow in
subjects with sore muscles after eccentric exercise of the elbow flexors. This force increased
in a similar time course to the development of muscle soreness. Similarly, Howell and
colleagues (1993) extended the elbow in a stepwise progression from 90 to 180 degrees. The
biceps and triceps muscles were electrically silent. The slope of the relationship between
Associations between Muscle Soreness, Damage, and Fatigue 465

elbow angle and the force required for extension was used as a measure of stiffness. The
slope doubled immediately after eccentric exercise and had not fully recovered 10 days later.
The mechanism to explain increased muscle stiffness is unclear. Either spontaneous
contraction of the elbow flexors or a shortening of the non-contractile tissue has been
proposed. Jones and colleagues (1987) found no surface EMG activity in the shortened elbow
flexor muscles. MRI data show that greater changes in signal intensity generally occur in
the brachialis, and surface EMG electrodes may not pick up electrical signals of the brachialis
(Shellock et ai., 1991).
Howell and colleagues (1993) assessed the influence of swelling on stiffness and
found that swelling of the upper arm (from circumference measures) and flexor muscles
(from ultrasound image analysis) was greatest 3-7 days after exercise. This time course
differed from the peak change in stiffness at 1-4 days after exercise. However, MRI indicated
the greatest increase in swelling of the forearm flexors at 1-5 days (Nosaka & Clarkson,
unpublished observation). This coincides better, but not perfectly, with the changes in muscle
shortening. Although swelling may be one contributing factor it is not the only factor. It has
also been suggested that the spontaneous shortening could be due to an abnormal accumu-
lation of calcium inside the muscle cell (Ebbeling & Clarkson, 1989), due either to a loss of
sarcolemmal integrity or a dysfunction of the sarcoplasmic reticulum (SR). Damaged SR
has been shown in biopsies taken from horses subjected to high-intensity exercise, along
with a depression in calcium uptake ability, diminished release of calcium, and increased
intracellular levels of free calcium (Byrd, 1992). Biopsies taken from a patient suffering
from spontaneous contractures induced by exercise showed a dysfunction in calcium uptake
by the SR (Brody, 1969). The role of calcium in exercise damage is still unclear.

Neuromuscular Dysfunction
Most studies of neuromuscular function after eccentric exercise focused on gross
changes such as force generation and range of motion. Recently Saxton and colleagues
(1994) examined changes in muscle tremor and proprioception after performance of high
force eccentric exercise. Tremor was assessed using an accelerometer attached to the forearm
while subjects maintained the elbow joint angle at 90 degrees. Tremor amplitude increased
immediately after and gradually returned to baseline by 72 hours. The power frequency
spectrum showed two dominant frequencies at approximately 3-5 and 8-12 Hz, which
remained largely unchanged by exercise.
In another experiment, a force-matching task was used with the control (non-exer-
cised) arm acting as the reference (Saxton et ai., 1994). The target force was 35% of the
pre-exercise maximal isometric contraction. Prior to exercise, subjects were well able to
match this force. However, in the days afterwards they consistently undershot the target;
they were generating less force than they thought. One subject could only produce 30% of
her baseline strength and thus was unable to achieve the target of35%, yet claimed that she
was exerting sufficient force to match the control arm. Instead of matching the target force,
subjects were producing 35% of the maximum that could be achieved with the weakened
exercised arm. This persisted throughout recovery.
How muscles sense force is not understood, although muscle sensations probably
arise from some combination of peripheral feedback and central feedforward mechanisms
(see Cafarelli, 1988; see Jones, Chapter 22). The feedforward mechanisms are important in
making relative judgments. For example, when two muscle groups have differing force
ability, subjects can produce the same relative force in a matching task. Apparently this holds
true when a muscle is damaged. From a performance standpoint, the CNS believed that the
forces generated by the contralateral muscle groups were similar when they were not. When
individuals have sore muscles from performance of strenuous exercise in the preceding days,
466 P. M. Clarkson and D. J. Newham

they may be unaware of the accompanying strength loss and may overestimate how much
force they are capable of achieving.
Another indication of neuromuscular dysfunction after eccentric exercise results
from studies of joint position sense (Saxton et aI., 1994). Testing was done using the
exercised limb to 1) match specific joint angles ofthe control limb; and 2) to replicate specific
angles set on the same limb. When the control arm acted as a reference, the exercised arm
consistently overshot the target joint angle. However, when the exercised limb acted as its
own reference, the joint positions were more accurately reproduced. The CNS appears to
have difficulty integrating information between the damaged and control limbs. Further
studies of neuromuscular control of exercise damaged limbs are warranted.

SUMMARY

Eccentric exercise results in muscle soreness and muscle damage. During eccentric
contractions more force and greater heat is produced compared with isometric and concentric
contractions, but the net energy generated and chemical changes are smaller. The high
mechanical tensions, muscle strain, and raised muscle temperatures appear to be factors
contributing to damage. Muscle biopsies taken from muscles that performed eccentric
contractions show histological and ultrastructural evidence of damage. This damage is
progressive until several days after exercise then regeneration ensues. A consequence of
damage is the development of muscle soreness that peaks 24-48 hours after eccentric
exercise. The exact cause of soreness is not known but inflammation and swelling are
considered factors as increased bradykinin, histamines, prostaglandins, and potassium ions
that are found in damaged muscle.
Eccentric exercise causes more fatigue, both maximal force generation and low-fre-
quency fatigue, than other forms of activity. There is a slow recovery of force that may take
longer than 3 weeks. Power output is also impaired, although this has been infrequently
studied. There is a prolonged loss in range of motion and stiffness such that the ability to
fully contract or lengthen the muscle is compromised. Evidence of neuromuscular dysfunc-
tion in the control of exercise-damaged muscle includes an increase in tremor amplitude and
a decrease in proprioception.

ACKNOWLEDGMENTS

The authors' work has been supported by the Wellcome Trust (D.J.N), the Ono Sports
Science Foundation and Yokohama City University, Japan (PM c., with K. Nosaka), the
University of Waiverhampton, England, and the University of Massachusetts (PM c., with
J.M. Saxton). Attendance of D.J.N at the 1994 Bigland-Ritchie conference was supported,
in part, by the Physiological Society, the Royal Society, the Wellcome Trust, the United States
Public Health Service (National Center for Medical Rehabilitation Research, National
Institute of Child Health and Human Development), and the Muscular Dystrophy Associa-
tion (USA).

REFERENCES
Abbott Be & Bigland B (1953). The effects of force and speed changes on the rate of oxygen consumption
during negative work. Journal of Physiology (London) 120, 319-325.
Associations between Muscle Soreness, Damage, and Fatigue 467

Abbott BC, Bigland B & Ritchie JM (1952). The physiological cost of negative work. Journal of Physiology
(London) 117, 380-390.
Aldridge R, Cady EB, Jones DA & Obletter G (1986). Muscle pain after exercise is linked with an inorganic
phosphate increase as shown by 31P NMR. Bioscience Reports 6, 663-667.
Beelan A & Sargeant AJ (1991). Effect of fatigue on maximal power output at different contraction velocities
in humans. Journal ofApplied Physiology 71, 2332-2337.
Bigland Ritchie B & Woods J (1973). Oxygen consumption and integrated electrical activity of muscle during
positive and negative work. Journal ofPhysiology (London) 234, 40P.
Bonde-Petersen F, Henriksson J & Knuttgen HG (1973). Effect of training with eccentric muscle contractions
on skeletal muscle metabolites. Acta Physiologica Scandinavica 88, 564-570.
Brody IA (1969). Muscle contracture induced by exercise. New England Journal ofMedicine 281, 187-192.
Byrd SK (1992). Alterations in the sarcoplasmic reticulum: a possible link to exercise-induced muscle damage.
Medicine and Science in Sports and Exercise 24, 531-536.
Cafarelli E (1988). Force sensation in fresh and fatigued human skeletal muscle. Exercise and Sport Sciences
Reviews 16, 139-168.
Clarkson PM, Nosaka K & Braun B (1992). Muscle function after exercise- induced muscle damage and rapid
adaptation. Medicine and Science in Sports and Exercise 24, 512-520.
Clarkson PM & Tremblay I (1988). Exercise induced muscle damage, repair and adaptations in humans.
Journal ofApplied Physiology 65, 1-6.
de Haan A, Lodder MAN & Sargeant AJ (1990). Influence of active pre stretch on fatigue of skeletal muscle.
Journal ofApplied Physiology 62, 268- 273.
Duan C, Delp MD, Hayes PD, Delp PD & Armstrong RB (1990). Rat skeletal muscle mitochondria [Ca2+]
and injury from downhill running. Journal ofApplied Physiology 68, 1241-1251.
Ebbeling CB & Clarkson PM (1989). Exercise-induced muscle damage and adaptation. Sports Medicine 7,
207-234.
Enoka RM & Stuart DG (1992). Neurobiology of muscle fatigue. Journal of Applied Physiology 72,
1631-1648.
Faulkner JA, Brooks SV & Opiteck JA (1993). Injury to skeletal muscle fibers during contractions. Conditions
of occurrence and prevention. Physical Therapy 73, 911-921.
Fleckenstein JL, Canby RC, Parkey RW, Peshock RM (1988). Acute effects of exercise on MR images of
skeletal muscle in normal volunteers. American Journal ofRadiology 151, 480-485.
Friden J & Lieber RL (1992). Structural and mechanical basis of exercise-induced muscle injury. Medicine
and Science in Sports and Exercise 24, 521- 530.
Fried R, Jolesz FA, Lorenzo AV, Francis H & Adams DF (1988). Developmental changes in proton magnetic
resonance relaxation times of cardiac and skeletal muscle. Investigative Radiology 23, 289-293.
Gandevia SC & McKenzie DK (1988). Activation of human muscles at short muscle lengths during maximal
static efforts. Journal of Physiology (London) 407, 599-613.
Giilch RW (1994). Force-velocity relations in human skeletal muscle. International Journal ofSports Medicine
15, S2-SI0.
Hough T (1902). Ergographic studies in muscular soreness. American Journal of Physiology 7, 76-92.
Howell IN, Chleboum G & Conatser R (1993). Muscle stiffness, strength loss, swelling and soreness following
exercise-induced injury in humans. Journal ofPhysiology (London) 464,183-196.
Huxley AF (1957). Muscle structure and theories of contraction. Progress in Biophysics and Biophysical
Chemistry 7, 255-318.
Jones DA, Newham DJ & Clarkson PM (1987). Skeletal muscle stiffness and pain following eccentric exercise
of the elbow flexors. Pain 30, 233-242.
Jones DA, Newham DJ, Round 1M & Tolfree SEJ (1986) Experimental human muscle damage: morphological
changes in relation to other indices of damage. Journal ofPhysiology (London) 375, 435-448.
Jones DA, Newham DJ & Torgan A (1989). Mechanical influences on long lasting human muscle fatigue and
delayed onset pain. Journal ofPhysiology (London) 412, 415-427.
Katz B (1939). The relation between force and muscular contraction. Journal of Physiology (London) 96,
45-64.
Knuttgen HG, Bonde-Peterson F & Klausen K (1971). Oxygen uptake and heart rate responses to exercise
performed with concentric and eccentric muscle contractions. Medicine and Science in Sports 3, 1-5.
Knuttgen HG, Nadel ER, PandolfKB & Patton IF (1982). Effects of training with eccentric muscle contractions
on exercise performance, energy expenditure and body temperature. International Journal of Sports
Medicine 3, 13-17.
468 P. M. Clarkson and D. J. Newham

Komi PV (1973). Relationship between muscle tension, EMG and velocity of contraction under concentric
and eccentric work. In: Desmedt JE (ed), New Developments in Electromyography and Clinical
Neurophysiology, vol 1, pp 596-606. Basel: Karger.
Kroon GW & Naejie M (1991). Recovery of the human biceps electromyogram after heavy eccentric,
concentric or isometric exercise. European Journal ofApplied Physiology 63, 444-448.
Kukulka CG (1992). Human skeletal muscle fatigue. In: Currier DP, Nelson RM (eds), Dynamics of Human
Biological Tissues, pp 164-181. Philadelphia: FA Davis.
Lieber RL & Friden J (1993). Muscle damage is not a function of muscle force but active muscle strain. Journal
ofApplied Physiology 74, 520-526.
Lowe DA, Warren GL, Hayes DA, Farmer MA & Armstrong RB (1994). Eccentric contraction-induced injury
of mouse soleus muscle: effect of varying [Ca2+]o. Journal ofApplied Physiology 76, 1445-1453.
Mair J, Koller A, Artner-Dworzak E, Haid C, Wicke K, Judmaier W & Puschendorf, B (1992). Effects of
exercise on plasma myosin heavy chain fragments and MRI of skeletal muscle. Journal of Applied
Physiology 72, 656-663.
McCully KK, Argov Z, Boden BP, Brown RL, Bank WJ & Chance B (1988). Detection of muscle injury with
31-P magnetic resonance spectroscopy. Muscle & Nerve 11, 212-216.
McCully KK & Faulkner JA (1985). Injury to skeletal muscle fibers of mice following lengthening contrac-
tions. Journal ofApplied Physiology 59, 119- 126.
Miles MP & Clarkson PM (1994). Exercise-induced muscle pain, soreness, and cramps. Journal of Sports
Medicine and Physical Fitness 34,203-216.
Miles MP, Ives J & Vincent KR (1993). Neuromuscular control following maximal eccentric exercise.
Medicine and Science in Sports and Exercise 25, S 176.
Morgan DL (1990). New insights into the behavior of muscle during active lengthening. Biophysics Journal
57,209-221.
Nadel ER, Bergh U & Saltin B (1972). Body temperatures during negative work exercise. Journal ofApplied
Physiology 33, 553-558.
Nakazawa K, Kawakami Y, Fukunaga T, Yano H & Miyashita M (1993). Differences in activation patterns in
elbow flexor muscles during isometric, concentric and eccentric contractions. European Journal of
Applied Physiology 66,214-220.
Newham DJ & Cady E (1990). A 31P study offatigue and metabolism in human skeletal muscle with voluntary,
intermittent contractions at different forces. NMR in Biomedicine 3,211-219.
Newham DJ, Jones DA, Ghosh G & Aurora P (1988). Muscle fatigue and pain after eccentric contractions at
long and short length. Clinical Science 74,553-557.
Newham DJ, Jones DA& Clarkson PM (1987). Repeated high force eccentric exercise; effects on muscle pain
and damage. Journal ofApplied Physiology 63, 1381-1386.
Newham DJ, Mills KR, Quigley BM & Edwards RHT (1983a). Pain and fatigue following concentric and
eccentric muscle contractions. Clinical Science 64, 55-62.
Newham DJ, McPhail G, Mills KR & Edwards RHT (1983b). Ultrastructural changes after concentric and
eccentric contractions of human muscle. Journal of the Neurological Sciences 61,109-122.
Nosaka K & Clarkson PM (1993). Plasma creatine kinase response to a subsequent bout of eccentric exercise
with the contralateral limb. Medicine and Science in Sports and Exercise 25, S33.
Nosaka K, Clarkson PM & Apple FS (1992). Time course of serum protein changes after strenuous exercise
of the forearm flexors. Journal of Laboratory and Clinical Medicine 119, 183-188.
Nurenberg P, Giddings C, Stray-Gundersen J, Fleckenstein JL, Gonyea WJ & Peshock RM (1992). MR
imaging-guided muscle biopsy for correlation of increased signal intensity with ultrastructural change
and delayed-onset muscle soreness after exercise. Radiology 184, 865-869.
Rodenburg JB, deBoer RW, Schiereck P, van Echeld CJA & Bar PR (1994). Changes in phosphorus compounds
and water content in skeletal muscle due to eccentric exercise. European Journal of Applied
Physiology 68, 205-213.
Rutherford OM, Jones DA & Newham DJ (1986). Clinical and experimental application of the percutaneous
twitch superimposition technique for the study of human muscle activation. Journal of Neurology,
Neurosurgery and Psychiatry 49,1288-1291.
Sargeant AJ & Dolan P (1987). Human muscle function following prolonged eccentric exercise. European
Journal ofApplied Physiology 56, 704-711.
Saxton JM, Clarkson PM, James R, Miles M, Westerfer M, Clark S & A.E. Donnelly (1994). Neuromuscular
function following maximum voluntary eccentric exercise. Medicine and Science in Sports and
Exercise 26, S115.
Shellock FG, Fukunaga T, Mink JH & Edgerton VR (1991). Exertional muscle injury: evaluation of concentric
versus eccentric actions with serial MR imaging. Radiology 179, 659-664.
Associations between Muscle Soreness, Damage, and Fatigue 469

Smith LL (1991). Acute inflammation: the underlying mechanism in delayed onset muscle soreness? Medicine
and Science in Sports and Exercise 23, 542-551.
Stauber WT (1989). Eccentric action of muscles, physiology, injury and adaptation. Exercise Sport Sciences
Reviews 17, 157-185.
Tidball JG (1991). Myotendinous junction injury in relation to junction structure and molecular composition.
Exercise and Sport Sciences Reviews 19,419-445.
Vol finger L, Lassourd V, Michaux JM, Braun JP & Toutain PL (1994). Kinetic evaluation of muscle damage
during exercise by calculation of amount of creatine kinase released. American Journal ofPhysiology
266, R434-44I.
Warren GI, Hayes DA, Lowe DA & Armstrong RB (1993a). Mechanical factors in the initiation of eccentric
contraction-induced injury in rat soleus muscle. Journal ofPhysiology (London) 464, 457-475.
Warren GI, Hayes DA, Lowe DA, Prior BM & Armstrong RB (1993b). Materials fatigue initiates eccentric
contraction-induced injury in rat soleus muscle. Journal of Physiology (London) 464, 477-489.
Warren GI, Lowe DA, Hayes DA, Karowski CJ, Prior BM & Armstrong RB (1993c). Excitation failure in
eccentric contraction-induced injury of mouse soleus muscle. Journal of Physiology (London) 468,
487-499.
Westing SH, Cresswell AG & Thorstensson A (1991). Muscle activation during maximal voluntary eccentric
and concentric knee extension. European Journal ofApplied Physiology 62, 104-108.
Woledge RC, Curtin NA & Homsher E (1985). Energetic aspects of muscle contraction. Monographs of the
Physiological Society No 41. London: Academic Press.
 34

MUSCLE FATIGUE IN OLD ANIMALS

Unique Aspects of Fatigue in Elderly Humans

1. A. Faulkner and S. V. Brooks

Institute of Gerontology and Bioengineering Program

University of Michigan
Ann Arbor, Michigan 48109-2007

ABSTRACT

Muscle atrophy, weakness, injury, and fatigue are inevitable and immutable concomitants of
old age. Atrophy results from a gradual process of fiber denervation with loss of some fibers
and atrophy of others. Fast fibers show more denervation and atrophy than slow fibers. Some
fast fibers are reinnervated by axonal sprouting from slow fibers resulting in remodeling of
motor units. With aging, the decreases in strength and power are greater than expected from
the loss in muscle mass. Contraction-induced injury is proposed as a mechanism of the fast
fiber denervation. With atrophy and weakness, human beings show a dramatic decrease in
endurance and increase in fatigability with aging, but strength and endurance training slows
the process.

INTRODUCTION

In old age, skeletal muscle structure and function are impaired significantly in all
animals, including human beings (Brooks & Faulkner, 1994a). Through a gradual process
of fiber denervation, fiber loss, and motor unit remodeling, muscle mass declines by 20%
to 40% between 30 and 80 years of age in human beings and between 12 and 28 months of
age in rodents (Brooks & Faulkner, 1994a). The muscle atrophy is caused by a combination
of a reduction in the number of muscle fibers and in the mean single fiber cross-sectional
area (CSA). Each factor appears to make an equal contribution to muscle atrophy in untrained
elderly human beings (Lexell et aI., 1988). Early studies of human beings concluded that
muscle atrophy explained the totality of the decrease in strength with aging (Grimby & Saltin,
1983). A subsequent study on EDL and soleus muscles of male mice indicated that compared
with the maximal force normalized for total fiber CSA of muscles in young (3 months) and
adult (12 months) mice, the maximal specific force of those in old (28 months) mice was
20% lower (Brooks & Faulkner, 1988). More precise assessments of the total fiber CSA of
muscles in human beings have resulted in new data which are in agreement with the data on

471
472 J. A. Faulkner and S. V. Brooks

mice and rats and conclude that muscles in elderly human beings are about 20% weaker than
those in adults (Bruce et aI., 1989).
In all species, the decreases in muscle mass and strength appear similar for males and
females (Brooks & Faulkner, I 994a). For female subjects, a sharp drop in maximal strength
is associated in time with the onset of menopause. The loss in strength at menopause is not
observed for women who receive hormone replacement (Phillips et aI., 1992). The mecha-
nism of the weakness in muscles of old animals, or in the absence of specific hormones is
not known. When maximally activated by a high calcium concentration, single permeabilized
skeletal muscle fibers from muscles of old mice develop the same maximal specific forces
as those from adult mice (Brooks & Faulkner, 1994b). This suggests that either some
substance in the cytosol of fibers in the muscles of old animals inhibits the force development
of cross-bridges, or that cross-bridges in intact fibers are not fully activated by calcium
(Brooks & Faulkner, 1994b). Following isometric contractions, a prolonged half relaxation
time has been attributed to a delayed removal of calcium from the cytosol of fibers in the
muscles of old animals (Larsson & Salviati, 1989). Compared with the deficits in specific
force associated with increasing age, even greater decreases are reported for maximal
normalized power (Kadhiresan, 1993). For muscles in old compared with adult animals, the
decrease in maximal normalized power results in part from an equivalent loss of 20% for
force during shortening compared with the loss observed for isometric force. In addition,
muscles in old animals have a lower optimal velocity for the generation of power than those
in adult animals. Consequently, for old animals the product of force and velocity, the power,
is impaired more than force alone.

MECHANISM OF FIBER DENERVATION AND MOTOR UNIT

REMODELING

Morphological, histochemical and electrophysiological data support the premise that

throughout the life span of animals, terminals of nerve degeneration occurs at the motor
end-plates of muscle fibers and the dennervation of the muscle fibers stimulates nerve
terminal and nodal sprouting (Brown et aI., 1981; Rosenheimer, 1990). The large fast
motoneurons appear to be more likely to degenerate with aging than the small slow ones
(Kanda & Hashizume, 1989). The increased vulnerability to degeneration of the large fast
motoneurons is attributed to their greater surface area, higher metabolic enzyme activity,
and more complex synaptic organization (Kanda & Hashizume, 1989). Coupled with the
greater probability of denervation of fast fibers, slow motoneurons are more effective at
reinnervating denervated muscle fibers than fast motoneurons, due in part to the more rapid
growth of the slow axons (Desypris & Parry, 1990). In muscles of old animals, the incidence
of denervated muscle fibers is increased and concurrently the ability of nerves to sprout and
regenerate is decreased (Pestronk et aI., 1980). Rosenheimer (1990) has proposed that the
limited response ofEDL muscles in old animals to partial denervation may indicate that the
muscle has already achieved its maximal capacity for sprouting. The withdrawal of trophic
factors has been proposed as the mechanism for both the denervation of muscle fibers and
the impairment in reinnervation, but no compelling evidence supports the premise (Brown
et aI., 1981; Kanda & Hashizume, 1989).
As a result of the age-related remodeling, large fast motor units decrease in number
and in their innervation ratios; whereas, slow motor units maintain motor unit number, but
increase their innervation ratio dramatically (Kadhiresan, 1993; Kanda & Hashizume, 1989).
The potential for motor unit remodeling is a function of the number of fast fibers available
for denervation and the degree of competition between fast and slow fibers in the reinner-
Muscle Fatigue in Old Animals 473

Figure 1. Maximal isometric 120

force of extensor digitorum lon-
gus muscles (EDL) from young u
(circles) and old (triangles) mice
e 0
100
~
at selected times after a protocol
of repeated stretches during con- .~ :::::- 80
e ~
-
tractions. Values are expressed as
'" .....80 60
0
a percentage of the maximal
forces of the uninjured control e ~
muscles. The data were collected ~ ~ 40
on an in situ EDL muscle pre- ·2 Young mice
paration with the nerves and :::g Old mice
20
blood flow intact. Stimulation of
the muscle was by electrical O'-----'L---'-----'---'---'--....I...---'------I.
stimulation of the peroneal
nerve. Open symbols represent
o 10 20 30
data from a protocol of 75 con- Time after Injury (days)
tractions which resulted in sig-
nificantly greater injury to the muscles in old compared with young mice. Filled symbols represent data from
a 225 contraction protocol which resulted in prolonged deficits in the muscles of old mice (data from Brooks
& Faulkner, 1990; Zerba et aI., 1990).

vation of denervated fast fibers (Desypris & Parry, 1990). Consequently, the potential for
remodeling is greatest in muscles with 50% fast and 50% slow fibers which is the charac-
teristic distribution of fiber types for untrained muscles of human beings (Faulkner et aI.,
1986). Such an interpretation is supported by the greater remodeling observed in the 70%
fast and 30% slow medial gastrocnemius muscle (Kanda & Hashizume, 1989; Kadhiresan,
1993) compared with the 97% fast and 3% slow EDL muscle (Edstrom & Larsson, 1987).
The heterogeneous fiber population of most of the 400 muscles in human beings amplifies
the remodeling of motor units that occurs in all mammalian muscles with aging and
contributes to the greater atrophy and weakness observed in the muscles of human beings
compared with those of most other mammalian species.
The mechanisms responsible for the transient state of nerve-muscle contacts through-
out life and the tenuous nature of the contact in muscles of old animals are not known. Similar
to the disruption of nerve-muscle contacts, contraction-induced injury to skeletal muscle
fibers occurs intermittently throughout life and increases in frequency and severity in
muscles of old animals (Fig. 1). In addition, recovery from contraction-induced injury,
although complete in the muscles of young and possibly adult animals, is incomplete in
muscles of old animals (Fig. 1). Furthermore, damaged and dennervated muscles in old
animals reinnervate less well than those in young animals (Carlson & Faulkner, 1995). Based
on these observations, we hypothesize that contraction-induced injury to skeletal muscle
fibers contributes to the transient nature of nerve-muscle contacts throughout life and
initiates the denervation of fibers in old animals.

CONTRACTION-INDUCED INJURY

Contraction-induced injury is defined as morphological damage to small focal groups

of sarcomeres as a result of mechanical disruption of the interdigitation of the thick and thin
filament arrays or of the Z-lines of single sarcomeres (Faulkner et aI., 1993). The mechanical
injury initiates a cascade of events that produce a more severe secondary injury after two or
three days. The secondary injury involves an inflammatory response, free radical damage,
appearance of cytosolic enzymes in the serum, and phagacytosis of elements within the
474 J. A. Faulkner and S. V. Brooks

cytosol of the damaged sarcomeres. Human beings report muscle pain associated with the
secondary injury which, as a consequence, has been termed late onset muscle soreness
(Faulkner et ai., 1993; see Clarkson & Newham, Chaper 33). When muscles in old and adult
mice are injured by comparable protocols of repeated stretches during maximal contractions,
the muscles in old mice show a more severe injury at 3 days than the muscles in the young
and adult mice. In addition, when muscles in adult and old mice are injured to the same
degree, the recovery of the muscles in the old mice is slower and not complete (Fig. 1). For
muscles in old compared with young animals, the greater susceptibility to contraction-in-
duced injury (Zerba et ai., 1990) combined with the impaired ability of muscle fibers to
regenerate following injury (Brooks & Faulkner, 1990) contributes to muscle atrophy and
has the potential to induce motor unit remodeling. Damage to muscle fibers may make fibers
more susceptible to denervation. Since denervated muscle fibers in muscles of old animals
reinnervate less successfully than those in young animals (Carlson & Faulkner, 1995), such
a mechanism could explain the lack of complete recovery from contraction-induced injury
for muscles in old animals (Brooks & Faulkner, 1990).

FATIGABILITY OF MUSCLES IN OLD ANIMALS

Fatigue may occur in a fiber, motor unit, whole muscle, or a group of muscles. Fatigue
is highly specific to different types of fibers to such an extent that the characteristics of
fatigue are an intrinsic part of fiber classifications (Burke et ai., 1973). During repeated
isometric contractions, slow (S) fibers or motor units are highly resistant to fatigue,
fast-fatigable (FF) fibers or motor units fatigue rapidly to almost zero force within a few
minutes, and fast fatigue resistant (FR) fibers or motor units are intermediate, fatiguing to
low forces over a period of hours. The differences among the three fiber types are due to
biochemical (Peter et ai., 1972), metabolic (Peter et ai., 1972) and mechanical (Brooks &
Faulkner, 1991) properties of the fibers that determine cycling rates and efficiency of force
or power production (Barclay et ai., 1993) and consequently the capability of the fiber to
generate and sustain force and power. Although challenged regarding the discrete nature of
the three fiber types, the basic utility of these three functional types of fibers has withstood
twenty years of rigorous investigation (Faulkner et ai., 1994). Dependent on the circum-
stances, each element of the motor system in young, adult, and old animals may be involved
in the development of fatigue, from the drive of suprasegmental centers to the interaction of
the contractile proteins (Enoka & Stuart, 1992). In terms of task dependency, the force- fati-
gability relationship, the hyperbolic force or power-endurance relationships, and the sense
of effort (Enoka & Stuart, 1992), the difference between the fatigue response of muscles in
old animals, compared with those in young or adult animals, appears to be quantitative rather
than qualitative. Furthermore, the multicausitive nature of fatigue and the mechanisms
responsible for fatigue described for muscles in young animals appear to be equally
appropriate to the phenomenon of fatigue in muscles of old animals.
Atrophy and weakness of muscles in animals of any age increase the susceptibility
of the muscles to fatigue when performing a task of absolute force or power (Faulkner et ai.,
1994). Under circumstances of muscle atrophy and weakness, the task represents a greater
proportion of maximal strength or power and based on the force-fatigue relationship this
translates into a shorter endurance time (Enoka & Stuart, 1992). Consequently, when
presented with the same task, fatigability is increased in the elderly (Makrides et ai., 1985;
Overend et ai., 1992). Surprisingly, compared with control muscles atrophied muscles often
show a greater resistance to fatigue when tests of fatigability are designed relative to maximal
force or power (Larsson & Karlsson, 1978). The mechanism responsible for the resistance
to fatigue of atrophied muscles is not clear, but the atrophy of fibers increases capillary
Muscle Fatigue in Old Animals 475

density and may also increase the concentration of mitochondria and metabolic enzymes. In
the muscles of old animals, the shift in fiber types from fast to slow (Kanda & Hashizume,
1989) has implications for the maintenance of force development because slow fibers are
more resistant to fatigue during isometric contractions than are fast fibers (Burke et aI., 1973).
Consequently, the increased proportion of slow fibers even coupled with muscle and mean
single fiber atrophy may increase endurance and decrease fatigability during low force tasks
or tests of endurance at relative values of maximal voluntary contraction strength (Larsson
& Karlsson, 1978). These factors explain the scattered reports of older subjects performing
better than young subjects in relative tests of strength and endurance (Vandervoort et aI.,
1986).
Fatigue is most often observed during physical activities that require high sustained
force, high power short term repetitive contractions, or low power long term repetitive
contractions (Faulkner et aI., 1994). The development of fatigue when muscles shorten
during contractions is more complex than fatigue during isometric contractions (Brooks &
Faulkner, 1991). The complexity arises from a fatigue due to decreases in both shortening
force and the velocity of shortening (Faulkner et aI., 1994). Despite the usefulness of the
concept of fatigue in understanding the underlying mechanisms of impaired development of
force and power, in general animals avoid inducing fatigue. This is evident in diaphragm
function wherein both patients with chronic obstructive pulmonary disease and athletes
performing maximal exhaustive exercise avoid inducing fatigue of the diaphragm muscle
except under the most demanding conditions (Johnson et aI., 1993). Exceptions wherein
fatigue is rapid and profound are the dramatic chases of carnivores, the equally dramatic
flight of the prey and the performance of human beings during sprint and endurance sports
events. Even under these circumstances, the most direct correlate of the success of the hunt,
the flight, or the athletic performance is on the maintenance of a high power output, rather
than on the induction of fatigue (Faulkner et aI., 1994).

MAXIMAL SUSTAINED NORMALIZED POWER OF MUSCLES IN

OLD ANIMALS

We have proposed that the maximal sustained normalized power (sustained power)
is the most valid measurement of the functional capability of a muscle, a limb, or a total
organism (Brooks & Faulkner, 1991; Faulkner & Brooks, 1993). Sustained power output
was selected as the most appropriate measure because movement, in addition to being critical
for all animals, is the essence of the human experience. Although postural stability and the
stretching of agonists are each a part of most, if not all, movements, the essential element is
the driving stroke of cross-bridges during shortening (Faulkner et ai. 1994). The sustained
power can be measured by beginning at low levels of power output and increasing the output
systematically with regard to both duration and intensity (Fig. 2). Beginning at a low power
is necessary to allow the appropriate responses in muscle respiration, circulation and
metabolism. A transition from rest to a high power output results in a premature onset of
fatigue without attaining maximal sustainable power output (Brooks & Faulkner, 1991). The
increments in the power required and the duration at each power must be designed to allow
power output to plateau at each level. Consequently, the test must be designed for the
capacities of the tissue or organism being tested.
For human beings with muscle temperatures of -35°C, the estimates for the values
for maximal normalized power during a single contraction are 225 W /kg for fast fibers and
65 W/kg for slow fibers (Faulkner et aI., 1994). The difference results from the three-fold
greater maximal velocity of unloaded shortening of fast compared with slow fibers and the
476 J. A. Faulkner and S. V. Brooks

12
..
CI)
10
Figure 2. The relationship of
the cycle rate of repeated con-
tractions (cycles per minute)
~
Q., and the maximal sustained
"8 8 normalized power for exten-
.;a
'Iil
::s
~...1;; 6 EDL muscles of mice:
sor digitorum longus muscles
in young (circles), adult
til (squares), and old (triangles)
e ~ 0 Young mice. As cycle rate is in-
~ 4 creased from very low values,
.~ 0 Adult
the maximal power that can be
::g 2 6. Old sustained increases until a
maximum is reached. Further
0 increases in cycle rate result in
0 100 200 300 400 500 600 rapid decline in power output
(data from Brooks & Faulk-
Cycles per minute ner, 1991).

less curved force-velocity relationship of fast fibers. The optimal velocity for power of fast
and slow fibers occurs at -30% of the maximal velocity of unloaded shortening and -30%
of the maximal force (Faulkner et aI., 1986). With aging, the decrease in the optimal velocity
for power in muscles in old animals results, at least in part, from motor unit remodeling
which reduces the fast to slow fiber ratio (Kanda & Hashizume, 1989; Larsson & Karlsson,
1978; Lexell et aI., 1988). Other factors that upregulate the expression of slower and
downregulate the expression of faster isoforms of myosin, or both may also contribute to a
decrease in maximal normalized power.
The sustained power of a muscle is a function of the fiber composition and the
oxidative capacity of the muscle. The capability of a muscle to sustain power output depends
on the balance between energy output and energy supply (Brooks & Faulkner, 1991). The
ability of the blood flow to supply oxygen to the mitochondria in muscle fibers and the
capacity of the muscle fibers for oxidative metabolism will determine the percentage of
maximal power output attained during a single contraction that can be sustained over time
(Faulkner et aI., 1994). After a warm-up at a moderate intensity of -10% of maximal power
during a single contraction, highly trained persons can maintain 15% of maximal power, or
-30 W/kg, for several hours. The dramatic decrease in power from the value during a single
maximal contraction to a sustained power is a function of the introduction rest periods
between each of a series of repeated contractions, a duty cycle, and the decrease from
maximal force development to a metabolically determined equilibrium force development
determined by the duty cycle and the metabolic capabilities of the muscle (Faulkner &
Brooks, 1993).
The decreased contractile (Brooks & Faulkner, 1991) and metabolic (McCully et aI.,
1993) capabilities of muscles in old mice combine to produce a 50% decrease in the sustained
power compared with the value for muscles in young mice and 25% of the value for muscles
in adult mice (Fig. 2). These decrements in maximal sustained normalized power are in
excellent agreement with the loss in endurance performance capacity of untrained human
beings between the ages of30 and 80 years (Fig. 3) The decrement in sustained power with
age is due to decreases in specific force, optimal velocity for power, and energy delivery and
conversion. A second phenomenon evident in the test of sustained power is the decrease in
the cycles per minute at which maximal sustained power was attained (Fig. 2). The cycle
rate at which sustained power is maximal decreased from 500 cycles per minute (cpm) for
young mice to 350 cpm for adult mice and finally to 150 cpm for old mice. In addition, as
cycle rates are further increased, the ability of muscles in mice of all ages to generate power
Muscle Fatigue in Old Animals 477

decreases rapidly. For human beings, this decrease in the ability of muscles to perform a
contraction and recover and be ready to perform another identical contraction is reflected in
a decrease in strides per minute walking, stair-climbing, running, or skiing; strokes per
minute in swimming; or revolutions per minute cycling. The loss in the sustained power with
aging is a result of both the decreased cycling rate and the lower power generated with each
cycle. Regardless of the measure made of the functional capability of skeletal muscles, the
muscles of old animals demonstrate declines of up to 50% in endurance or an equivalent
increase in fatigability.

ATROPHY, WEAKNESS, INJURY AND FATIGUE INTRINSIC

FUNCTIONS OF AGING
At any age throughout the life span of animals, a decrease in physical activity initiates
decreases in muscle mass and in the capacity to develop force and power (Faulkner et aI.,
1994). Consequently, a key issue is whether the muscle atrophy and weakness observed in
old animals is simply a function of the decrease in physical activity or whether these
impairments in skeletal muscle structure and function are immutable and inevitable conse-
quences of basic biological changes that occur in skeletal muscles with aging. Three widely
disparate sets of data strongly support the premise that some degree of atrophy and weakness
is immutable and an inevitable consequence of aging. First, at any given age the muscle
atrophy and weakness observed with immobilization of limbs, or a decrease in physical
activity, is reversed completely by appropriate conditioning of the atrophied and weak
skeletal muscles (Faulkner et aI., 1994). Although some reversibility is evident even in very
old people (Frontera et aI., 1988; Meredith et aI., 1989), as evidenced by gains in strength,
power, and endurance, the older person never attains the status possible at an earlier age
(Fig. 3).
Second, athletes competing in masters' track and field events are highly motivated
to train as hard as younger athletes. Despite this high level of motivation and rigorous
training, the records for performance in all events that involve strength, power, or endurance
show an inevitable decline with increasing age (Fig. 3). The decline appears to be more rapid
for events requiring high power than for the endurance events (Schulz & Curnow, 1988).
The slower decline in endurance than in maximal power has given rise to the premise that
aerobic power deteriorates less rapidly than muscle power. With regard to data on track and
field events, this premise is suspect because of the varying roles of relative and absolute
loads in the performance of track compared with field events, respectively. In field events,
the muscle power involves tossing or throwing an absolute mass, the shot or the discus,

___ ShotputlDiscus
120
........ Marathon
~100 -.- Basketball
Figure 3. Physical performance data OS 80
plotted against the age of the per-
former for the marathon run. the shot ..
~
60 ..........•
put/discus throw, and the number of ~
points scored per game by a basket- E 40
ball player. The data for the marathon ~ 20
run and the shot put and discus throw ~
are from Moore (1975) and the data 0
for the basketball player are from the 10 20 30 40 50 60
NBA Register 1992-1993 edition. Age (years)
478 J. A. Faulkner and S. V. Brooks

which constitutes an increasing percentage of maximal muscle power with aging. In contrast,
the endurance events in track require movement of the total organism which reflects, at least
in part, the smaller muscle mass in older runners. Clearly, the relative rates of decline are
highly complex and subject to multiple influences. An even more precipitous decline is
observed for the performance of professional athletes in the team sports of baseball,
basketball, and football (Fig. 3). In the team sports, the decline in performance might be
heightened by injuries, direct competition with younger athletes, or both.
Finally, during the 100 years of the modem Olympiad, the Olympic records for
different power and endurance events have improved by 20% to 90% (Schulz & Curnow,
1988). The improvements in olympic records are attributable to improvements in training
and equipment, intensity of training, and the skill of the competitors. Despite the improve-
ments in the record-setting performances of both power and endurance events, in all cases,
the ages of the record-setting athletes have remained between 16 and 31 years (Schulz &
Curnow, 1988). Thus, declines in strength, power and endurance are due to immutable and
inevitable changes in the basic properties of skeletal muscles which are not reversible with
training.
With aging, lower absolute values of strength, power, and endurance for an untrained
compared with a trained person will bring the untrained person to the threshold of impaired
mobility, inability to perform the activities of daily living, and the increased probability of
experiencing accidents and falls at an earlier age than trained subjects. The impairments in
muscle function arise from muscle atrophy, weakness, fatigue, and injury (Brooks &
Faulkner, 1994a; Faulkner et aI., 1994). Consequently, the lower values of strength, power
and endurance of untrained people initiate a cascade of events that lead to an earlier onset
of disabilities that affect the quality of life. Even in very old untrained subjects, programs
in strength (Frontera et aI., 1988) and endurance (Overend et aI., 1992) training result in
significant improvements. Furthermore, careful progressive training in activities that involve
stretches of maximally activated muscles enable old persons to perform such activities
without injury. The conclusion is that strength and endurance training for old people provide
the opportunity of more years free of immobility and disability. The solution for the
individual and for health professionals administering programs for elderly people is the
maintenance of sufficient strength, power, and endurance to remain above the threshold of
an impaired ability to perform the activities of daily living throughout one's life. Such a life
style will not prevent the inevitable decline in the functional capabilities of skeletal muscle,
but it will defer the impingement on the quality of life to fewest possible years.

ACKNOWLEDGMENTS

This research was supported by United States Public Health Service grant AG 06157
(to J.A.F.), and a postdoctoral fellowship award, AG 00114 (to S. V.B.). Attendance of J.A.F.
at the 1994 Bigland-Ritchie conference was supported, in part, by the University of Arizona
Regents' Professor funds of Douglas Stuart.

REFERENCES

Barclay CJ, Constable JK & Gibbs CL (1993). Energetics of fast-and slow-twitch muscles of the mouse.
Journal of Physiology (London) 472, 61-80.
Brooks SV & Faulkner JA (1988). Contractile properties of skeletal muscles from young, adult, and aged mice.
Journal of Physiology (London) 404, 71-82.
Muscle Fatigue in Old Animals 479

Brooks SV & Faulkner JA(1990). Contraction-induced injury: recovery of skeletal muscles in young and old
mice. American Journal of Physiology 258 (Cell Physiology 27), C436-C442.
Brooks SV & Faulkner JA (1991). Maximum and sustained power of extensor digitorum longus muscles from
young, adult and old mice. Journal of Gerontology: Biological Science 46, B28-33.
Brooks SV & Faulkner JA (1994a). Skeletal muscle weakness in old age: underlying mechanisms. Medical
Science and Sports Exercises 26, 432-439.
Brooks SV & Faulkner JA (1994b). Isometric, shortening and lengthening contractions of muscle fiber
segments from adult and old mice. American Journal of Physiology 267 (Cell Physiology 36),
C507-C513.
Brown MC, Holland RL & Hopkins WG (1981). Motor nerve sprouting. Annual Reviews of Neuroscience 4,
17-42.
Bruce SA, Newton D & Woledge RC (1989). Effect of age on voluntary force and cross-sectional area of
human adductor pollicis muscle. Quarterly Journal ofExperimental Physiology 74, 359-362.
Burke RE, Levine DN, Zajak FE, Tsairis P & Engel WK (1973). Physiological types and histochemical profiles
in motor units of the cat gastrocnemius. Journal of Physiology (London) 234, 723-748.
Carlson BM & Faulkner JA (1995). Skeletal muscle regeneration and aging: the influence of innervation.
Journal of Gerontology In press.
Desypris G & Parry DJ (1990). Relative efficacy of slow and fast-motoneurons to reinnervate mouse soleus
muscle. American Journal of Physiology 258 (Cell Physiology 27), C62-C70.
Edstrom L & Larsson L (1987). Effects of age on contractile and enzyme-histochemical properties of fast-and
slow-twitch single motor units in the rat. Journal of Physiology (London) 392. 129-145.
Enoka RM & Stuart DG (1992). Neurobiology of muscle fatigue. Journal of Applied Physiology 72,
1631-1648.
Faulkner JA & Brooks SV (1993). Fatigability of mouse muscles during constant length, shortening, and
lengthening contractions: interactions between fiber types and duty cycles. In: Sargeant T, Kernell D
(eds.) Neuromuscular Fatigue, pp 116-123. Amsterdam: Royal Netherlands Academy of Arts and
Sciences.
Faulkner JA, Brooks SV & Opiteck JA (1993). Injury to skeletal muscle fibers during contractions: conditions
of occurrence and prevention. In: Binder-Macleod SA (ed.), Physical Therapy. Special Edition, pp
911-921. Alexandria, VA: American Physical Therapy Association.
Faulkner JA, Claflin DR & McCully KK (1986). Power output of fast and slow fibers from human skeletal
muscles. In: Jones NL, McCartney N, McComas J, (eds.), Human Power Output, pp. 81-91.
Champaign, IL: Human Kinetics Publishers Inc.
Faulkner JA, Green HJ & White TP (1994). Skeletal muscle responses to acute and adaptations to chronic
physical activity. In: Bouchard C, Shephard RJ, Stephens T (eds.), Physical Activity, Fitness & Health,
pp. 343-357. Champaign, IL: Human Kinetics Publishers, Inc.
Frontera WR, Meredith CN, O'Reilly KP, Knuttgen HG & Evans WJ (1988). Strength conditioning in older
men: skeletal muscle hypertrophy and impaired function. Journal of Applied Physiology 64, 1038-
1044.
Grimby G & Saltin B (1983). The aging muscle. Clinical Physiology 3,209-218.
Johnson BD, Babcock MA, Suman OE & Dempsey JA (1993). Exercise-induced diaphragmatic fatigue in
healthy humans. Journal of Physiology (London) 460, 385-405.
Kadhiresan VA (1993). Functional Properties of Motor Units in Medial Gastrocnemius Muscles of Rats:
Remodelling in Old Age. Ph.D. Dissertation, Univ. Microfilm [Link]. No. 9409723. Ann Arbor, MI:
University of Michigan.
Kanda K & Hashizume K (1989). Changes in properties of the medial gastrocnemius motor units in aging.
Journal of Neurophysiology 61, 737-746.
Larsson L & Karlsson J (1978). Isometric and dynamic endurance as a function of age and skeletal muscle
characteristics. Acta Physiologica Scandanavica 104, 129-136.
Larsson L & Salviati G (1989). Effects of age on calcium transport activity of sarcoplasmic reticulum in fast-
and slow-twitch rat muscle fibres. Journal ofPhysiology (London) 419, 253-264.
Lexell J, Taylor CC & Sjostrom M (1988). What is the cause of the ageing atrophy? Total number, size and
proportion of different fiber types studied in whole astus lateralis muscle from 15- to 83-year-old
men. Journal of Neurological Science 84, 275-294.
Makrides L, Heigenhauser GJ, McCartney N & Jones NL (1985). Maximal short term exercise capacity in
healthy subjects aged 15-70 years. Clinical Science 69, 197-205.
McCully KK, Fielding RA, Evans WJ, Leigh Jr, JS & Posner JD (1993). Relationships between in vivo and
in vitro measurements of metabolism in young and old human calf muscles. Journal of Applied
Physiology 75,813-819.
480 J. A. Faulkner and S. V. Brooks

Meredith CN, Frontera WR, Fisher EC, Hughes VA, Herland JC, Edwards J & Evans J (1989). Peripheral
effects of endurance training in young and old subjects. Journal ofApplied Physiology 66, 2844-2849.
Moore DH (1975). A study of age group track and field records to relate age and running speed. Nature 253,
264-265.
Overend n, Cunningham DA, Paterson DH & Smith WD (1992). Physiological responses of young and elderly
men to prolonged exercise at critical power. European Journal ofApplied Physiology 64, 187-193.
Pestronk A, Drachman DB & Griffin J (1980). Effects of aging on nerve sprouting and regeneration.
Experimental Neurology 70, 64-82.
Peter JB, Barnard RJ, Edgerton VR, Gillespie CA & Stemple KE (1972). Metabolic profiles of three fiber
types of skeletal muscle in guinea pigs and rabbits. Biochemistry 14, 2627-2633.
Phillips SK, Rook KM, Siddle NC & Woledge RC (1992). Muscle weakness in women occurs at an earlier
age than in men, but strength is preserved by hormone replacement therapy. Clinical Science 84,
95-98.
Rosenheimer JL (1990). Ultraterminal sprouting in innervated and partially denervated adult and aged rat
muscle. Neuroscience 38, 763-770.
Schulz R & Curnow C (1988). Peak performance and age among superathletes: track and field, swimming,
baseball, tennis, and golf. Journal of Gerontology: Psychological Sciences 43, PI13-120.
Vandervoort AA, Hayes KC & Belanger AY (1986). Strength and endurance of skeletal muscle in the elderly.
Physiotherapy Canada 38, 167-173.
Zerba E, Komorowski TE & Faulkner J A (1990). Free radical injury to skeletal muscles of young, adult, and
old mice. American Journal of Physiology 258 (Cell PhYSiology 27), C429-C435.
 35

HISTORICAL PERSPECTIVE: A
FRAMEWORK FOR INTERPRETING
PATHOBIOLOGICAL IDEAS ON HUMAN
MUSCLE FATIGUE

R. H. T. Edwards, V. Toescu, and H. Gibson

Muscle Research Centre

Department of Medicine
University of Liverpool
P.O. Box 147
Liverpool L69 3BX
United Kingdom

ABSTRACT

The flow of ideas on the causes of human muscle fatigue appear to have been established in
the literature during the last century. Critical analysis had to await innovative experimental
designs or techniques. Progress has come particularly from the recognition of inconsisten-
cies, particularly in clinical conditions in which there are alterations in the supply of energy
or contractile function. While there is a continuing search for a unique cause offatigue, much
evidence points to there being different causes according to the type of muscular activity or
clinical condition. "Nature's experiments" (patients exhibiting isolated defects of function
or metabolism) offer unique opportunities for understanding the relative importance of
particular levels of metabolic organization or physiological control. Theories of limiting
biochemical processes and the Ca- kinetic basis of electromechanical coupling defects are
both essentially "single-cell" models offatigue. Functional requirements appear to determine
the diversity of structure and organization of motor units. This would suggest that a "muscle
cell population" approach would take into account the consequences of disease altering the
number, type offunctioning fibers or intrinsic strength of individual fibers. A graphical model
is offered to allow a possible interpretation of the cause of fatigue in different forms of
exercise and clinical conditions.

SETTING THE SCENE

This chapter begins by identifying a few of the most evident milestones in develop-
ment of ideas on muscle fatigue.

481
482 R. H. T. Edwards et al.

ORIGINS OF BIOLOGICAL ELECTRICITY AND THE ROLE THAT

STIMULATION HAD IN UNDERSTANDING FATIGUE

In 1786, Luigi Galvani studied the effects of atmospheric electricity upon

dissected frog muscles. His 1791 Commentary on the Effects of Electricity on Muscular
Motion (cited in Rasch & Burke, 1967) is probably the earliest explicit statement of
the presence of electrical potentials in nerve and muscle, although it was not until
1838 that the word tetanize was introduced (Matteucci, 1838; cited in Rasch & Burke,
1967). The study of animal electricity at once became the absorbing interest of the
physiological world. A prominent name among the early students of the subject was
Emil DuBois-Reymond (1818-1896) to whom Rosenthal (1883) dedicated his book,
Muscles and Nerves, because it was DuBois-Reymond who laid the foundations of
modern electrophysiology.
Following the discoveries of Luigi Galvani (1737-1798) and those of Michael
Faraday (1791-1867) in establishing the principles of electricity and magnetism, electrical
stimulation became well established as a tool for studying muscle function and fatigue.
Angelo Mosso, in the English translation of his book, Fatigue (Drummond & Drummond,
1915), describes how his interest was kindled in the subject after he met Professor Hugo
Kronecker in Leipzig in 1873. He observed Kronecker's latest experiments on isolated
muscles and these formed a model for Mosso's subsequent experiments in man. In muscles
from frogs, Kronecker " ... succeeded in obtaining 1,000 and even 1,500 contractions, one
after the other, with the greatest regularity. As the contractions follow one another, their
height diminishes in proportion as the fatigue increases, and goes on diminishing with
regularity until it disappears altogether."
This is know as the fIrst of Kronecker's Laws of Fatigue i.e., "The curve of fatigue
of a muscle which contracts at regular intervals, and with equally strong induction shocks
is represented by a straight line." While more recent observations have not confIrmed the
straight line relationship between force and time in all forms of fatigue, this is, nevertheless,
a useful starting point. There are forms of muscular activity in which the loss of force with
time in human muscle is fairly precipitate as in some individuals doing voluntary muscular
activity (as illustrated in Mosso's book), and with some forms of exercise involving ischemia
(e.g., Merton, 1954). The probability is that the isolated muscles in which Kronecker
demonstrated this Law were somewhat anoxic as would be suggested today by the study of
muscles treated with a metabolic inhibitor.
The second is: "The difference in the height of the contractions is less when the
intervals of time are greater. In other words, the height of the contractions diminishes the
more rapidly, the more rapid is the rhythm in which they are produced and vice versa" The
second Law clearly establishes the dependence of fatigue on the frequency of electrical
stimulation and thus on the frequency of contractions in voluntary muscular activity such as
isometric contractions in which the tendency to fatigue is related to the frequency (and
duration) of contractions.
In Julius Althaus' book, A Treatise on Medical Electricity (1873), which summa-
rizes techniques for examining neuromuscular function in patients, early controlled
experiments on fatigue in man are cited. A more analytical approach had already been
developed by Duchenne and published in 1872 in what was probably his life's work (De
Electrisation Localisee et de son application a la pathologie et la therapeutique par
courants induits et par courants galvaniques interompus et continus). Duchenne, who is
well known for his clinical accounts of the muscular dystrophy which bears his name, is
less well known for his important physiological experiments in which he systematically
stimulated virtually all the muscle groups in the human body in order to establish their
Historical Perspective 483

actions. He devised cloth-covered electrodes for percutaneous stimulation and was the
first to use alternating current, suggesting the name faradic for this form of stimulation.
His work had earlier been published in 1867 under the title Physiologie des Mouvements
(translated by Kaplan, 1959).
Jolly (1895), using a Marey capsule placed on the skin with a smoked drum
kymograph for recording mechanical responses and a double coil stimulator for faradic
stimulation (30-60 Hz) demonstrated for the first time that the fatigue and paresis of
Erb-Goldflam disease depended on a failure in the neuromuscular system. He applied the
faradic current at repeated short intervals or maintained it for about 1 min. This resulted in
an unusually rapid deterioration of the size of the muscle responses which recovered after
rest. These observations led him to propose the name myasthenia gravis (myasthenia gravis
pseudoparalytica) for the disease.
These examples illustrate the early forging of ideas for the use of electricity in
analyzing the function of muscle, now commonplace in physiology and clinical labora-
tories.

CONCEPTS OF SITES OF MUSCLE FATIGUE

Michael Foster, one of the founders of the Physiological Society, expressed the
contemporary views on the neural contribution to fatigue and opens consideration of the
metabolic factors responsible for the development of human muscle fatigue in his Textbook
ofPhysiology (1883).
The sense of fatigue of which, after prolonged or unusual exertion, we are conscious in our own
bodies, is probably of complex origin, and its nature, like that of the normal muscular sense of
which we shall have to speak hereafter, is at present not thoroughly understood. It seems to be in
the first place the result of changes in the muscles themselves, but is possibly also caused by
changes in nervous apparatus concerned in muscular action, and especially in those parts of the
central nervous system which are concerned in the production of voluntary impulses. In any case
it cannot be taken as an adequate measure of the actual fatigue of the muscles; for a man who says
he is absolutely exhausted may under excitement perform a very large amount of work with his
already weary muscles. The will in fact rarely if ever calls forth the greatest contractions of which
the muscles are capable .... A certain number of single induction-shocks repeated rapidly, say
every second or oftener, bring about exhaustive loss of irritability more rapidly than the same
number of shocks repeated less rapidly, for instance every 5 or 10 seconds. Hence tetanus is a
ready means of producing exhaustion.... In exhausted muscles, the elasticity is much diminished,
the tired muscle returns less readily to its natural length than does the fresh one .... The exhaustion
due to contraction may be the result: either of the consumption of the store of really contractile
material present in the muscle. Or of the accumulation in the tissue of the products of the act of
contraction. Or of both of these causes.

The sense of effort was approached in an analytical if rather discursive manner by

Augustus Waller (1891), another early member of the Physiological Society, who argued as
follows:
The question is: does normal voluntary fatigue depend upon central expenditure of energy, or upon
peripheral expenditure of energy, or upon both factors conjointly; and if upon both, in what
proportion upon each? The form of this question is justified as follows; a maximum voluntary
effort may decline: 1) by decline of cerebral motility; 2) by sub-central or spinal block; Central.,
3) by motor end-plate block; 4) by decline of muscular energy; Peripheral. On the human subject
it is impossible to separate factors I and 2, we must therefore embrace them under the term central.
The third factor can be more easily demonstrated upon isolated parts oflower animals; its existence
can also - if less assuredly - be demonstrated upon man; but for the cleared appreciation of the
central versus the peripheral factors I do not dwell upon this sub-distinction, and include the effects
of factors 3 and 4 under the common term peripheral.
484 R. H. T. Edwards et al.

Following careful experimentation Waller concludes:

That in voluntary fatigue the degree of change is on decreasing ratio from centre to periphery. In
other words that the cell of higher function is, relatively to the amount of effect which it can
produce, more exhaustible than the cell which is subordinate to it in the cerebro-muscular chain.
So that central fatigue is protective from peripheral fatigue; and although normally in the body the
dynamic effect of the physiological stimulus emitted by a "motor" centre far exceeds the maximal
effect which can be elicited by direct experimental excitation, yet that physiological stimulus, by
virtue of the rapid exhaustibility of the organ from which it proceeds, cannot normally "overdrive"
and exhaust the subordinate elements of the motor apparatus.
Mosso devised the ergograph which he used on a colleague to repeat Kronecker's
work and found that "We have a great change from the straight line found by Kronecker
as the expression of fatigue in frog's and dog's muscles after separation from the body.
This shows that in man the phenomenon is considerably more complex." However he
goes on:
To eliminate the mental element which might alter the fatigue curve of the muscle, I thought of
stimulating the nerve of the arm, or rather the flexor muscles of the fingers .... Fatigue was
produced with the same curve as when the muscles contracted voluntarily... one must conclude
that the mental factor does not exercise a preponderating influence, and that fatigue may even be
a peripheral phenomenon.
In Bainbridge's Physiology ofMuscular Exercise (Bainbridge, 1931), the following
summary of his views of the site of fatigue in different forms of muscular activity is to be
found.
Fatigue after exertion is characterised by diminished capacity for performing work, this being
usually accompanied by certain subjective sensations. The sensation offatigue does not necessarily
correspond with the actual fatigue of an individual when this is measured by his capacity for work.
There are two types of fatigue, one originating entirely within the central nervous system, the other
arising partly in the nervous system and partly within the active muscles. The former is of common
occurrence, whereas the latter occurs comparatively infrequently. Industrial fatigue is usually of
the first type. There is no clear evidence that the products of muscular activity take any part in
bringing about fatigue of the central nervous system.
Perhaps the most convincing argument for a peripheral cause of fatigue this century
came from a study by Merton on a single individual. " ... anyone who possesses a
sphygmomanometer and an open mind can readily convince himself that the site of fatigue
is in the muscles themselves" (Merton, 1956). "If it were central, recovery should occur
when effort ceases, for there is no interference with the circulation to central structures."
(Merton, 1954). This study has proved to have been a particularly important milestone in
the understanding of human muscle fatigue. Not only did it show beyond any doubt (in
isometric contractions under ischemic conditions) that fatigue was peripheral, but it also
showed that force need not follow alterations in action potential amplitude (see below), and
that recovery from the state of fatigue required restoration of the circulatory oxygen supply
(see also Gandevia et aI., Chapter 20).
The debate as to whether the site of fatigue was central or peripheral was thus
well established by the early part of the twentieth century. Established too was the use
of electrical stimulation as a tool for analyzing fatigue in particular forms of activity.
Thereafter, debate has been ongoing as to the specific causes of fatigue; those looking
for peripheral causes and other workers identifying the role of central mechanisms. While
major milestones in the understanding of fatigue mechanisms have come about through
the advance in the tools for the study of muscle, others have come about through the
study of muscle disease, the pathobiological processes that result in the premature and
delayed development of fatigue. The following paragraphs examine the evolution of
concepts, the hypotheses that lay behind fatigue development in normal muscle through
analysis of muscle disease.
Historical Perspective 485

CENTRAL MECHANISMS OF FATIGUE

The work of Waller, The sense ofeffort: an objective study (1891), alluded to above
illustrates the extensive efforts put in by early workers in unraveling the complexities of the
central processes contributing to fatigue. While much work has gone into identifying and
quantifying central contributions, the manifestation of generalized fatigue and myalgia has
been well documented over the last 200 years. This is well illustrated by an example from
Lewis' Soldier's Heart and Effort Syndrome (1918).
Fatigue is an almost universal complaint of our patients. In the "effort syndrome" group it varies
in degree and, when it can be gauged, is an excellent index of the severity of the affection. The
mildest cases suffer such fatigue on exercise as would a healthy man who is out of training, in the
severer cases it is experienced after very brief and simple exercises. Lassitude is especially
prominent in the early morning and late afternoon. More rarely, fatigue on exercise proceeds to
exhaustion; actual collapse from exhaustion on exercise is not seen because it is guarded against.
Fatigue and the early signs of exhaustion appear objectively in the expression of the face and in
the droop of the body; a material feeling of weakness is usually accompanied by uncontrollable
tremor of the hands, or shaking of the legs. The symptoms are those which are found in healthy
subjects submitted to strenuous exercise. That they have the same origin is rendered probable by
their association with a general feeling of malaise and in some patients with a rapidly developed
and severe "stiffness" of the muscles.

Now known as the chronic fatigue syndrome, similar symptoms have been described
over the years under many various names including irritable heart (Da Costa, 1871), Soldier's
heart (Lewis, 1918), neurocirculatory asthenia (Oppenheimer et aI, 1918), effort syndrome
(Lewis, 1918; Wood, 1941) and autonomic imbalance (Kessel & Hyman, 1923). Shorter in
his book, From Paralysis to Fatigue (1992), takes an historical perspective of psychosomatic
illness that may contribute to the perception of this form of fatigue. Stokes and colleagues
(1988) and Gibson and coworkers (1993a) have shown that this type of fatigue is centrally
mediated and is not due to peripheral failure of force generation. However, as yet, no agreed
explanation can be offered for the fatigue described in the chronic fatigue syndrome.
The role of peripheral inhibitory influences in the form of afferent reflexes altering
central activity have long been debated. This century has seen a plethora of information on
the alterations in motor control, particularly from the work of Marsden and colleagues (1971 ;
1983) who demonstrated the decline in motor unit discharge rates during contraction in
relation to slowing of relaxation (muscle wisdom), and Bigland-Ritchie and colleagues
(1986) who examined the origins of the reflex mechanisms involved. It is interesting to note
that the methods enjoyed now to determine the central contribution of fatigue such as
interpolation of stimuli on voluntary efforts (Merton, 1954; Belanger & McComas, 1981;
Bigland-Ritchie et aI., 1986) were, in fact, well established in the last century as reported by
workers such as Waller (1891) in his studies of central fatigue.
While metabolic factors involved in the reflex hypotheses of central fatigue appear
to receive much attention, a central metabolic cause has generally been excluded owing to
the unlikely situation of changes in brain metabolism such as acidosis except in pathological
conditions and that exercise lactacidosis has a negligible effect on brain function (Siesjo,
1982; see Newsholme & Blomstrand, Chapter 23). Nevertheless, the importance of the role
of the reticular formation in sleep has received attention more latterly with Newsholme and
colleagues (1987) suggesting a pivotal role for altered levels of the neurotransmitter
5-hydroxy-tryptamine and its close association with the plasma tryptophan:branched chain
amino acid ratio. It is possible that such a mechanism may be involved in exercise of long
duration, but whether such an explanation can be offered for fatigue in short held contractions
is doubtful.
486 R. H. T. Edwards et al.

FAILURE OF MUSCLE EXCITATION AS A CAUSE OF FATIGUE

The credit for detennining the motor nerve firing frequency is generally given to
Adrian and Bronk (1929). Historical researches would suggest that the phenomenon of
recording electrical potentials may have its origin far earlier with the existence of muscle
action potentials, negative variations, being first proven in 1843 and action currents recorded
from the ann of a man who contracted his muscle by using jars of liquid as electrodes by
(DuBois-Reymond in 1851, Fig. 1).
The discovery of muscle action potentials led to the exclusion of failure of the nerve
fiber in the genesis of fatigue as several workers in the latter part of the 19th century used
novel methods to detect action potentials and to demonstrate that nerve was indefinitely
excitable: Wedenski (1884, cited in Waller 1891) " ... used a telephone to detect the negative
variation offaradised nerve", while Waller (1891) employed a galvanometer to record true
negative variation and electrotonic currents.
The identification of transmission failure of excitation in myasthenia gravis perhaps
gave important clues as to the possible involvement of the neuromuscular junction in fatigue.
With developments in electromyograpy (Adrian & Bronk, 1929) applied to muscular
disorders together with an understanding of the chemical nature of the neuromuscular
junction, Lindsley (1935) concluded that the early fatigue seen in myasthenia gravis was
indeed due to " ... a blocking of the transmission of excitation at the 'myoneural junction'''.
Techniques of supramaximal electrical stimulation with recording of the action potential

Figure 1. DuBois-Reymond's demonstration of deflection of the magnetic needle by the will. " . .. the man
strongly contracts the muscle of one arm, the result is an immediate deflection of the multiplier, which indicates
the presence of a current ascending in the contracted arm from the hand to the shoulder. If the muscles of the
other arm are contracted, a deflection occurs in the opposite direction. We are, therefore, able by the mere
power of the will to generate an electric current and to set the magnetic needle in motion." (reprinted with
permission from Rosenthal, \883).
Historical Perspective 487

were developed further by Harvey and Masland (1941) as a diagnostic tool. Various programs
of repetitive nerve stimulation to permitted the identification of disorders of neuromuscular
activation forming the basis of the electrophysiological diagnosis of myasthenia gravis, and
later were refined with the introduction of single unit EMG methods.
There remained the question as to whether there was evidence of impaired neuromus-
cular transmission in any form of muscular activity in normal SUbjects. The finding that there
was close correlation between the force and smoothed rectified EMG suggested to Edwards
and Lippold (1956) and later Stephens and Taylor (1972) that in sustained isometric
contractions of the first dorsal interosseous muscle, the site of fatigue was indeed at the
neuromuscular junction. But earlier work by Merton (1954) which was supported later by
Bigland-Ritchie and colleagues (1983) systematically examined the question ofneuromus-
cular junction failure. By interpolating stimuli on a maximum voluntary contraction, these
investigations were unable to demonstrate any evidence of failure during at least 60 s of
sustained maximal isometric activity, clearly highlighting the involvement of processes that
lie beyond the neuromuscular junction.

BIOCHEMICAL BASIS FOR FATIGUE

The growth of ideas about the biochemical basis of muscular contraction is covered
in Dorothy Needham's classic Machina Carnis (Needham, 1971). Various clinical studies
can be perceived in retrospect to strongly support the hypothesis that muscle pain and fatigue
are closely related to failure of energy supply or undue accumulation of lactate. Direct
evidence to support these hypotheses dates back to 1904 when Weichardt suggested an
accumulation of a substance, which he called kenotoxin, was responsible for fatigue. Muller
(1935) also demonstrated a faster onset of fatigue with circulatory occlusion. The finding
that the disappearance of lactic acid showed a consistent correlation with recovery from
fatigue led Hill and coworkers (1924) to propose that lactic acid was the fatigue substance.
This had important implications in the future work on fatigue with many researchers seeking
to show the importance of lactic acid and H+ in the genesis of fatigue.
The description of the experimental procedure by Lewis (1942, p. 96) emphasizes
these ideas:
In the standard test, almost maximal voluntary gripping movements, which develop a tension of
20 to 281bs. weight, are made they are carried out rhythmically at the rate of usually one a second.
Such rhythmic movements can be undertaken painlessly for very many minutes, provided that the
circulation to the limb is free; but, ifthe circulation is stopped before exercise begins, pain is soon
felt. It begins after 24 to 45 sec. (or the same number of contractions) and quickly grows in intensity
until it renders the exercise so disagreeable that the exercise is stopped at a point between 60 and
80 sec. Using uniform conditions and adequate periods of intervening rest between tests, it is
surprising how constant is the time taken to reach the intolerable point in repeated tests on a given
individual. The pain is rather diffuse, once it has come, and it gathers steadily in intensity as the
test proceeds. It is easy to ascertain that the pain, though diffuse, appears largely in the region of
the muscles used. Thus, in the standard test, the pain is felt mainly over the flexor surface of the
forearm; some may be felt over the thenar eminence. The most convincing examples are those in
which small muscles and simple movements are employed, such as opposition of the thumb or
abduction of the little finger .... It is concluded that the stimulus responsible for pain when
muscular exercise is taken in the absence of blood supply arises directly or indirectly out of the
contraction process.

The practical application of this clinical physiological approach of Sir Thomas Lewis
came with the publication in 1951 of the description of the ischemic forearm lactate test.
This test was applied in a patient who complained of muscle fatigue and cramps and who
was unable to produce lactate during this type of exercise (McArdle, 1951). The implication
488 R. H. T. Edwards et al.

of this finding for the study of energy metabolism was to provide a paradigm disorder
consistent with isolated poisoned muscle models. Similarly, Lunsgaard (1930) demonstrated
that contraction occurred in frog muscle despite inhibiting glycolysis with iodoacetate. The
same observations are made in these patients in whom some exercise is possible, but the
premature fatigue seen cannot be a consequence of the accumulation of lactate since none
is produced.
Another group of individuals with a defect in metabolism are those with mito-
chondrial disorders (De Jesus, 1974, Morgan-Hughes et aI., 1977). Profound exercise
intolerance and marked fatigability with a tendency to develop lactic acidosis illustrate
the primary defect is in oxidative phosphorylation. Such individuals again belong to
Nature s experiments. Their counterpart, isolated cyanide poisoned preparations (Uittgau,
1965), where there is a reliance on anaerobic metabolism for supply of energy. Another
group of patients which have given important clues as to the metabolic dependence of
fatigue are those with hyper- and hypothyroidism. Hypothyroid patients are able to sustain
force for longer at a lower ATP cost than normal individuals or patients with hyperthy-
roidism (Wiles et aI., 1979). Though attributed largely to an increase in prevalence of
type I fatigue resistant fibers, the resistance to fatigue cannot be explained by this fact
alone in view of changes in relaxation characteristics before changes in fiber composition
with treatment.
Simultaneously and mostly in Scandinavia, there had progressed a major drafting
of the principal changes in the energy metabolism associated with fatigue in muscular
exercise. This had been possible only because of the introduction of the needle biopsy
technique (Bergstrom, 1962). Such studies identified the importance of high energy
phosphates in brief exercise (Hultman et aI., 1967) as well as muscle glycogen as a
determinant of endurance time in heavy dynamic exercise. The needle biopsy technique
again allowed direct observation of the extent of recruitment of motor units in particular
forms of dynamic exercise, from determination of the glycogen depletion patterns (Goll-
nick et aI., 1974), thus confirming the motor unit recruitment patterns according to the
size principle (Henneman, 1957), and fatigability of fiber types as demonstrated in animal
muscle (Burke et aI., 1973).
Recently, there has been much interest as well as rapid progress in our understanding
of muscle chemistry due to the introduction of magnetic resonance spectroscopy. The use of
this tool to study muscle energy metabolism in health and disease (Edwards et aI., 1982;
Kemp & Radda, 1994), and its applications to fatigue are discussed in Section 4, this volume.
This major advance in technology spurred once again an interest in analyzing fatigue in
patients with mitochondrial and glycolytic defects, elegantly demonstrating specific blocks
in metabolism, and is now finding applications in other muscle diseases. The true value of
magnetic resonance in the study of fatigue is only likely to be realized, however, when
combined with objective measurement of muscle function (Gibson & Edwards, 1993;
Vestergaard-Poulsen et aI., 1994).

EXCITATION-CONTRACTION COUPLING FAILURE CAUSING

FATIGUE

The demonstration in human muscle that the equivalent of the impairment of

excitation-contraction coupling could occur with ischemic isometric exercise (Edwards et
aI., 1977) raised yet another candidate for the site of fatigue. More dramatically, this
impairment can be induced by eccentric activity suggesting the involvement of mechanical
Historical Perspective 489

0.90

.~ 0.77
fII
a:
II 0.64
u
....
0
u-
N 0.51
:I:
0
LO
0.38
0
C\I

0.25

0 20 40 60 80 100

per (lot total PI

Figure 2. Voluntary, submaximal exercise involving repeated isometric contraction at 40% of maximum force
generation (6 s with 4 s rest) was performed by six subjects Electrically evoked frequency-force relations are
plotted against energy status determined by magnetic resonance spectroscopy (data from each subject are
shown with a different symbol). Note that while there is a considerable variation among the subjects in their
use of muscle per, there is consistent reduction in the ratio of forces achieved with 20 and 50 Hz electrical
stimulation. These data indicate that impairment in electromechanical coupling was more important than
energy depletion as a cause of fatigue in this form of exercise. (Gibson et aI., 1993b)

damage (Newham et aI., 1983). However, fatigue during isometric activity also occurs more
rapidly in patients with mitochondrial disorders indicating a metabolic cause (Wiles, 1981).
The interrelations between energy utilization and excitation/excitation-contraction
failure were first explored in an intuitive manner at the conclusion of the milestone Ciba
Foundation Symposium on Human Muscle Fatigue: Physiological Mechanisms. Merton
proposed "that muscle fatigue has everything to do with electricity and nothing to do with
chemistry". That debate continues today helped by rapid progress in magnetic resonance
spectroscopy (e.g., Miller et aI., 1987; see Miller, Chapter 14; Bertocci, Chapter 15). There was
further debate in the Chairman's concluding section. These discussions provided the foundation
for the subsequent Catastrophe Theory offatigue (Edwards, 1983). At the Amsterdam meeting
on muscle fatigue (1992), more evidence was forthcoming as to the greater importance of the
impairment of electromechanical coupling compared with energy limitations as determining
the onset of fatigue. Fig. 2 illustrates in normal quadriceps muscle for six subjects the greater
relative importance of these processes compared to energy lack. In this study, there were
electrophysiological assessments and spectroscopic measures of metabolism in a whole body
MR system during muscle fatigue resulting from prolonged submaximal activity. These results
contrast with those of VIII lIestad and colleagues (1988) who, using serial biopsy measures of
energy metabolism, showed little change in metabolism until exhaustion.

IMPORTANCE OF RECOVERY PROCESSES IN THE ANALYSIS

OF FATIGUE

There has been much interest (and there still is today) in the time course of recovery
processes which are apparently associated with fatigue. In 1883, Foster wrote
Absolute (temporary) exhaustion of the muscles, so that the strongest stimuli produce no contrac-
tion, may be produced even within the body by artificial stimulation; recovery takes place on rest.
490 R. H. T. Edwards et al.

Out of the body absolute exhaustion takes place readily. Here also recovery may take place.
Whether in any given case it does occur or not, is determined by the amount of contraction causing
the exhaustion, and by the previous condition of the muscle. In all cases recovery is hastened by
renewal (natural or artificial) of the blood-stream. The more rapidly the contractions follow each
other, the less the interval between any two contractions, the more rapid the exhaustion.
This quotation highlights an early understanding of the importance of blood borne
processes in recovery, eventually leading to a plethora of studies which examine the role of
metabolism in fatigue. The major advances, however, have come about not through identi-
fying those factors that correlate with fatigue, but rather from the divergence of processes
during recovery from fatigue. Hence an early paradoxical finding by Davis and Davis (1932)
on cat soleus was the observation that fatigue during high frequency stimulation could be
reversed on acutely reducing the frequency, where force is seen to increase dramatically.
This has subsequently been confirmed for human muscle by Bigland-Ritchie and colleagues
(1979) who were also able to demonstrate a rapid concomitant recovery in the action
potential. Clearly metabolic factors are not involved owing to the marked differences in the
time course of recovery of excitation and metabolism, the reduction in frequency permitting
time for the rapid redistribution of ionic fluxes.
Further recent insight into the processes of failure of excitation in fatigue have come
from ischemic studies in McArdle's patients. A decline in the action potential amplitude is
observed not to recover following activity when ischemia is maintained (Gibson & Edwards,
1988). The converse is seen in normal subjects. Moreover, the action potential does not
appear to increase in duration, a phenomenon associated with a slowing of muscle conduction
velocity, often attributed to accumulation of potassium. Such observations suggest the
involvement of an energy-dependent system, and supports the notion that dropping out of
fibers may occur during fatigue. These results give an opportunity to examine alternative
concepts of fatigue. Selective dropping out of fibers may also explain the phenomenon
described above where type II fibers may have become fatigued more readily. Such a scenario
is illustrated in Fig. 3 where a fiber population paradigm of fatigue is presented for an
admixture of type I and II fibers with differing fatigue characteristics alluded to earlier. High
frequency stimulation results in a decline in force, but as stimulation frequency is reduced,
force recovers. The insets illustrate the relative changes in two populations of fibers where
one popUlation fatigues more rapidly than the second (fibers drop out). Reduction of
stimulation frequency allows rapid recovery of excitatory processes with subsequent recov-
ery of both fiber popUlations. This is, of course, an oversimplification; the gradual decline

100Hz

Figure 3. Hypothetical representation of

differential fatigability of muscle fiber pop-
Force/Fiber
ulations (insets) as a possible explanation of
the paradoxical increase in force as stimu-
lation frequency is reduced (experimental
TIME example based on Jones, 1981).
Historical Perspective 491

in force is likely to represent a statistical statement of events for many fibers. Thus fatigue
in a whole muscle (often treated as a single cell model) in which fibers pull in parallel could
be considered as 1) a shift in active fiber popUlation (derecruitment); or 2) a reduction in the
fiber force-generating performance.

A COMBINED CONCEPTUAL MODEL FOR CLASSIFYING

SPECIFIC FORMS OF FATIGUE IN DIFFERENT MUSCLE OR
ACTIVITY CIRCUMSTANCES

Fig. 4 summarizes the principal mechanisms resulting from the milestone advances
that have contributed to our understanding of fatigue which may be applied to all forms of
activity in health and disease and illustrates the contribution of the many elements offatigue
often studied in isolation.
The model incorporates the principal components:
• Parallel Fatigue i.e., loss of overall force generation due to loss of function of
contractile elements (i.e., individual fibers) pulling in parallel
• Contractile Fatigue comprises interrelations between energy depletion and elec-
tromechanical coupling processes.

Parallel fatigue

Mechanism: Mechanism:
(motor unit activation)
Metabolic
Derecuitment e Ischaemic exercise
eChronic fatigue syndrome eMcArdles's disease
ePartial denervation?
ePartial curarisation? Neuromuscular transmission
eMyasthenia gravis
eHyperthyroidism

Sarcolemmal excitation
eMyotonia
Central fatigue Peripheral fatigue
Mechanism: Mechanism:

'Natural Wisdom' Metabolic

elschaemic exercise
eMcArdles's disease
e Mitochondrial abnormalities

Neuromuscular transmission
e Eccentric exercise
eSubmaximal exercise
eDantrolene sodium

Contractile fatigue

Figure 4. A four quadrant diagram to help classify type of fatigue in different types of exercise or clinical
condition.
492 R. H. T. Edwards et al.

These have been combined to form a four quadrant diagram (Fig. 4) to illustrate how
these components may contribute to:
• Central Fatigue resulting in parallel fatigue, due to derecruitment of motor units
- (e.g., reduced motivation for continuing voluntary muscular activity)
• Central Fatigue with impaired force generation of population of muscle fibers (this
represents the natural wisdom in which there is apparent matching of the firing
frequency to match alterations in the contractile properties of fatiguing muscle
fibers.
• Peripheral Fatigue (Parallel) due to selective fatigue of particular fiber popula-
tions either because of specific metabolic characteristics or because of preexisting
alteration in the distribution of fiber types e.g., Type II fiber dominance and,
• Peripheral Fatigue (Contractile) due to impaired intrinsic force generating capac-
ity of individual muscle fibers owing to impaired excitation-contraction coupling
defect or energy depletion affecting all fibers.

CONCLUSIONS

This cursory review of the history of pathophysiological studies in fatigue shows that
several of the fatigue phenomena that are studied in isolated experimental preparations can
be studied in human muscle. All these studies have given clues as to the sites and mechanisms
of fatigue in diseased muscle. Concepts clearly thought out by the latter part of the last
century required new tools to verify the ideas. Findings in neuromyopathies such as
myasthenia gravis, Nature s experiments (single enzyme defects or specific abnormalities of
function) or altered central drive give the opportunity to understand the factors controlling
neuronal firing frequency and to discover ways to optimize force generation of fatiguing
muscle fibers to improve overall performance.

ACKNOWLEDGMENT

The authors gratefully acknowledge laboratory support from the Muscular Dystrophy
Group of Great Britain and Northern Ireland, Association Franc;:aise contre les Myopathies,
and ICI Pharmaceuticals. Attendance of [Link].E. and HG. at the 1994 Bigland-Ritchie
conference was supported, in part, by the Wellcome Trust, the American Physical Therapy
Association (Research and Analysis Division for R.H T.E.; Section on Research for H G.),
and the Muscular Dystrophy Association (USA),

REFERENCES
Adrian ED & Bronk DW (1929). The discharge of impulses in motor nerve fibres. Journal of Physiology
(London) 67,119-151.
Althaus J (1873). Theoretical and practical paralysis, neuralgia and other diseases In: A Treatise on Medical
Electricity, 3rd Ed. London: Longmans, Green, and Co.
Bainbridge FA (1931). The Physiology ofMuscular Exercise, p. 235. London: Longmans, Green & Co.
Belanger AY & McComas J (1981). Extent of motor unit activation during effort. Journal ofApplied Physiology
51, 1131-1135.
Bergstrom J (1962). Muscle electrolytes in man. Scandinavian Journal of Clinical Laboratory Investigation
14 (suppl 68), 9-88.
Historical Perspective 493

Bigland-Ritchie B, Dawson NH, Johansson RS & Lippold OCJ (1986). Reflex origin for the slowing of
motoneurone firing rates in fatigue in human voluntary contractions. Journal ofPhysiology (London)
379,451-459.
Bigland-Ritchie B, Johansson R, Lippold OCJ & Woods JJ (1983). Contractile Speed and EMG changes during
fatigue of sustained maximal voluntary contractions. Journal ofNeurophysiology 50, 313-324.
Bigland-Ritchie B, Jones DA& Woods JJ (1979). Excitation frequency and muscle fatigue: electrical responses
during human voluntary and stimulated contractions. Experimental Neurology 64, 414-427.
Burke RE, Levine DN, Tsairis P & Zajac FE (1973). Physiological types and histochemical promes in motor
units of the cat gastrocnemius. Journal ofPhysiology (London) 234, 723-748.
Da Costa JKM (1871). On irritable heart: a clinical study of a form of functional cardiac disorder and its
consequences. American Journal ofMedical Science 61, 17-52.
Davis H & Davis PA (1932). Fatigue in skeletal muscle in relation to the frequency of stimulation. American
Journal of Physiology (London) 101,339-356.
De Jesus PV (1974). Neuromuscular physiology in Luft's syndrome. Electromyography and Clinical Neuro-
physiology 14, 17-27.
Drummond M & Drummond WB (1915) (English translation of Mosso A book). Fatigue. pp. 81-102. London:
George Allen & Unwin Ltd.
Duchenne L (1872). L'electrisation localisee et de son application a la pathologie et a la therapeutique par
courants induits et par courants galvaniques interrompus et continus. Paris: Bailliere, JB et ms.
Edwards RHT (1983). Biochemical basis of fatigue in exercise: Catastrophy theory of muscular fatigue In:
Knuttgen HG, Vogel JA, Poortmans J (eds.), Biochemistry ofExercise, pp. 3-28. Champaign: Human
Kinetics.
Edwards RHT, Dawson MJ, Wilkie DR & Gordon RE (1982). Clinical use of nuclear magnetic resonance in
the investigation of myopathy. Lancet 1(8274}, 725-731.
Edwards RHT, Hill DK, Jones DA & Merton PA (1977). Fatigue of long duration in human skeletal muscle
.after exercise. Journal ofPhysiology (London) 272, 769-778.
Edwards RG & Lippold OCJ (1956). The relation between force an integrated electrical activity in fatigued
muscle. Journal ofPhysiology (London) 132,677-681.
Foster M (1883). A Textbook of Physiology. 4th Ed, pp. 96-97. London: Macmillan and Co.
Gibson H & Edwards RHT (1988). No ischaemic recovery of the evoked compound muscle action potential
in McArdle's disease. Clinical Science 75, 40P.
Gibson H & Edwards RHT (1993). Inter-relation of electro-mechanical coupling and chemistry in the study
of fatigue by 31p_NMR spectroscopy In: Sargeant AJ, Kernell D (eds.), Neuromuscular Fatigue, pp.
30-31. Amsterdam: Royal Netherlands Academy of Arts and Sciences.
Gibson H, Carroll N, Clague IE & Edwards RHT (1993a). Exercise performance and fatiguability in patients
with the chronic fatigue syndrome. Journal ofNeurology, Neurosurgery, and Psychiatry 56,993-998.
Gibson H, Saugen E, Martin PA, Vollestad NK & Edwards RHT (l993b). Individual pathways for ele-
tromechanical-coupling and energy utilization in submaximal exercise. International Union of
Physiological Sciences, Glasgow, August 1-6,216.
Gollnick PD, Karlsson J, Piehl K & Saltin B (1974). Selective glycogen depletion in skeletal muscle fibres in
man following sustained contractions. Journal of Physiology (London) 241, 59-67.
Harvey AM & Masland RL (1941). A method for the study of neuromuscular transmission in human subjects.
Bulletin of the Hopkins Hospital 68, 81-93.
Henneman E (1957). Relation between size of neurons and their susceptibility to discharge. Science 126,
1345-1347.
Hill AV, Long CNH & Lupton H (1924). Muscular exercise lactic acid, and the supply and utilization of oxygen,
parts I-III. Proceedings of the Royal Society ofLondon 96, 438-479.
Hultman E, Bergstrom J & McLennan AN (1967). Breakdown and resynthesis of adenosine triphosphate in
connection with muscular work in man. Scandinavian Journal of Clinical Laboratory Investigation
19,56-66.
Jolly F (1895). Uber myasthenia gravis pseudoparalytica. Berliner Kliniscshe Wochenschrift 32, 1-7.
Jones DA (1981). Muscle fatigue due to changes beyond the neuromuscular junction In: Porter R, Whelan J
(eds.), Human Muscle Fatigue: Physiological Mechanisms. Ciba Foundation Symposium 82, pp.
178-212. London: Pitman Medical.
Kaplan EB (1959) (English translation ofDuchenne GB book). PhYSiology ofMotion Demonstrated by Means
of Electrical Stimulation and Clinical Observation and Applied to the Study of Paralysis and
Deformities. Philadelphia: WB Saunders.
 494 R. H. T. Edwards et al.

Kemp GJ & Radda GK (1994). Quantitative interpretation of bioenergetic data from 31p and IH magnetic
resonance spectroscopic studies of skeletal muscle: an analytical review. Magnetic Resonance
Quarterly 10, 43-63.
Kessel L & Hyman HT (1923). The clinical manifestations of disturbances of the involuntary nervous system
(autonomic imbalance). American Journal ofMedical Sciences 165, 513.
Lewis T (1918). The Soldier's Heart and The Effort Syndrome. London: Shaw & Sons.
Lewis T (1942). Pain. New York: The Macmillan Company.
Lindsley DB (1935). Myographic and electromyographic studies of myasthenia gravis. Brain 58, 470-482.
Liittgau HC (1965). The effect of metabolic inhibitors on the fatigue of the action potential in single muscle
fibres. Journal ofPhysiology (London) 178, 45-67.
Lundsgaard E (1930). Uber die Einwirkung der Monoiodoessigsiiure aud den spaltungs - und Oxydationsstof-
fwesche!. Biochemishe Zeitschrijt 220,8-12.
Marsden CD, Meadows JC & Merton PA (1971). Isolated single motor units in human muscle and their rate
of discharge during maximal voluntary effort. Journal ofPhysiology (London) 217, 12-13P.
Marsden CD, Meadows JC & Merton PAC 1983). "Muscular wisdom" that minimizes fatigue during prolonged effort
in man: peak rates of motomeuron discharge and slowing of discharge during fatigue In: Desmedt IE (ed.),
Motor Control Mechanisms in Health and Disease, pp. 169-211. New York: Raven Press.
McArdle B (1951). Myopathy due to a defect in muscle glycogen breakdown. Clinical Science 10,13-33.
Merton PA (1954). Voluntary strength and fatigue. Journal of Physiology (London) 123, 553-564.
Merton PA (1956). Problems of muscular fatigue. British Medical Bulletin 12, 219-239.
Miller RG, Giannini D, Milner-Brown HS, Layzer RB, Koretsky AP, Hooper D & Weiner MW (1987). Effects
of fatiguing exercise on high-energy phosphates, force, and EMG: evidence for three phases of
recovery. Muscle & Nerve 10, 810-821.
Morgan-Hughes JA, Darveniza P, Kahn SN, Landon DN, Sherratt RM, Land 1M & Clark JP (1977). A mitochondrial
myopathy characterised by a deficiency in reducible cytochrome b. Brain 100, 617-640.
Muller EA (1935). Die Erholung nach statischer Haltearbeit. Arlbeitsphysiologie 8, 72-75.
Needham DM (ed) (1971). Machina Carnis. Cambridge: Cambridge University Press.
Newham DJ, Mills KR, Quigley BM & Edwards RHT (1983). Pain and fatigue after concentric and eccentric
muscle contractions. Clinical Science 64, 55-62.
Newsholme EA, Acworth IN & Blomstrand E (1987). Amino acids, brain neurotransmitters and a functional
link between muscle and brain that is important in sustained exercise In: Benzi G (ed.), Advances in
Myochemistry, pp. 127-133. London: John Libbey, Ltd.
Oppenheimer B, Levine SA, Moneson RA, Rothechild MA, St. Lawrence W & Wilson FN (1918). Appendix
of illustrative cases of neurocirculatory asthenia. Military Surgery 42, 409-420.
Rasch PJ & Burke RK (eds.) (1967). The history of kinesiology In: KineSiology and Applied Anatomy. The
Science ofHuman Movement, 3rd Ed., pp. 1-17. Philadelphia: Lea & Febiger.
Rosenthal I (ed.) (1883). General Physiology ofMuscles and Nerves. London: Kegan Paul, Trench & Co.
Shorter E (1992). From Paralysis to Fatigue. New York: Free Press.
Siesjo BK (1982). Lactic acidosis in the brain: occurrence, triggering mechanisms and pathophysiological
importance In: Porter R, Lawrenson G (eds.), Metabolic Acidosis, Ciba Foundation Symposium 87,
pp. 77-88. Edinburgh: Churchill Livingstone.
Stephens JA & Taylor A (1972). Fatigue of maintained voluntary muscle contraction in man. Journal of
PhYSiology (London) 220, 1-18.
Stokes M, Cooper RG & Edwards RHT (1988). Normal muscle strength and fatigability in patients with 'effort
syndromes'. British Medical Journal 297, 1014-1017.
Vestergaard-Poulsen P, Thomsen C, Sinkjaer T & Henriksen 0 (1994). Simultaneous 31p NMR spectroscopy
and EMG in exercising and recovering human skeletal muscle: technical aspects. Magnetic Reso-
nance in Medicine 31, 93-102.
V 0llestad NK, Sejersted OM, Bahr R & Bigland-Ritchie B (1988). Motor drive and metabolic responses during
repeated sub-maximal contractions in man. Journal ofApplied PhYSiology 63, 1-7.
Waller AD (1891). The sense of effort: an objective study. Brain 14, 179-249.
Weichardt W (1904). Uber das ermudungtoxin und-antitoxin. Muenchener Medizinsche Wochenschrijt 51,
2121-2126.
Wiles CM, Jones DA & Edwards RHT (1981). Fatigue in human metabolic myopathy In: Porter R, Whelan J
(eds.), Human Muscle Fatigue: PhYSiological Mechanisms, Ciba Foundation symposium 82, pp.
264-282. London: Pitman Medica!.
Wiles CM, Young A, Jones DA & Edwards RHT (1979). Muscle relaxation rate, fibre-type composition and
energy turnover in hyper- and hypo-thyroid patients. Clinical Science 57, 375-384.
Wood P (1941). Da Costa's syndrome (or effort syndrome). British Medical Journal i, 767- 770.
 36

FATIGUE BROUGHT ON BY MALFUNCTION

OFTHECENTRALANDPEIDPHERAL
NERVOUS SYSTEMS

A. J. McComas I ,R. G. Miller,2 and S. C. Gandevia3

IDepartment of Biomedical Sciences

McMaster University, Hamilton, Canada
2Department of Neurology, California Pacific Medical Center
San Francisco, California
3 Prince of Wales Medical Research Institute
Sydney, Australia

ABSTRACT

Increased fatigability necessarily occurs in every patient with muscle weakness, regardless
of whether the latter is due to a central or peripheral neurological disorder. The tendency for
disuse to increase fatigability, as a secondary phenomenon, must also be considered; disuse
affects both motoneuron recruitment and the biochemical and physiological properties of the
muscle fibers. In recent studies impaired recruitment has been observed in postpolio patients,
while patients with multiple sclerosis or spinal cord injury have shown, in addition, altered
neuromuscular function. Findings are also presented in ALS and the chronic fatigue syn-
drome. In general, the most dramatic increases in fatigability take place in disorders of the
peripheral nervous system and almost any cell component can be incriminated. There is a
need to study fatigability systematically in neurology and rehabilitation.

INTRODUCTION

Symptomatology

Patients with neurological disorders often have symptoms which suggest that they
suffer from undue fatigability. Commonly heard complaints include the following: "My legs
get tired", "I cannot walk as far as I used to", "My arms start to feel heavy", "My eyelids
droop as the day goes on", "I start to see double when I am tired", and others. The choice of
verb in these statements is significant, in that it leaves open whether or not the affected
muscles are initially weak; instead, there is an insistence that some change takes place in the
course of an activity which would normally be accomplished with relative ease. Indeed, the
ability to distinguish between absolute weakness and fatigability may be crucial in leading

495
496 A. J. McComas et al.

towards the correct diagnosis. An example is the fatigability of the neuromuscular disorder,
myasthenia gravis (see below). Mildly affected patients will be quite clear that they have no
problems in the mornings, but that the drooping of the eyelids, the tendency to see double
and the tiring of the limbs declare themselves later in the day and then progress. Such
information strongly suggests that the available muscle mass is normal.
Ifmuscles are weak to begin with, however, fatigability will be increased if the same
demanding tasks are attempted as before the illness. This conclusion depends on the fact that
there is normally an inverse relationship between the percentage of maximal force attempted
and the time to a set fatigue level (e.g., Bigland-Ritchie et aI., 1986; Dolmage & Cafarelli,
1991). Suppose, then, that a patient with a neuromuscular disorder can only develop half the
initial force that he/she was previously capable of. Clearly the time to fatigue will be less
for a MVC (maximal voluntary contraction) in these circumstances than it would have been
in the premorbid state, or in a control subject (Fig. lA). On the other hand, if the diminished
MVC is taken as the reference value, there is no a priori reason for fatigue to be faster for
a MVC than in the premorbid state. Since the force that can be developed in a MVC varies
markedly from one subject to another, depending on muscle bulk, training and effective lever
arm, performance in a patient should be assessed against previous abilities, rather than in
absolute terms. Thus a steelworker may have genuine fatigability and still be stronger than
the examining physician.
Furthermore, regardless of whether or not a set physical task can be accomplished
normally, patients with neuromuscular disorders commonly have more protracted recoveries
with greater discomfort than before their illnesses.
Finally, although this review is concerned mainly with the changes in muscle
physiology which accompany fatigue, the problem of mental fatigue must not be overlooked.
Some patients appear to have low mental energy and would previously have been labeled as
neurasthenic. Alternatively, or in addition, patients may find that the mental cost of a physical
activity becomes disproportionately large, such that they feel drained afterwards. Such
considerations apply especially to the chronic fatigue syndrome (see below).

Central and Peripheral Fatigue in Neurological Disorders

It is customary to divide fatigue into central and peripheral components (see elsewhere
in this volume); in neurology such a division is analogous to the simple classification of motor
disorders into those of the upper motoneuron (UMN) and lower motoneuron (LMN) types. A
UMN disorder is one in which a CNS lesion may cause diminished effort and impaired
organization or transmission of the motor command to motoneurons in the brainstem and spinal
cord. A lower motoneuron lesion, in contrast, is one which involves the motoneuron, its axon,
the neuromuscular junction or the muscle fiber. The correlation between UMN and LMN
lesions, on the one hand, and central and peripheral fatigue on the other, is not exact, however,
for dysfunction of the motoneuron soma would be seen as part of a LMN lesion but also as a
cause of central fatigue. In contrast, dysfunction of the motor axon, for example at the branch
points or at the nerve terminal, would be regarded as contributing to peripheral fatigue. This
distinction is largely artificial, however, since dying-back changes in the nerve endings may be
the first manifestation of a cell body disorder, presumably because of impaired axoplasmic
transport and inadequate trophic support.
On general grounds, it would be expected that the LMN lesion would cause peripheral
fatigue and that the UMN lesion would result in central fatigue. Closer inspection, however,
suggests that such correlations may be too simple (Fig. IE). Thus, weakness from a LMN
lesion will require greater than normal effort to accomplish a set task and this would be
expected to induce central fatigue. Again, weakness from either UMN or LMN lesions may
be complicated by the effects of disuse, as the patient tries to conserve strength by resting
Fatigue and Malfunction ofthe Nervous Systems 497

Maximum
voluntary
force (N)

Healthy subject

(A) Patient

_________________________________________________ Failure level

..... "
Time

(8)

PERIPHERAL
FATIGUE

Figure 1. A, if a patient has become weak and attempts the same task as a healthy subject or as previously
undertaken, fatigue must occur sooner_ D, relationships between central and peripheral fatigue, and upper and
lower motoneuron lesions.

or no longer attempts tasks that have become particularly tiring; the most extreme disuse
effects would involve those patients who are largely confined to bed or to wheelchairs_ On
the one hand, disuse will cause or accentuate fatigue by reducing the ability of the motor
centers to recruit motoneurons maximally. Perhaps the most unequivocal demonstration of
this effect, unaccompanied by disease, was the finding of a reduction in the density of the
EMG interference pattern when patients with previous bone fractures attempted to make
MVCs in limbs that had been recently freed from immobilizing casts (Fuglsang-Frederiksen
& Scheel, 1978; Duchateau & Hainaut, 1987). The transient nature of the EMG changes,
with resumption of normality in a few days, indicated that impaired recruitment was unlikely
to be associated with overt degeneration of the descending fibers, but the possibility of
microscopic changes in the nerve terminals remains_ On the other hand, disuse of UMN or
LMN origin would also be expected to produce excessive fatigability in muscle fibers on
the basis of animal experiments in which surgically denervated muscles were subjected to
different regimens of direct electrical stimulation (Westgaard & Lemo, 1988). Although
increased fatigability after immobilization has not been observed in some animal studies
(Maier et aI., 1976; Robinson et aI., 1991; see, however, Petit & Gioux, 1993), the findings
in humans are consistent with the results ofWestgaard and Lemo. Thus, tetanic stimulation
498 A. J. McComas et al.

at 30 Hz produced faster fatigue in the adductor pollicis muscles of previously immobilized

arms than in the contralateral control muscles (Duchateau & Hainaut, 1987). Similar results
have recently been noted by Miller and coworkers (1990) in leg muscles of patients with
spastic paraparesis of varying etiologies (Gordon, Chapter 32).

INVESTIGATION OF FATIGABILITY IN PATIENTS

Force Measurements
It is true to say that the rate of fatigue, and the possible peripheral and central
components, are almost never investigated in the neuromuscular clinic. The reason for this
unfortunate state of affairs is that very few clinics, even in university centers, have incorpo-
rated quantitative strength measurements as part of the diagnostic armamentarium, despite
the commercial availability of hand-held myometers with chart recorders. Consequently the
study of muscle fatigue has been left to neurological research laboratories; alternatively gross
performance may be tested on treadmills or cycle ergometers in exercise laboratories.
Whenever possible, it is important that stimulated force also be recorded. Although pro-
longed repetitive maximal stimulation of a muscle through its peripheral nerve is too
uncomfortable for most patients, much useful information can be obtained from twitches.
Not only is the twitch tension itself a useful index of muscle function in the resting state, but
the interpolation of one or more stimuli can be used to detect the presence of central fatigue
during a MVC (see Gandevia et aI., Chapter 20). The presence of a substantial interpolated
twitch, early in the MVC, is not conclusive evidence of muscle fatigue, since such results
also occur in patients who, for one reason or another, choose not to make a full effort
(McComas et aI., 1983). The muscles in which contractions have been evoked by stimulation
include the ankle dorsiflexors and plantarflexors, adductor pollicis, quadriceps, biceps
brachii and diaphragm.

Subsequent Strategies
Further investigation will depend on whether or not a central or peripheral etiology
is suspected. In the case of a central disorder, MR imaging should reveal a structural lesion,
with electrophysiology playing a useful adjunct role. For example, cortical evoked potentials
tend to be delayed and prolonged in multiple sclerosis, while transcranial magnetic stimu-
lation of each motor cortex, in tum, will test the integrity of the corticospinal pathways. In
a peripheral disorder, further examination may require some or all of the following proce-
dures, together with muscle enzymes, and possibly blood lactate levels after ischemic
forearm exercise.

Electromyography
Any patients suspected of having increased fatigue should undergo needle EMG and
nerve conduction studies in an attempt to reveal an underlying neuromuscular disorder. If
impaired neuromuscular transmission is suspected, repetitive nerve stimulation and single
fiber electromyography (SFEMG) should be performed (see Trontelj & StAlberg, Chapter 7).

Muscle Biopsy
Muscle biopsy, preferably by needle, should be undertaken for diagnosis in any
patient suspected of harboring a neuromuscular disorder. Leaving aside the usefulness of
Fatigue and Malfunction of the Nervous Systems 499

assisting in the diagnosis of denervation or myopathy, a biopsy may suggest a more specific
fatigue-inducing pathology as in myophosphorylase deficiency, camitine deficiency or a
mitochondrial disorder (see below). Detailed testing for an enzyme disorder is best done by
one of the laboratories offering this type of service.

NMR Spectroscopy
This powerful non-invasive technique enables changes in muscle ATP, creatine
phosphate and H+ to be followed during voluntary or stimulated contractions, and has also
been used to explore the cellular nature of fatigue in healthy subjects (Miller et al., 1987;
see Miller, Chapter 14; Bertocci, Chapter 15).

FATIGUE STUDIES IN SELECTED CENTRAL NERVOUS SYSTEM

DISORDERS

Postpolio Patients
A proportion of patients who have had poliomyelitis many years previously appear
to make full recoveries and, indeed, have normal strength values and motor unit estimates
when tested quantitatively. Others, more severely affected by the initial illness, are left with
weakness and undue fatigability. A substantial fraction of the latter patients will complain
that their muscle deficits are becoming worse, and are associated with further muscle wasting
and with soreness after exercise. It is only this last group who should be considered to have
the postpolio syndrome. Two of the authors have made detailed studies of muscle perform-
ance in postpolio patients with varying degrees of clinical involvement. In both instances
endurance was tested using voluntary contractions which were 30% of maximal and which
lasted for 6 s out of every 10 s; the exercise was maintained for 45 min in the Sydney patients
(Allen et al., 1994) and for 15 min, with an additional 10 min at 50% of maximal force, in
the San Francisco patients (Sharma et al., 1994). The Sydney study was performed on the
elbow flexors, while the ankle dorsiflexors (especially tibialis anterior) were investigated in
the San Francisco patients.
In both studies MVCs showed greater rates of decline in the postpolio patients than
in controls. In the Sydney patients an important contributing factor was impaired voluntary
activation of motoneurons, as tested by a double-pulse technique. Thus, after 45 min of
fatiguing exercise, activation was 90.4 ± 9.5% (mean ± SD) in controls and 80.0 ± 17.7% in
patients (Fig. 2). In an extension of the Sydney series (now greater than 100 patients),
approximately one quarter of patients have been found to have impaired activation of
unfatigued muscles. In keeping with earlier remarks the deficit in activation may be ascribed
to the effects of disuse. Although no sign of impaired voluntary activation was found in the
San Francisco study, this may have been due to the smaller sample size and the different
technique employed for assessment (tetanic force instead of interpolated twitch).
Both the San Francisco and Sydney studies have been important in demonstrating
significantly greater relative decreases in stimulated force in postpolio patients than in
controls. In addition, the tetanic responses of the San Francisco patients were found to have
significantly slower rates of tension development and relaxation. These findings, presumably
due to differences in intracellular Ca2+ handling, raise the possibility that there may have
been increased incidences of type I muscle fibers in the postpolio patients. However, had
this been the case, fatigability to electrical stimulation should have been smaller, rather than
greater.
500 A. J. McComas et aI.

100

9
90
f T f f f
j
Voluntary

f
activation 80
(%)

70

o control
60 • polio

50
o 5 10 15 20 25 30 35 40 45
Time (min)
Figure 2. Progressive failure of voluntary activation in postpolio patients attempting repetitive elbow flexion
against resistance, compared with controls. Values expressed as means ± SDs (reproduced from Allen et aI.,
1994, by permission of Oxford University Press).

Although there was no evidence of abnormal M-wave decrement to repetitive

stimulation in the San Francisco patients, signs of synaptic dysfunction have been noted by
others (Trojan et aI., 1993). A feature of the San Francisco study was the application ofNMR
spectroscopy in some of the patients during and after the fatiguing exercise (Miller et aI.,
1988). No significant differences between postpolio patients and controls were detected in
terms of phosphocreatine (Per), inorganic phosphate (Pi) and intracellular pH. Interestingly,
in both subject groups there was an almost complete recovery of metabolites within 5 min
of stopping the exercise, whereas tetanic force was still considerably depressed after 15 min,
especially in the patients. Such observations are, of course, crucial for an understanding of
the cellular mechanisms responsible for prolonged low-frequency fatigue (Moussavi et aI.,
1989).

Amyotrophic Lateral Sclerosis Patients

Patients with amyotrophic lateral sclerosis (ALS) also have symptoms of increased
fatigue, although the functional limitations imposed by the progressive weakness often
dominate the clinical picture. In ALS, there are lesions involving both the upper motoneuron
(corticospinal tract pathways) and the lower motoneuron (anterior hom cells). Thus, fatigue
in these patients is even more complicated than postpolio syndrome patients in whom lower
motoneuron disturbances predominate (see, however, Bruno et aI., 1991). Fatigue in patients
with ALS should involve both central and peripheral components because of the nature of
the disease process. Few studies have examined the mechanisms of fatigue in patients with
ALS. 31p nuclear magnetic resonance spectroscopic studies of resting muscles in patients
with ALS or peripheral neuropathy demonstrated a higher inorganic phosphate/phosphate
ratio and higher intracellular pH compared to controls (Zochodne et aI., 1988; Argov & Bank,
1991). The only metabolic study that has been performed during exercise in patients with
ALS utilized bicycle ergometry and demonstrated a reduced maximum oxygen consumption
and reduced work capacity (Sanjak et aI., 1987). Moreover, increased muscular fatigability
was found in the shoulder abductor muscles of ALS patients compared to controls, but no
Fatigue and Malfunction of the Nervous Systems 501

measures of metabolism or force evoked by electrical stimulation were made (Nicklin et aI.,
1987).
Ten San Francisco patients with ALS have now been examined with a protocol similar
to that employed for postpolio patients (see above); all the patients had moderate weakness
of ankle dorsiflexors, with force values 25-30% of control. As with the postpolio patients,
the rate of fatigue was higher in ALS patients than in controls, and this was true both for
MVCs and for tetanic stimulation. The ALS results resembled the postpolio ones in another
respect, in that the rates of rise and relaxation of tetanic force were reduced, compared with
those of control subjects. Again, no significant differences from controls were found in 31p
metabolites or in pH, using NMR spectroscopy; this was true before, during and after
fatiguing exercise.
Surprisingly, in view of the pathophysiology of ALS, little evidence of impaired
motoneuron activation in fatigue was found when MVCs and tetanic forces were compared.
However, the NMR spectroscopy results suggested that there may have been some impair-
ment, since an exercise-induced increase in muscle water could be demonstrated in controls
but not in patients. Although the M-wave was maintained during tetanic stimulation, other
investigators have noted decrements in a proportion of ALS patients (e.g., Denys & Norris,
1979); further declines in amplitude can sometimes be seen in the potentials of single motor
units during voluntary contractions (McComas, unpublished observations).

Multiple Sclerosis Patients

Multiple sclerosis (MS) is a chronic neurologic disorder with demyelination often
involving the long motor and sensory pathways of the central nervous system. Patients with
MS who have motor control disturbances have both reduced firing rate modulation and
impaired recruitment of motor units (Rice et aI., 1992). In a recent San Francisco study it
was hypothesized that fatigue, which is the most common complaint in patients with MS,
was primarily central, partly as a result of the effort involved in activating and recruiting
motor units via dysfunctional corticospinal pathways and partly because of the constitutional
impact of the disease. The experimental approach was therefore to bypass the central nervous
system by employing tetanic, rather than voluntary, contractions. The ankle dorsi flexors were
the muscles chosen and the stimulation protocol was 50 Hz for 240 ms every 3 s, continued
for 9 min; 29 patients with MS were examined.
From Fig. 3 it can be seen that fatigue was considerably greater in the MS patients
than in controls, and that recovery offorce was less complete after 15 min of rest. An example
of the rapid reduction in force is given in Fig. 5 (right, middle), taken from another
investigation in which intermittent tetanic stimulation of the ankle dorsiflexor muscles was
employed to test fatigability in MS (Lenman et aI., 1989). These results were not associated
with excitation failure, as reflected in the M-waves of the San Francisco patients, and must
therefore have been due to changes in the muscle fibers. This conclusion is supported by the
results ofNMR spectroscopy in a subgroup of patients, who showed greater depression of
PCr, a larger rise in Pi and a more marked fall in pH, compared to control subjects (Fig. 4).
However, less obvious changes were observed in a subsequent study, in which a delay was
found in the resynthesis of PCr during the recovery period after fatiguing stimulation
(Kent-Braun et aI., 1994).
The most likely explanation for the altered physiological and metabolic properties
of the muscle fibers is that they stem from the effects of disuse or altered patterns of use, a
conclusion which is supported by the finding of similar fiber abnormalaties in patients with
spastic paraparesis from causes other than MS (Miller et aI., 1990). In this context, the results
of the study by Lenman and colleagues (1989) are pertinent; they found very significant
 502 A. J. McComas et al.

100

90
]
:8 80

...
&
.
u

70

60

Telanic stimulation Recovery

o 5 10 15 20 25

Time (min)

Figure 3. Greater fatigability of29 MS patients (open circles) compared with control subjects (fined circles)
during intermittent tetanic stimulation of ankle dorsiflexors. Values expressed as means ± SEMs. Note also
the delayed and incomplete recovery in the patients.

reductions in force during tetanic stimulation in 16 patients with paraplegia resulting from
spinal cord injury. An example from the study is included in Fig. 5.

Chronic Fatigue Syndromes

Numerous reports exist of syndromes which include chronic fatigue as a prominent
symptom. It may follow a viral or other illness. However, a strict etiological sequence which
has stood the test of time and multidisciplinary investigation has yet to be devised for the
fatigue or the subjectively reduced endurance. Clinically, the patients may describe disabling
subjective fatigue and compared with those patients with definite reasons for reduced
performance (e.g., polio, or polymyositis), their symptoms often seem disproportionately
severe relative to their usual daily activities. In particular, SUbjective recovery from simple
daily activities is markedly prolonged. Not surprisingly, some have been recently exposed
to viral infections, some have mild depressive symptoms, some have jobs with repetitive
manual tasks and others have generalized pain disorders (such as fibromyalgia).
Patients with presumed post-viral fatigue syndromes have no diagnostic features on
routine clinical EMG and nerve conduction studies. No specific findings occur in muscle
biopsies. However, the highly sensitive technique of SFEMG reveals abnormal jitter in a
proportion of patients (e.g., Jamal & Hansen, 1985; 1989; Connolly et al., 1993). The
limitation of this finding is that it is non-specific, only involves the lowest threshold motor
units, is not usually measured after objective fatigue, and is not accompanied by significant
blocking. Blocking is mandatory if any peripheral force reduction is to be ascribed to
defective neuromuscular transmission. Slightly increased fiber density has been reported in
a subgroup of those presenting with unexplained fatigue in whom myalgia and muscle
tenderness are major features (Connolly et al., 1993). An initial study suggesting an
abnormality of muscle metabolism based on NMR spectroscopy (Arnold et al., 1984) has
not been confirmed (Kent-Braun et al., 1993; Barnes et al., 1993). Immunological abnor-
malities (including reduced T cell subsets and impaired delayed hypersensitivity), together
Fatigue and Malfunction of the Nervous Systems 503

40

~..,
~
~

Ie
V
t. 20
~
1; 15
V
~

10

7.1

7.0

6.9
i
6.8

6.7

6.6

10 15 20 25

Time (min)
Figure 4. Results of 31p NMR spectroscopy in MS patients during intermittent tetanic stimulation of ankle
dorsiflexors. Values are means ± SEMs. Note that all three metabolites studied (Pi' per, and H1 show
significantly greater changes in patients (open circles) than in controls (filled circles).

with persistence of enteroviral RNA have been reported (Yousef et aI., 1988), but not
crucially linked to impaired peripheral muscle performance.
Several studies of small sample size have assessed muscle performance in these
patients and relatively normal results have been obtained. Lloyd and colleagues (1988) found
normal strength, maximal isometric endurance and recovery for the elbow flexors in 20
patients. In an extended test using interpolated stimulation and 45 min of submaximal
exercise, voluntary activation to more than 95% occurred in all patients and the decline in
maximal voluntary strength, voluntary drive, and the twitch force was not excessive.
504 A. J. McComas et al.

,
,
100

,
~

~ 80
x
v
"0
c::
60 ,,
v
:::I ,,
...
01) /

'"
'-
c:: 40
'"v
~

20
Con MS SCI

0
Figure 5. Left, mean fatigue indices (+ SEMs) of healthy subjects (Con), and of patients with MS (MS) or
spinal cord injury (SCI). Values derived from intermittent tetanic stimulation of ankle dorsiflexors. Right,
examples of typical responses from the three groups of subjects; the tetanic trains lasted a total of 3 min and
different amplifications were employed, since the initial force was considerably greater for the control subject
than for the two patients (reproduce from Lenman et ai., 1989, by permission of John Wiley & Sons, Inc.).

Interestingly, the SUbjective effort measured on a Borg scale during the test did not grow
abnormally during exercise in the patients (Lloyd et aI., 1991; see also Gibson et aI., 1993).
Almost all patients in the two studies described above had documented immunological
abnormalities. Quadriceps strength was normal in I I patients with post-viral fatigue (mostly
following infectious mononucleosis; Rutherford & White, 199 I; see also Stokes et aI., 1988)
and endurance over 20 min was not impaired. Two patients had reduced ability to drive the
muscles at rest. Performance in cycle ergometry was not impaired (Gibson et aI., 1993).
Kent-Braun and colleagues have studied the ankle dorsi flexors in seven patients with
chronic fatigue. The patients had lower maximal voluntary and twitch forces than controls
(by about 30%), although these changes were not statistically significant, and nor were the
patients' changes in pH and PCr with 25 min of submaximal isometric exercise excessive
(see also Barnes et aI., 1993). However, in this study, voluntary activation failure was present
in the patients prior to exercise and had increased abnormally at the end of the exercise: This
study highlights the difficulty in interpretation of voluntary endurance studies when volun-
tary activation is initially low. The level of exercise is overestimated for the patients (as
%MVC) because their true MVC was underestimated. Unfortunately, it is not simple to
correct for this, but one implication is that the sensitivity for detection of a peripheral
abnormality is reduced. The discrepancy between this result and that of Lloyd and colleagues
is likely to reflect patient selection because the twitch interpolation technique is likely to be
sensitive in both studies (see Gandevia et aI., Chapter 20). Another recent study using NMR
Fatigue and Malfunction of the Nervous Systems 50S

and dynamic plantarflexion contractions found the usual peak and pattern of changes in PCr,
Pi and pH in the patient group, but the changes occurred much more rapidly, suggesting
accelerated glycolysis, relative to oxidative metabolism, in the patients (Wong et aI., 1993).
In summary, biochemical, functional and electrophysiological indices of muscle
performance are abnormal in comparatively few individuals with prominent subjective
fatigue but no supporting pathology elsewhere. The fatigue in the so-called chronic fatigue
syndrome is more a result of a misperception of what is a normal level of fatigue than a
manifestation of abnormal muscle physiology at the level of the muscle or motor cortex and
motoneurons responsible for voluntary muscle activation. Measurements of actual perform-
ance incorporating some type of electrical stimulation to assess the adequacy of voluntary
drive are essential for accurate patient evaluation. Future studies must provide an adequate
description of patient groups so that abnormal findings can be investigated by others. Second,
given the difficulty in definition of anything resembling a homogeneous group, study groups
must be sufficiently large so that some reliance can be placed on either positive or negative
results. Third, the control groups must be appropriate for the patient group. This is easier to
recommend than to achieve, but is critical when small changes consistent with decreased
muscle conditioning are found.

FATIGUE STUDIES IN PERIPHERAL NERVOUS SYSTEM

DISORDERS
Although, as already observed, fatigue has been rather poorly investigated, even in
the research laboratory, there is a clinical impression that it may be a feature of almost any
peripheral nervous system disorder. Rather than survey each disorder in tum, this review
will divide the disorders into four classes and will look for general pathophysiological
principles within each (Fig. 6).

Denervating Disorders
Ifthere is a loss of excitable muscle mass as a result of denervation, fatigue will result
for the reasons given in the introduction. However, there are four additional considerations
which are applicable to all denervating disorders regardless of etiology.
First, in disorders of the motoneuron soma (polio, ALS, spinal muscular atrophy) or
of its axon (peripheral neuropathy) there will be a strong tendency on the part ofthe surviving
neurons to develop axonal sprouts and to innervate the muscle fibers which have been
relinquished by the degenerating motoneurons. This process of collateral reinnervation is
remarkably efficient and may be adequate to sustain muscle bulk and twitch tension until
80% of the motoneuronal population is lost (McComas et aI., 1971). In the early stages of
reinnervation, however, impaired impulse conduction in the new axon sprouts and dysfunc-
tion in their endings may be responsible for increased neuromuscular jitter, as revealed by
SFEMG, and the latter phenomenon is probably also present in the early stages of a
dying-back neuropathy. Impaired synaptic function in neuropathies can also be determined
by monitoring M-wave responses to repetitive nerve stimulation, or even by assessing
individual motor unit potentials during voluntary contraction. (e.g., McComas et aI., 1971).
A second consideration, applicable to all chronic neuropathies, concerns the propor-
tions of fast-twitch and slow-twitch motor units which will remain in the affected muscle,
and which will have increased their sizes by collateral reinnervation. According to Rafuse
and colleagues (1992), all motor units increase proportionately in size so that those which
were previously the largest will adopt the greatest numbers of denervated muscle fibers. If
 506 A. J. McComas et al.

Chronic de nervation
<with sproutjne) .
Fibre type disproportion .../
Branch POint}. / ....
Terminal failure ...""
K+ effects .........
....

.i l

/
Junction disorders
Myasthenia etc.

1--..·.. _
,
I
.. __ · _ - - - _ · _ - - ,
!
I '
I
.. __.__ . __...__. . __. J
I

Ene,&), djsorders
Glycolytic (e.g. McArdle)
Lipid transportL ' .
Respiratory chain} mltochoodna
etc.

Figure 6. Summary of peripheral neuromuscular disorders capable of inducing fatigue, arranged in four
categories.

the loss offast-twitch and slow-twitch motor units is random, the expansionist proclivity of
the fast-twitch units may result in normal proportions of fast-twitch and slow-twitch muscle
fibers, albeit with fiber type grouping. In some cases, by chance, the proportion of fast-twitch
fatigable muscle fibers may be distinctly higher or lower, and hence patients would be
expected to demonstrate greater or lesser susceptibility to peripheral fatigue respectively.
The same considerations might be expected to apply to those muscle disorders in which there
was preferential involvement of one or other main fiber types. However, in congenital type
I fiber predominance any difference in fatigability is over-ridden by significant weakness
(Linssen et aI., 1991).
A third consideration arises out of recent awareness of the importance of Na+,
K+-pumping in offsetting K+ effects during exercise (see Sj0gaard & McComas, Chapter 4).
In a normal muscle the fibers of one motor unit are intermingled with those of20 or so others
and, hence, during submaximal contractions, K+, released by the contracting fibers, can
diffuse into the interstitial spaces surrounding the quiescent fibers. Further, the latter fibers
can reduce the interstitial K+ concentration by increasing Na+, K+ exchange through en-
Fatigue and Malfunction ofthe Nervous Systems 507

hanced Na+, K+ -pumping (Kuiack & McComas, 1992). In chronic denervation, however, the
juxtaposition of active and inactive fibers is disrupted due to the presence of fiber type
grouping and it seems likely that decreased excitability of muscle fibers, due to K+ -induced
depolarization, may result.
Finally, if the denervating disorder is one that affects sensory neurons and fibers, as
well as motor ones, the diminished impulse traffic from type Ia and II endings would be
expected to reduce oliogosynaptic excitation ofmotoneurons during voluntary contractions,
contributing to central fatigue. Similarly, in a severe neuropathy, the lack ofafferent feedback
would be anticipated to interfere with the sense of effort and with the planning, amplification
and transmission of the motor command.

Neuromuscular Junction Disorders

The archetypal neuromuscular disorder is myasthenia gravis, in which fatigability,
especially of the external ocular muscles, is the cardinal symptom. It has been known for
some time that the disorder is due to an immunological attack on the acetylcholine (ACh)
receptors. However, it is still not clear why fatigability should become worse as the day
progresses, especially since there is no evidence for an associated deficiency of ACh in the
motor nerve terminals (Ito et aI., 1976). The diagnosis of myasthenia gravis may be
confirmed in a number of ways, of which the most sensitive is SFEMG (Ekstedt & St111berg,
1967), particularly as applied to muscles of the face. Estimation of ACh receptor antibodies
in the serum is slightly less sensitive, followed by repetitive nerve stimulation. The ability
of anticholinesterase medication to relieve weakness and fatigability may be dramatic,
especially in the Tensilon test. The improvement in muscle function is due to the ability of
the diminished ACh receptors to have multiple hits from ACh molecules which have escaped
hydrolysis; thus the endplate potential is prolonged and may better reach threshold or
produce multiple firing. In the Lambert-Eaton myasthenic syndrome, although there is severe
weakness, the problem of fatigue is reversed inasmuch as performance improves at the task
progresses. This condition resembles myasthenia gravis in being autoimmune in origin, with
antibodies are directed against the Ca2+ channels of the motor nerve terminal (Vincent et aI.,
1989).

Disorders of the Muscle Plasmalemma

In the hereditary myotonic disorders, myotonia congenita and myotonic dystrophy,
patients may have an exceptionally fast form of fatigue which appears within seconds of
voluntary or stimulated contractions and can reduce force to less than half the initial value.
As the contractions continue, however, the paresis gradually wears off (warm-up phenome-
non). In both conditions, recordings of EMG activity or of M-waves evoked by nerve
stimulation, show that the problem is one of impaired plasmalemmal excitability, due to
K+ -mediated depolarization of the contracting fibers. In myotonia congenita the increased
sensitivity is known to result from the low Cl· conductance of the muscle plasmalemma as
a consequence of the expression of abnormal Cl· channel genotypes (Koch et aI., 1992). The
nature of the problem in myotonic dystrophy is more obscure, since no consistent channel
abnormalities have been found so far. Fenton and colleagues (1991) have suggested, in
keeping with various lines of investigation, that defective Na+, K+-pumping may be respon-
sible.
In hyperkalemic familial periodic paralysis and the closely related paramyotonia
congenita, fatigue is also due to diminished plasmalemmal excitability but it is rather less
rapid than in myotonia, running a time-course measured in minutes rather than seconds.
Further, the fatigue tends to become worse after the voluntary contractions cease, and the
508 A. J. McComas et al.

M-waves may decline to as little as one quarter of the initial value within 30 min (McManis
et aI., 1986). In these conditions the genetic abnormality is one that affects the Na+ channels
(Fontaine et aI., 1990).

Energy Disorders
In the contracting muscle fiber a renewed supply of ATP is essential, not only for the
contractile process itself but also for the vigorous ion pumping which takes place across the
membranes of the muscle fiber surface, T-tubules and sarcoplasmic reticulum.
The first report of an hereditary metabolic disorder characterized by rapid fatigue
was that of McArdle (1951). The absence of a rise in blood lactate concentration during
exercise led McArdle to postulate that muscle glycogen could not be broken down, a
prediction confirmed 8 years later, when a deficiency of myophosphorylase was discovered
(Mommaerts et aI., 1959). A feature of McArdle's disease is severe muscle cramping on
exercise. The absence of associated EMG activity suggests that the muscle fibers enter a
rigor state, as would be expected if ATP was unavailable to detach the myosin cross-bridges
from the actin filaments. There is also a marked decline in the amplitude of the M-wave
(Dyken et aI., 1967; their Fig. 7; also Cooperet aI., 1989), a finding which could be attributed
to diminished Na+, K+-pumping secondary to ATP deficiency. However, both biopsy and
NMR spectroscopy studies have shown, surprisingly, that ATP concentration is not dimin-
ished in the muscle fibers (Wiles et aI., 1981; Ross et aI., 1981). Since there is no obvious
reason for any inhibition of ATP hydrolysis, the biochemical basis for the physiological
abnormalities remains enigmatic. In this context, the suggestion that there may be a
compartmental reduction in ATP within the fiber, which does not reveal itself in whole muscle
studies, seems rather unlikely. Clearly, more can be learned about the biochemical basis of
muscle fatigue from further studies of this disorder.
After McArdle's seminal contribution, reports of other types of hereditary glycolytic
disorders followed. Like myophosphorylase deficiency, these conditions are rare and most
are associated with fatigue. A different type of metabolic disorder, also characterized by

100
_::e'-----~;::,O-- -0----0._ --0."'0.

"'Q"
"0_
"a
""'0
...o "
"0,
c "''0"
8 50 "'OM-wave
~
Force
Force

Patients Controls

o 2 3

Time (min)
Figure 7. Fatigue studies in patients with McArdle's disease (filled circles) and control subjects ( open circles),
Mean values are shown without error bars (for clarity), Note that there is not only a dramatic loss of force, but
that muscle excitability is also affected (adapted from Cooper et ai., 1989),
Fatigue and Malfunction of the Nervous Systems 509

excessive fatigability, is carnitine deficiency. As carnitine is required for the transport of

fatty acids into mitochondria, there is an inability to use lipid as a fuel for the later stages of
exercise, after the muscle glycogen has been consumed.
A third class of metabolic disorders comprises those in which there are defects in the
mitochondrial respiratory chain or in which oxidative metabolism is dissociated from ATP
production, as in Luft's disease (Luft et aI., 1962). In the latter condition, the energy from
oxidative metabolism is released as heat. Within this class of disorders mention should be
made of cytochrome b deficiency, in view of the fatigue studies which were undertaken when
this condition was first recognized (Morgan-Hughes et aI., 1977). Mitochondrial dysfunction
appears to be more frequent in old age, as the organelle population declines within the muscle
fibers; it is probable that multiple enzyme defects result. A mitochondrial etiology should,
however, be suspected in any patient with undue fatigability and in whom no other cause
can be demonstrated (cf. Van Ekeren et aI., 1991).
At the present time, it appears that different genetic abnormalities may affect almost
all the enzymes involved in glycogenolysis, lipid metabolism, the citric acid cycle and the
cytochrome chain, and that peripheral fatigue, with or without muscle pathology, is the
inevitable consequence.

CONCLUSIONS

It is evident from this review that, regardless of the nature of the primary pathophysi-
ological deficit, fatigue may be a complex process, with important contributions from one
or more secondary factors. In this respect, the findings of the various studies have verified
and emphasized themes which had been anticipated in the introduction. One example of this
complexity is the impaired voluntary activation of motoneurons which occurs in those
postpolio patients with moderate disability, and which must be attributed to disuse. The
effects of disuse and/or altered patterns of use are also evident in the muscle fibers of patients
with central lesions, as in those patients with MS or spinal cord injury. In the MS patients,
not only the force recordings but the results of NMR spectroscopy also, show striking
abnormalities. So far as the primary defects are concerned, the most dramatic fatigue effects
have been seen in those patients with peripheral neuromuscular disorders, such as McArdle's
disease, periodic paralysis and myotonic dystrophy.
Sufficient is now known about the cellular and body system mechanisms in fatigue,
to justify more frequent application of fatigue testing to patients. Such studies will certainly
tell us more about physiological perturbations in disease but, of equal importance, they will
provide a firm reference point for assessing the results of therapeutic interventions. In this
context, it is clear that considerable functional improvement would be expected from
appropriate training, to overcome the central and peripheral effects of disuse. Indeed, we
consider that there is a need for carefully designed training studies in neurological patients
with fatigue, in whom the different elements in the physiological chain of muscle contraction
can be dissected out for analysis. It is to be hoped that the results of such trials will become
regular features of future fatigue symposia.

ACKNOWLEDGMENTS

The authors wish to thank Dr. Christine Thomas for her helpful comments on the
manuscript. The authors also express their gratitude to their respective technical and
secretarial staffs. The authors' laboratories have been supported by: the Muscular Dystrophy
Association and the Natural Sciences and Engineering Research Council (NSERC) of
 510 A. J. McComas et al.

Canada, and the De Groote Foundation, Canada ([Link].); the Multiple Sclerosis Society
and the Muscular Dystrophy Association (USA; R. G.M.); and the National Health & Medical
Research Council of Australia, and the Asthma Foundation of New South Wales, Australia
(SC.G.). Attendance of two of the authors at the 1994 Bigland-Ritchie conference was
supported by NSERC ([Link].), the University of Miami (S C. G. and [Link].), and by
the Muscular Dystrophy Association (USA; Sc.G.).

REFERENCES
Allen GM, Gandevia SC, Neering IR, Hickie I, Jones R & Middleton J (1994) Muscle performance, voluntary
activation and perceived effort in normal subjects and patients with prior poliomyelitis. Brain 117,
661-670.
Argov Z & Bank WI (1991). Phosphorus magnetic resonance spectroscopy (3lp MRS) in neuromuscular
disorders. Annals ofNeurology 30, 90-97.
Arnold DL, Bore PJ, Radda GK, Styles P & Taylor D (1984). Excessive intracellular acidosis of skeletal
muscles in exercise in a patient with a post viral exhaustion fatigue syndrome. Lancet 1(8391),
1367-1369.
Barnes PRJ, Taylor DJ, Kemp GJ & Radda GK (1993). Skeletal muscle bioenergetics in the chronic fatigue
syndrome. Journal of Neurology Neurosurgery and Psychiatry 56, 679-683.
Bigland-Ritchie B, Cafarelli E & V 0llestad NK (1986). Fatigue of submaximal contractions. Acta Physiologica
Scandinavica 128 (suppI556), 137-148.
Bruno RL, Frick NM & Cohen J (1991). Polioencephalitis, stress, and the etiology of post-polio sequelae.
Orthopedics 14, 1269-1275.
Connolly S, Smith DG, Doyle D & Rowler CJ (1993). Chronic fatigue: electromyographic and neuropathologi-
cal evaluation. Journal ofNeurology 240, 435-438.
Cooper RG, Stokes MJ & Edwards RHT (1989). Myofibrillar activation failure in McArdle's disease. Journal
ofNeurological Sciences. 93,1-10.
Denys EH & Norris FH (1979). Amyotrophic lateral sclerosis: impairment of neurologic transmission. Archives
ofNeurology 36, 202-205.
Dolmage T & Cafarelli E (1991). Rate of fatigue during repeated submaximal contractions of human
quadriceps muscle. Canadian Journal of Physiology and Pharmacology 69, 1410-1415.
Duchateau J & Hainaut K (1987). Electrical and mechanical changes in immobilized human muscle. Journal
of Applied Physiology 62,2168-2173.
Dyken ML, Smith DM & Peake R (1967). An electromyographic diagnostic screening test in McArdle's disease
and a case report. Neurology 17, 45-50.
Ekstedt J & Stalberg E (1967). Myasthenia gravis. Diagnostic aspects by a new electrophysiological method.
Opuscula Medica 12, 73-76.
Fenton J, Gamer S & McComas AJ (1991). Abnormal M-wave responses during exercise in myotonic muscular
dystrophy. A Na+, K +-pump defect? Muscle & Nerve 14, 79-84.
Fontaine B, Khurana TS, Hoffman EP, Bruns GAP, Hains JL, Trofatter JA, Hanson MP, Rich J, McFarlane H,
McKenna Yasek D, Romano D, Gusella JF & Brown RH (1990). Hyperkalemic periodic paralysis
and the adult muscle sodium channel a-subunit gene. Science 250, 1000-1002.
Fuglsang-Frederiksen A & Scheel U (1978). Transient decrease in number of motor units after immobilization
in man. Journal of Neurology Neurosurgery and Psychiatry 41, 924-929.
Gibson H, Carroll M, Clague JE & Edwards RH (1993). Exercise performance and fatiguability in patients
with chronic fatigue syndrome. Journal of Neurology Neurosurgery and Psychiatry 56, 993-998.
Gordon T & Stein RB (1982). Reorganization of motor unit properties in reinnervated muscles of the cat.
Journal of Neurophysiology 48, 1175-1190.
Ito Y, Miledi R, Molenaar PC, Vincent A, Polak RL, Van Gelder M & Davis IN (1976). Acetylcholine in human
muscle. Proceedings of the Royal Society of London B Biological Series 192, 475-480.
Jamal GA & Hansen S (1985). Electrophysiological studies in the postviral fatigue syndrome. Journal of
Neurology Neurosurgery and Psychiatry 48,691-694.
Jamal GA & Hansen S (1989). Postviral fatigue syndrome: evidence for underlying organic disturbance in the
muscle fibre. European Neurology 29, 273-276.
Kent-Braun JA, Sharma KR, Miller RG & Weiner MW (1994). Post-exercise phosphocreatine resynthesis is
slowed in multiple sclerosis. Muscle & Nerve 17, 835-841.
Fatigue and Malfunction of the Nervous Systems 511

Kent-Braun JA, Shanna KR, Weiner MW, Massie B & Miller RG (1993). Central basis of muscle fatigue in
chronic fatigue syndrome. Neurology 43, 125-131.
Koch MC, Steinmeyer K, Lorenz C, Ricker K, Wolf F, Otto M, Zoll B, Lehmann-Hom F, Grzechik KH &
Jentsch TJ (1992). The skeletal muscle chloride channel in dominant and recessive human myotonia.
Science 257,797-800.
Kuiack S & McComas AJ (1992). Transient hyperpolarization of non-contracting muscle fibres in anaesthe-
tized rats. Journal of Physiology (London) 454, 609-618.
Lenman AJR, Tulley F, Vrbova G, Dimitrijevic M & Towle JA (1989). Muscle fatigue in some neurological
disorders. Muscle & Nerve 12, 938-942.
Linssen WH, Stegeman DF, Joosten EM, Merks HJ, ter Laak HJ, Binkhorst RA, & Notennans SL (1991).
Force and fatigue in human type I muscle fibres. A surface EMG study in patients with congenital
myopathy and type I fibre predominance. Brain 114, 2123-2132.
Lloyd AR, Gandevia SC & Hales JP (1991). Muscle perfonnance, voluntary activation, twitch properties and
perceived effort in nonnal subjects and patients with the chronic fatigue syndrome. Brain 114, 85-98.
Lloyd AR, Hales JP & Gandevia SC (1988). Muscle strength, endurance and recovery in the post-infection
fatigue syndrome. Journal of Neurology Neurosurgery and Psychiatry 51, 1316-1322.
Luft R, Ikkos D, Palmieri G, Ernster L & Afzelius B (1962). A case of severe hypennetabolism of nonthyroid
origin with a defect in the maintenilnce of mitochondrial control - a correlated clinical, biochemical
and morphological study. Journal of Clinical Investigation 41, 1776-1804.
Maier A, Crocket JL, Simpson DR, Saubert CW & Edgerton VR (1976). Properties of immobilized guinea pig
hindlimb muscles. American Journal ofPhysiology 231,1520-1526.
McArdle B (1951). Myopathy due to a defect in muscle glycogen breakdown. Clinical Science 10, 13-33.
McComas AJ, Kereshi S & Quinlan J (1983). A method for detecting functional weakness. Journal of
Neurology Neurosurgery and Psychiatry 46, 280-282.
McComas AJ, Sica REP, Campbell MJ & Upton ARM (1971). Functional compensation in partially denervated
muscles. Journal ofNeurology Neurosurgery and Psychiatry 34, 453-460.
McManis PG, Lambert EH & Daube JR (1986). The exercise test in periodic paralysis. Muscle & Nerve 9,
704-710.
Miller RG, Boska MD, Moussavi RS, Carson PJ & Weiner MW (1988). 31p nuclear magnetic resonance studies
of high energy phosphates and pH in human muscle fatigue. Journal of Clinical Investigation 81,
1190-1196.
Miller RG, Giannini D, Milner-Brown HS, Layzer RB, Koretsky AP, Hooper D, & Weiner MW (1987). Effects
of fatiguing exercise on high-energy phosphates, force and EMG: evidence for three phases of
recovery. Muscle & Nerve 10,810-821.
Miller RG, Green AT, Moussavi RS, Carson PJ & Weiner MW (1990). Excessive muscle fatigue in patients
with spastic paraparesis. Neurology 40, 1271-1274.
Mommaerts WFHM, Illingworth B, Pearson CM, Guillorg RJ & Seraydarian K (1959). A functional disorder
of muscle associated with the absence of phorsphorylase. Proceedings of the National Academy of
Sciences of the United States ofAmerica 45,791-797.
Morgan-Hughes JA, Darveniza P & Kahn SA (1977). A mitochondrial myopathy characterized by a deficiency
in reducible cytochrome b. Brain 100, 617-640.
Moussavi RS, Carson PJ, Boska MD, Weiner MW & Miller RG (1989). Nonmetabolic fatigue in exercising
human muscle. Neurology 39, 1222-1226.
Nicklin J, Kami Y & Wiles CM (1987). Shoulder abduction fatigability. Journal of Neurology Neurosurgery
and Psychiatry 50,423-427.
Petit J & Gioux M (1993). Properties of motor units after immobilization of cat peroneus longus muscles.
Journal ofApplied Physiology 74, 1131-1139.
Rafuse VF, Gordon T & Oroczo R (1992). Proportional enlargement of motor units after partial denervation
of cat triceps surae muscles. Journal of Neurophysiology. 68, 1261-1276
Rice CL, Volmer TL & Bigland-Ritchie B (1992). Neuromuscular responses of patients with multiple sclerosis.
Muscle & Nerve 15, 1123-1132.
Robinson GA, Enoka RM & Stuart DG (1991). Immobilization-induced changes in motor unit force and
fatigability in the cat. Muscle & Nerve 14, 560-573.
Ross BD, Radda GK, Gadian DG, Rocker G, Esiri M & Falconer-Smith J (1981). Examination of a case of
suspected McArdle's syndrome by 31 P nuclear magnetic resonance. New England Journal ofMedicine
304, 1338-1343.
Rutherford OM & White PD (1991). Human quadriceps strength and fatiguability in patients with post-viral
fatigue. Journal ofNeurology Neurosurgery and Psychiatry 54, 961-964.
512 A. J. McComas et al.

Sanjak M, Paulson D, Sufit R, Reddan W, Beaulieu D, Erickson L, Shug A & Brooks BR (1987). Physiologic
and metabolic response to progressive and prolonged exercise in amyotrophic lateral sclerosis.
Neurology 37, 1217-1220.
Sharma KR, Kent-Braun J, Mynhier MA, Weiner MW & Miller RG (1994). Excessive muscular fatigue in the
postpoliomyelitis syndrome. Neurology 44, 642-646.
Stokes MJ, Cooper RG & Edwards RHT (1988). Normal muscle strength and fatiguability in patients with
effort syndromes. British Medical Journal 297, 1014-1017.
Trojan DA, Gendron D & Cashman NR (1993). Anticholinesterase-responsive neuromuscular junction
transmission defects in post-poliomyelitis. Journal o/Neurological Sciences 114, 170-177.
Van Ekeren GJ, Cornelissen EAM, Stadhouders AM & Sengers RCA. (1991). Increased volume density of
peripheral mitochondrion in skeletal muscle of children with exercise intolerance. European Journal
0/Pediatrics 150, 744-750.
Vincent A, Lang B & Newsom-Davis J (1989). Autoimmunity to the voltage-gated calcium channel underlies
the Lambert-Eaton myasthenic syndrome, a paraneoplastic disorder. Trends in Neuroscience 12,
496-502.
Westgaard R & L0mo T (1988). Control of contractile properties within adaptive ranges by patterns of impulse
activity in the rat. Journal o/Neuroscience 8, 4415-4426.
Wiles CM, Jones DA & Edwards RHT (1981). Fatigue in human metabolic myopathy. Ciba Foundation
Symposium 82, 264-282.
Wong R, Lopaschuk F, Zhu G, Walker D, Catellier D, Burton D, Teo K, Collins-Nakai R & Montague T (1993).
Skeletal muscle metabolism in the chronic fatigue syndrome. Chest 102, 1716-1722.
Yousef GE, Bell EJ, Mann GF, Murugesan V, Smith DG & McCartney RA (1988). Chronic enterovirus
infection in patients with post viral fatigue syndrome. Lancet 1(8578), 146-150.
Zochodne DW, Thompson RT, Driedger AA, Strong MJ, Gravelle D & Bolton CF (1988). Metabolic changes
in human muscle denervation: topical 31p NMR spectroscopy studies. Magnetic Resonance in
Medicine 7, 373-383.
EPILOGUE
 37

NEUROBIOLOGY OF MUSCLE FATIGUE

Advances and Issues

S. C. Gandevia, I R. M. Enoka,2 A. J. McComas,3 D. G. Stuart,4 and

C. K. Thomas 5

I Prince of Wales Medical Research Institute

Sydney, Australia
2 Department of Biomedical Engineering
Cleveland Clinic Foundation
Cleveland, Ohio 44195
3 Department of Biomedical Sciences, McMaster University
Hamilton, Ontario, L8N 3Z5, Canada
4 Department of Physiology
University of Arizona College of Medicine
Tucson, Arizona 85724
5 The Miami Project to Cure Paralysis
University of Miami School of Medicine
Miami, Florida 33136

When you cannot measure it.

When you cannot express it in numbers -
You have scarcely, in your thoughts,
Advanced to the stage of Science, whatever the matter may be.
-Lord Kelvin

In reading this volume one should ask what is new and true, and what will lead to
new approaches and insight. Clearly, as the topic muscle fatigue is increasingly studied by
muscle physiologists, neuroscientists and clinicians, it becomes more difficult to summarize
the state of knowledge, even in a volume with 36 chapters. Since a major symposium on the
physiology of muscle fatigue in London in 1980 (Porter & Whelan, 1981), much progress
has been made in understanding both the intracellular, muscle-fiber mechanisms and, to some
extent, mechanisms of CNS control as they relate to muscle fatigue. Some of this progress
has been summarized at recent symposia in Paris in 1990 (Atlan et aI., 1991), and Amsterdam
in 1992 (Sargeant & Kernell, 1993). Despite this, we are far from having an equation that
will predict the degree of fatigue under a variety of physiological circumstances, although
some of the boundary conditions are becoming well established.
In this epilogue, we have addressed some points that became evident at the 1994
Bigland-Ritchie Symposium in Miami at which meeting the majority of the chapters were

515
516 S. C. Gandevia et al.

partially presented, and other issues that arose when editing the volume and comparing it
with the 1980, 1990, and 1992 Symposia. These issues are examined in the conventional
framework, moving from the periphery to the central nervous system (CNS) while comments
on the motoneuron are considered in a separate section termed, "at the motoneuron inter-
face". Additionally, there is some consideration offatigue in adapted states, and pathophysi-
ological conditions. From this perspective, it is now widely accepted that sustained physical
activity causes a reduction in the maximal force a muscle can exert due to the concurrent
impairments of several (more or less) simultaneous processes which begin upstream of the
motor cortex and go down to the myofibril.
As it is now possible to measure many factors that change as fatigue develops, the
challenge is to determine if, when, and how much they are contributing to the reduction in
force. This is a difficult task. For example, some biochemical changes, such as an increase
in ADP concentration, affect force and the maximal velocity of shortening differentially;
ADP increases maximal force but depresses maximal shortening velocity. The reduced speed
of shortening combined with slowed relaxation diminish performance, and this effect can be
just as disabling as the loss of force. Doubtless it is theoretically possible to construct a
situation in which one, say, intracellular process is the crucial limit to force, but in typical
exercise it is inevitable that more than one process is involved. One of the major challenges
for the field of muscle fatigue will be to quantitate the mounting qualitative evidence (Section
VII; see also Enoka & Stuart, 1992) that the mechanisms of fatigue and recovery, from the
molecular to the whole organism level, are task dependent. This theme is prominent
throughout the volume, from the introductory consideration of single muscle fibers to the
concluding chapters on neuromuscular adaptations. In parallel with this emphasis on
task-dependency, there is need to continue work on the individual molecular processes that
generate force. Without such information, the fundamental mechanisms will remain un-
known.

IN THE PERIPHERY

Muscle produces force and movement through the interaction of myosin crossbridges
with actin filaments, and it is through these crossbridges that fatigue must be manifest.
Expressed most simply, fatigue is due to a reduction in the number of interactions, or in the
forces developed by individual crossbridges. In this context, it is remarkable that, with an
in vitro laser trap system, it is now possible to measure the movement in a single crossbridge
cycle, or the tension generated by a single crossbridge when movement is prevented (Finer
et aI., 1994). To further our understanding offatigue mechanisms, it should be a relatively
easy step to study the effects on the isolated crossbridge of alterations in the concentrations
ofATP, ADP, Pj, Ca2+ and H+, all of which alter in fatigue. These measurements would include
not only the amplitudes of movement or force, but also the speeds with which these are
generated. Equally relevant to an understanding offatigue are three-dimensional models of
the myosin head, as deduced by X-ray diffraction and computer analysis (Rayment et aI.,
1993).
Although future fatigue symposia will include observations at the level of the
individual myosin molecule or crossbridge, for the present summary we must begin with the
single fiber preparation. In Chapter 1, Edman has shown, by comparisons of electrical
stimulation and K+ contracture, that moderate fatigue appears to result from the reduced
forces developed by individual crossbridges and that acidification of the fiber interior is
likely to be one of the factors involved. Another of the intracellular candidates for force
reduction during fatigue is the excitation-contraction coupling mechanism, a topic discussed
in detail at the molecular level by Stephenson and colleagues in Chapter 2. Critical
Neurobiology of Muscle Fatigue 517

observations here are the measurements of Ca2+ flux within single muscle fibers by means
of fluorescent dyes. First achieved by Ashley and Ridgway (1970) in the large fibers of the
barnacle, the methodology was subsequently extended to amphibian muscle by Blinks and
colleagues (1978) and, finally, to the smaller and more vulnerable fibers of the mouse by
Allen, Westerblad and Lannergren (Chapter 3). The findings from these last studies agree
with those of Edman in showing that early fatigue is not due to any failure of excitation-con-
traction coupling. As fatigue progresses, however, there is diminished Ca2+ release from the
SR, as well as reduced sensitivity of the myofibrils to Ca2+.
The term excitation-contraction coupling, while remaining conceptually simple, is
being exposed as a cascade of events encompassing many processes, each modifiable in a
multitude of interacting ways (Chapter 2). For example, the sarcoplasmic reticulum Ca2+
release channel seems differentially controlled by ATP (increasing channel opening) and
Mg2+ (tonically inhibiting channel opening). Other biochemical factors, such as pH and
inorganic phosphate are also able to modify the Ca2+ release channel.
Given the large number of possibilities, as described by Stephenson in Chapter 2,
what is the cause of the decreased Ca2+ release? Here, there are major differences in
viewpoint. On the one hand, Allen and colleagues find no evidence for failure of the inward
spread of excitation in the T-tubules. On the other hand, Edman, using a wavy myofibrillar
preparation originally developed by Gonzalez-Serratos (1971), reaches the opposite conclu-
sion. If inward spread is indeed impaired, it seems likely that K+-induced depolarization of
the T-tubules is an important contributory factor, given the striking increases in K+ concen-
tration which have been observed, or deduced, in the interstial fluid of contracting muscle
(Chapter 4). It is perhaps surprising, in view of the size of the impulse-mediated K+ efflux
and the narrowness of the interfiber spaces and T-tubules, that K+ effects are not more
prominent in fatigue. However, as reviewed by Sjegaard and McComas, there are a number
of cellular mechanisms, including enhanced electrogenic Na+-K+ pumping, which stave off
K+ -induced paralysis. In addition to failure of inward excitation in fatigue, there are other
factors including intracellular acidosis that will reduce Ca2+ release from the SR through the
ryanodine channels.
Observations of the biochemical changes in muscle fibers are critical to an under-
standing of fatigue, and complement the findings of the various types of physiological
experiments. The greatest advance in this field is NMR spectroscopy which enables the
concentrations of high-energy phosphates, Pi and H+ to be measured in situ in contracting
human muscle. The technique Of3lp NMR has become an important part of the armamen-
tarium of those examining the metabolic performance of whole human muscles (Chapters
12-14). Such studies have been important in confirming that only slight reductions in ATP
concentrations occur in fatigue, at a time when phosphocreatine has all but disappeared and
the H+ concentration has markedly increased. Nevertheless, values obtained with NMR are
averages, and ignore the fact that the different types of fiber will have been employed to
greater or lesser extents in the fatiguing contractions and, because of their differing biochem-
istry, are likely to have considerable differences in metabolite concentration. Even in the
same fiber, a value for ATP says little about the conditions likely to exist at such critical sites
as the cross bridges, and the SR and surface membrane pumps.
Not only does NMR spectroscopy reveal the status of the high-energy phosphates,
but it promises more with measurements of water and ionic fluxes across the cell membrane.
Results with NMR appear to agree well with those obtained with muscle biopsies, although
neither method reveals the metabolite concentrations at the level of the actomyosin cross-
bridges. This is unlikely to be relevant for ATP concentrations which probably do not drop
below critical levels at the crossbridge under natural conditions, but may be important for
other factors that modify crossbridge action. Furthermore, with NMR and usually with
biopsy studies, the concentrations across the sampled tissue are averages so that the level of
518 S. C. Gandevia et al.

neural drive to the particular muscle or motor unit needs to be determined. At a molecular
level, the way in which the intracellular changes in fatigue interact at the level of the
crossbridge and actomyosin ATPases must be known. Of particular interest is what factors
differentially impair crossbridge force as opposed to crossbridge cycling and how these
change in eccentric exercise when the crossbridges sustain more force and do it more
efficiently. Eccentric exercise produces excessive muscle damage perhaps because of inho-
mogeneity in the strength of individual sarcomeres and thus local sarcomere dynamics may
generate fatigue, pain and altered muscle performance which take days to recover. In
addition, unexpected evidence has emerged that the energy cost of contractions (assessed by
oxygen consumption or heat output) remains constant during dynamic contractions but
increases with isometric contractions.
At the risk of being overwhelmed by the complexity of molecular controls on muscle
bioenergetics, Kushmerick (Chapter 14) attempts a simplification. He points out that human
muscles are relatively homogeneous from a biochemical point of view compared with those
of common experimental animals, and thus they can be regarded as consisting of a unimodal
metabolic continuum. This view, of course, bolsters the global approach offered by NMR
and reveals the likely way in which the products of the ATPases in muscle (ADP, Pi, H+ and
creatine) must ultimately regulate energy balance. Applicability of this view must be
tempered by evidence that activation of crossbridges may sometimes be a critical impairment
(Chapter 13). Furthermore, the concept of a unimodal metabolic continuum is at odds with
the literature on muscle-fiber and motor-unit types. In this scheme, typing is accomplished
by classifying muscle fibers and motor units on the basis of biochemical, morphological,
and physiological properties. These classifications are not rigid distinctions based on
absolute criteria but rather characterize the populations on the basis of relative values. These
schemes have proven to be robust (Chapters 8-10,16,17,24,26,27; Stuart et aI., 1984;
Rome, 1993; Stuart & Callister, 1993) and much remains to be done to assess the validity
of the Kushmerick model as it applies to less constrained systems.
In addition to the widespread use ofNMR spectroscopy in studies on muscle fatigue,
there are a number of developments that promise new avenues for exploring human
performance. Bertocci (Chapter 15) reviews the most promising of these new opportunities;
these include l3C magnetic resonance spectroscopy (MRS) to examine the regulation of
substrate into the citric acid cycle, 23N a and 39K MRS to determine the distribution of sodium
and potassium ions across the membrane, and IH protons combined with the mechanisms of
nuclear spin relaxation to provide spatial information on the location and movement of water.
There seems little doubt that the results obtained with these methods will advance our
understanding of the mechanisms underlying muscle fatigue. For example, the measurement
of longitudinal (Tl) and transverse (T2) relaxation times with magnetic resonance imaging
can provide spatial information, both static and dynamic spatial information on regions
within a muscle or group of muscles that have been activated in the performance of a
particular task. The information that can be obtained with these techniques may ultimately
be as significant as that obtained with the development of the EMG.
In studies of human muscles in vivo, isolation of some of the force failure to
excitation-contraction coupling is helpful but it says little about which processes are critically
involved. An accepted hallmark of coupling of contraction to neural excitation is the selective
impairment of force production with low compared with high rates of stimulation. This
frequency-dependent alteration in force means that the extent of fatigue must depend on the
exact range of frequencies used by motor units under natural conditions. Hence it will remain
necessary to obtain information about the naturally-occurring discharge rates of motor units.
A second reason for this approach is that factors favoring both fatigue and force potentiation
may coexist (Chapter 9), so that the liability of the force-frequency relationship needs to be
emphasized.
Neurobiology of Muscle Fatigue 519

The potential for neuromuscular transmission failure to reduce force (Chapters 5-7)
is undenied. However, the role of axonal factors, such as branch-point failure proximal to
the presynaptic terminal in muscle fatigue during normal voluntary activities is not estab-
lished. Although transmission across the neuromuscular junction is typically regarded to
have a high safety factor, Fuglevand (Chapter 6) suggests that experiments have not
adequately distinguished between a reduction in action potential amplitude and a depolari-
zation-induced contraction (due to the extracellular accumulation of potassium) to determine
if impairment of neuromuscular transmission contributes to the decline in force. Certainly,
fast-twitch, fatigable motor units exhibit greater reductions in action potential amplitude
with sustained activity. Neuromuscular propagation might become impaired in those
neuromuscular diseases in which there is chronic continuous reinnervation of muscle fibers
and the number of muscle fibers innervated by one motoneuron increases. Under these
circumstances, there is both an increase in the number of branch points due to the enlarged
motor unit territory and a reduction in the local sink provided by non-active fibers for a rising
interstitial potassium concentration due to grouping of fibers from the same motor unit (see
also Chapters 4-5).
One of the factors that seems to exacerbate impairment of neuromuscular propagation
is ischemia (Chapter 7). Blood flow will be restricted if the intramuscular pressure is
sufficiently high and hence the architecture of the muscle and its vasculature will be relevant.
Muscles vary in the extent to which their blood supply is restricted as muscle force increases
(Chapter 25). Here, thin, parallel-fibered muscles such as the diaphragm and some intrinsic
muscles have an advantage. Intramuscular pressure is being used increasingly to estimate
force in muscles during complex tasks such as walking, when EMG records may be hard to
interpret. However, its role in assessment of the acute changes during fatigue is not yet
established.

AT THE MOTONEURON INTERFACE

In the Sherringtonian sense, the motoneuron is the final common path and the last
site at which any CNS (supraspinal, segmental, sensory feedback) commands to mo-
toneurons can be modified. Conceding the powerful role of segmental interneurons in the
control of motoneuron discharge (Chapter 17), the extent of the neuromodulation of
discharge by factors other than synaptic input is remarkable (Hultborn & Kiehn, 1992; Binder
et aI., 1995). Furthermore, the motoneuron transduces the net input command or current into
a discharge rate that is then translated into the force produced by a single motor unit (Chapters
8-9). These transformations are not linear, although when considered solely in terms of
intrinsic properties and steady-state isometric contractions, the overall relation between input
current to motoneurons in the brainstemlspinal cord and force output by the motor units is
more linear than the component relationships (Binder et aI., 1995). A challenge in the
immediate future is to determine the extent to which these relationships are modified by
synaptic input and other forms of neuromodulation. For example, motoneurons exhibit a
bistable state in which discharge can be superimposed on a membrane potential that is
relatively near or far from (depolarized) the resting level (Hultborn & Kiehn, 1992; Binder
et aI., 1995). The discharge rate is significantly different under these two conditions. If such
bistability can occur during fatiguing contractions, it would confound interpretation of the
relationship between stimulus current-firing frequency-force. Interestingly, such bistability
has been inferred from the EMG firing patterns of single motor units in conscious experi-
mental animals (Kiehn, 1991) but not yet from human data.
It is also necessary to emphasize that the transformation of motoneuron discharge
frequency into motor unit force will also depend on the history of recent stimulation, the
520 s. C. Gandevia et al.
nature of the contraction (shortening vs. isometric vs. lengthening), the range of muscle
lengths, the temperature within the muscle and the relative effects of potentiating and
fatigue-inducing processes within the muscle fibers (Chapter 9). Despite the impressive
progress over the last few years on the interactions between motoneurons and muscle fibers
(Chapters 8 and 9), much remains to be learned about these factors during fatiguing
contractions.
Since its intriguing discussion at the 1980 symposium (Porter & Whelan, 1981),
several chapters in the present volume have addressed the substantial interest in the so-called
muscle wisdom phenomenon. According to this scheme, motoneurons discharge at progres-
sively lower rates during maximal voluntary contractions (Chapters 8, 9, 16-19,26,27). This
produces a convenient matching of speed because the contraction and relaxation times of
the motor units, particularly the fast-twitch ones, slows during maximal voluntary contrac-
tions. The term, muscle wisdom, is obviously a misnomer because if there is any wisdom in
the appropriateness of the motoneuron firing rate, it certainly does not reside in the muscle.
The decline in discharge rate during fatiguing contractions appears to be task dependent. For
example, it is not apparent during submaximal contractions and may not be evident during
eccentric exercise (Chapter 16). Also, the slowing of muscle relaxation and contraction,
arguably a requirement for muscle wisdom, do not occur with all forms of muscle fatigue
under even isometric conditions (Chapter 16).
Binder-Macleod (Chapter 16) emphasizes the need to distinguish between activation
by voluntary command and the artificial wisdom of contractions elicited by imposed
electrical stimulation. With imposed contractions, it is possible to capitalize on the catch-like
property of striated muscle which involves an enhancement of muscle force due to the
inclusion of a brief interpulse interval in the stimulus train. However, the extent to which
catch-like effects are observed under natural circumstances remains an open issue. Nonethe-
less, muscle fatigue remains a major limitation of motor prostheses that use electrical
stimulation to generate force. Further work on variable-frequency stimulus trains is neces-
sary to advance the field of functional electrical stimulation.
The role of the fusimotor system in muscle fatigue is enigmatic. Originally, the system
was conceived as automatically supplying excitation through the discharge of muscle spindle
afferents during fatigue (Merton, 1953; Chapter 18). However, the limited recordings in
conscious humans make this unlikely. In fact, the fusimotor-muscle spindle system appears
to produce less reflex support to motoneuron output during fatigue. Ultimately, the effects
of presynaptic inhibition and other segmental reflexes will need to be characterized before
the full implications of the observations on muscle spindle behavior in human fatigue can
be determined.
Finally, at the interface between the muscle fiber and the CNS, evolution appears to
have introduced a substantial safety margin under most circumstances, particularly for
axonal conduction and, to a lesser degree, for synaptic transmission across the neuromuscular
junction. Consequently, the major adaptations to fatigue, disuse, and disease must be focused
either in the muscle cell or in the segmental and supraspinal machinery.

CNSFACTORS
To what extent is the supraspinal and spinal circuitry concerned with the discharge
of single motoneurons? Apart from insufficient neural circuitry for individual control of
motoneurons, variable force-frequency relationships within and across motoneurons con-
found a rigid strategy that could be imposed by higher centers. These are far from trivial
issues for the CNS, and ones which may well have been solved by the evolution of some
basic strategies to organize a motoneuron pool, such as Henneman's size principle (Binder
Neurobiology of Muscle Fatigue 521

& Mendell, 1991). It is clear that independent control of the discharge rate of each
motoneuron is not a realistic strategy and this is probably why so many properties are
organized systematically from the low- to high-threshold motor units (Henneman, 1991).
Windhorst and Boorman (Chapter 17), describe the progress ofthe last four decades on CNS
control based on intracellular recordings from largely passive (non-firing) motoneurons in
deeply anesthetized, reduced in vivo (largely cat) preparations (see also lankowska, 1992;
McCrea, 1992). Within the past two decades, progress has also been made in understanding
CNS control strategies by: 1) recording the reflex forces of several muscles simultaneously
in less reduced, unanesthetized preparations (Nicholls, 1994); 2) measuring unitary firing
patterns of single motor and sensory axons in freely moving animals (Prochazka, 1995); and
3) resorting to modeling and simulation (e.g., Loeb & Levine, 1989; Fuglevand et aI., 1993;
Binder et aI., 1995).
In view of these developments, it is significant that Chapter 17 emphasizes how little
is known about the segmental and suprasegmental factors that may modify motoneuron
discharge rate during fatigue. This situation appears to be the consequence of several factors.
First, there is a paucity of experiments on segmental mechanisms of fatigue but, also, the
technical difficulty of studying the firing patterns of motoneurons and interneurons with in
in vivo preparations during fatiguing contractions in which force is also measured. While
there have been substantial advances on active firing properties using in vitro slice prepara-
tions (Binder et aI., 1995) these beg the question of force development and its associated
control. Second, there is the failure of investigators in this field to appreciate the full role of
proprioceptive input in the control of muscle contractions and movements (Hasan & Stuart,
1988; Gandevia & Burke, 1992; Prochazka, 1995). Third, although the control of mo-
toneuron discharge is influenced by many species of afferent feedback (Chapters 17-19),
there is insufficient specificity in individual afferent interneuronal pathways to effect
something like a precise decline in discharge rate during a fatiguing contraction.
Studies on CNS factors have underscored thatfatigue encompasses more than simply
the failure to maintain a particular level of force. When attempting to lift a weight that is
half of one's maximum, fatigue can begin at the start of the task (not only when it is
impossible to sustain the contraction) and the sensorium is soon alerted to the need to increase
the motor command to drive the fatiguing muscles. As documented in Chapter 22, this
increased command is manifest as the increased apparent heaviness of the load and the
increased recruitment and discharge frequencies of motoneurons. Between the start of the
lift and the detection of the need to increase the motor command, the altered sensory input
will have modified the state of spinal and supraspinal circuits controlling the muscles. Such
events will occur well before the task, which initially involved a submaximal contraction,
cannot be sustained. For tasks involving submaximal forces, the increase in the motor
command will occur before the force is reduced below the requisite amount. Most subjects
appear to base their judgments of force during fatiguing isometric contractions, in part, on
centrally generated signals (Chapter 22; Gandevia, 1995). More detailed examination of the
underlying mechanisms may be possible with recently developed imaging technology, such
as functional magnetic resonance imaging. This approach might well complement the
information that can be gleaned from studies on the readiness· potential (Chapter 21). These
two techniques have the advantage of being able to examine the behavior of supraspinal
centers in conscious human volunteers.
To what extent does the CNS act to prevent catastrophic changes in the muscle fibers?
This question has attracted interest for over a century, although techniques required to
examine it quantitatively were not devised until Merton's development of twitch interpola-
tion (Merton, 1954; Gandevia et aI., Chapter 20). Careful application of the technique has
revealed first, that voluntary activation of muscles is usually less than maximal even in highly
motivated subjects when a muscle is not fatigued, and second, that voluntary activation
522 s. C. Gandevia et al.
diminishes during exercise, i.e. central fatigue develops. A specific terminology to cover
these findings is presented in Chapter 20. Major differences between muscles and tasks have
been shown for the development of central fatigue (Chapter 20). Central fatigue, unlike many
steps involved in peripheral fatigue, need not scale with the level and duration of force
production. Indeed, they may be inversely related, with peripheral fatigue dominant in brief
high-force tasks and central fatigue more obvious during more prolonged low-intensity
exercise.
Transcranial magnetic stimulation of the motor cortex that is interpolated during
maximal voluntary contractions produces twitch-like increases in force, particularly when
central fatigue develops (Thomas, 1993; Gandevia et aI., 1995). Hence, the output of the
motor cortex and distal sites at the time of the stimulation is not maximal and thus limits
force. This observation suggests that some of the mechanisms underlying the reflex inhibition
of motoneuronal discharge during fatigue may involve supraspinal sites. Given that motor
cortical output to the fatiguing muscle probably increases, these observations focus some
attention on the causes for central fatigue that may reside upstream of the motor cortex. Here,
the interaction with motivation becomes crucial. Much remains to be done to check the
generality of these observations and investigate the supraspinal mechanisms. Already, some
have speculated which central transmitters are involved (Chapter 23, Newsholme & Blom-
strand). Of course, the further one moves from those areas within the CNS that directly
modulate the firing ofmotoneurons, the more difficult it is to establish the cause rather than
the effect of central fatigue.

NEUROMUSCULAR ADAPTATIONS

Fatigue is an acute response of the neuromuscular system to exercise. From a

mechanistic perspective, it is instructive to determine the interactions between chronic
adaptations and fatigue. Such an approach is adopted in Chapters 32-34 with an examination
of some paradigms that increase use, others that reduce use, and some that involve remod-
eling of the neuromuscular system. Studies of the chronic application of electrical stimula-
tion have demonstrated that the fatigability and strength of muscle appear to be controlled
reciprocally (Chapter 32). Frequent stimulation improves fatigue resistance by increasing
the mitochondrial content and oxidative-enzyme capacity of the activated muscle fibers.
However, this regimen decreases muscle-fiber diameter and hence the maximal force that a
muscle can exert; the retention of force capabilities depends on intermittent activation with
high-frequency trains of stimuli. These contrasting effects of the amount and pattern of
activity have also been observed with exercise training and with the models of compensatory
hypertrophy (removal of synergist muscles). However, the adaptive capabilities are limited.
Reinnervation studies (self- and cross-reinnervation) have indicated that reinnervated muscle
fibers retain some of their original phenotype and that the motor unit properties, including
fatigability, are determined by the combined effects of neuromuscular activity and muscle
loading. Little is known about the factors that determine the adaptive limits of muscle fibers,
motor units, and muscles.
The function of the neuromuscular system is impaired with increasing age, the
ultimate chronic adaptation. Many of these changes are attributable to the loss of neurons
and the subsequent reorganization of motor units (Chapter 34). For example, the decline in
muscle mass and strength in elderly persons is largely due to the denervation of muscle fibers
that occurs as motoneurons die. As in other paradigms that induce muscle atrophy and
weakness (Chapter 32), however, the muscles of elderly subjects either exhibit an increased
resistance to fatigue or, at the least, no effect of the chronic adaptation on fatigability (Chapter
34). This effect might be caused by three types of mechanisms: 1) an increased capillary
Neurobiology of Muscle Fatigue 523

density, concentration of mitochondria, or metabolic-enzyme activity due to muscle atrophy;

2) a reduced occlusion of blood flow because the absolute force exerted by the muscle, and
hence the intramuscular pressure, is less for a task where force is normalized as a percentage
of maximum; and 3) an increased proportion of slow-twitch motor units. However, absolute
measures of fatigability, such as the maximal sustainable power, decrease with age, appar-
ently due to reductions in specific force, optimum velocity for power, and energy delivery
and conversion (Chapter 34).
Faulkner and Brooks (Chapter 34) hypothesize that contraction-induced injuries
mediate such adaptations as muscle-fiber denervation with aging. Physical activity, espe-
cially when it involves eccentric contractions, can damage muscle and impair performance
(Chapter 33). The strain experienced by active muscle fibers appears to cause damage that
is perceived as soreness over the subsequent 2-3 days, followed by regeneration of the
damaged fibers. After a bout of eccentric exercise, the magnitude of low-frequency fatigue
is greater than after isometric or concentric contractions. Because motor unit discharge rates
during voluntary contractions occur in this domain, eccentric contractions may also have a
greater effect on the control of movements, as indicated by increased tremor and alterations
of force perception after such activity (Chapter 33).

PATHOPHYSIOLOGY

As highlighted in Chapters 35 and 36, a range of diseases will produce excessive

fatigue (and even no disease at all, merely a sedentary lifestyle) and the clinician's job is to
determine the critical site or sites at which performance is impaired. Objective measurement
of the peripheral and central components to fatigue is possible and has already been adopted
by several laboratories using similar, but not identical, procedures. Given that fatigue has
many task-dependent elements, the use of comparable testing procedures is important.
Results are fairly reproducible and, although their specificity for a particular disease is low,
this is outweighed by the practical information provided to the patient and the physician.
Because of the variability in strength and endurance between individuals, longitudinal
studies of peripheral and central fatigue are often helpful. Commonly, the patient presents
with a story suggesting both peripheral and central components to their fatigue and weakness.
As an example, patients with myasthenia gravis, with a frequency-dependent abnormality at
the neuromuscular junction, may not drive their muscles fully, because, to do so, would result
in excessive force loss. They, and others with disorders in which high motor unit firing rates
lead to excessive fatigue, have developed a longer-term central strategy which has taught
them to adopt low firing rates. In contrast, patients with muscular dystrophy show greater
levels of voluntary activation, and less central fatigue than expected. This, too, is an
appropriate CNS adaptation.
Clinically, absolute muscle strength should also be considered. After all, in many
instances of disease, disability and disuse, small increments in strength may make large
differences to the activities of daily life, such as the capacity to rise from a chair and to walk
unaided. Thus, from a therapeutic viewpoint, much interest surrounds the attempts to unravel
the underlying molecular signaling leading to adaptations within muscle fibers and their
motoneurons.
As Bigland-Ritchie and her colleagues emphasize in Chapter 27, evolution has had
a long time to sort out muscle fatigue. In a particular species, all aspects of muscle
performance have adaptability within some boundaries. The critical limits differ between
species, individuals and between muscles and even their constituent fibers. Ultimately, what
drives the muscles to start and to stop contraction is the CNS and it also changes the behavior
 524 S. C. Gandevia et al.

of reflexes and proprioceptors during fatigue - hence the importance of the phrase in the title
of the chapter "the neurobiology of muscle fatigue".

SUMMARY

Throughout this epilogue, we have emphasized that rapid advances in understanding

of neural and muscular aspects of fatigue have occurred since the 1980 London Symposium.
However, in each instance of progress, from the single muscle fiber to the forebrain, the
application of more precise techniques have raised important new questions. Neuroscientists
and muscle physiologists have expanded opportunities for rigorous study ofa topic of major
scientific and social importance.

ACKNOWLEDGMENTS

Research by the authors is currently supported by: the National Health & Medical
Research Council of Australia, and the Asthma Foundation of New South Wales, Australia
(S.c.G.); United States Public Health Service (USPHS) grants AG 09000 and NS 20544
([Link].); The Medical Research Council and Natural Sciences and Engineering Research
Council (NSERC) of Canada ([Link].); USPHS grants GM 08400, NS 20577, NS 01686
and NS 07309 (D. G.s.); and USPHS grant NS 30226 and the Miami Project to Cure Paralysis
(C.K.T.). The authors' attendance at the 1994 Bigland-Ritchie conference was supported, in
part, by NSERC ([Link].), NS 30226 (C.K. T.), the Cleveland Clinic Foundation ([Link].),
University of Arizona Regents' Professor funds (D. G.S.), the Muscular Dystrophy Associa-
tion (USA; s.c.G.), and the University of Miami (S.c.G. [Link].).

REFERENCES
Ashley CC & Ridgway EB (1970). On the relationships between membrane potential, calcium transient and
tension in single barnacle muscle fiber. Journal ofPhysiology (London) 209, 105-130.
Atlan G,BeJiveau L & Bouissou P (eds). (1991). Muscle Fatigue: Biochemical and Physiological Aspects.
Paris: Masson.
Binder MD, Heckman CJ & Powers RK (1995). The physiological control of motoneuron activity. In: Rowell
LB, Shepherd JT (eds.), Handbook of Physiology, Integration of Motor, Respiratory and Metabolic
Control During Exercise, Sect. A, J. Smith (Sec. Ed.), Neural Control of Movement, pp. 00-00.
Bethesda: American Physiological Society.
Binder MD & Mendell LM (eds.) (1990). The Segmental Motor System. New York: Oxford University Press.
Blinks JR, Rudel R & Taylor SR (1978). Calcium transients in isolated amphibian skeletal muscle fibers:
detection with aqueorin. Journal ofPhysiology (London) 277, 291-323.
Enoka R & Stuart DG (1992). Neurobiology of muscle fatigue. Journal ofApplied Physiology 72, 1631-1648.
Finer JT, Simmons RM & Spudich JA (1994). Single myosin molecule mechanics: piconewton forces and
nanometre steps. Nature 368, 113-119.
Fuglevand AJ, Winter DA & Patla AE (1993). Models of recruitment and rate coding organization in motor-unit
pools. Journal ofNeurophysiology 70, 2470-2488.
Gandevia SC (1995). Kinesthesia: roles for afferent signals and motor commands. In: Rowell LB, Shepherd
JT (eds.), Handbook on Integration ofMotor, Circulatory, Respiratory and Metabolic Control during
Exercise Sect. A, J. Smith (Sec. Ed.), Neural Control of Movement, pp. 00-00. Bethesda: American
Physiological Society.
Gandevia SC, Allen GM, Butler JE & Taylor JL (1995). Supraspinal factors in human muscle fatigue: evidence
for suboptimal output from the motor cortex. Journal of Physiology (London) In press.
Gandevia SC & Burke D. (1992). Does the nervous system depend on kinesthetic information to control natural
limb movements? Behavioral and Brain Sciences 15, 614-632.
Neurobiology of Muscle Fatigue 525

Gonzalez-Serratos H, Garcia M, Somlyo A, Somlyo AP & McClellan G (1981). Differential shortening of

myofibrils during development of fatigue. Biophysical Journal 33, 224a.
Hasan Z & Stuart DG (1988). Animal solutions to problems of movement control: the role ofproprioceptors.
Annual Review of Neuroscience 11,199-223.
Henneman E (1990). Henneman's contributions in historical perspective. In: Binder MD, Mendell, LM (eds.),
The Segmental Motor System, pp. 3-19. New York: Oxford University Press.
Hultborn H & Kiehn 0 (1992). Neuromodulation of vertebrate motor neuron membrane properties. Current
Opinion in Neurobiology 2, 770-775.
Jankowska E (1992). Interneuronal relay in spinal pathways from proprioceptors. Progress in Neurobiology
38,335-378.
Kiehn 0 (1991). Plateau potentials and active integration in the '[mal common pathway' for motoneuron
behaviour. Trends in Neurosciences 14, 68-73.
Loeb GE, He J & Levine WS (1989). Spinal cord circuits: are they mirrors of musculoskeletal mechanics?
Journal ofMotor Behavior 21,473-491.
McComas AJ (1995). Fatigue. Chapter 15. In: Skeletal Muscle: Form and Function, pp. 00-00. Champaign
(IL): Human Kinetics.
McCrea DA (1992). Can sense be made of spinal interneuron circuits? Behavioral and Brain Sciences 15,
633-643.
Merton PA (1953). Speculations on the servo-control of movement. In: Wolstenholme GEW (ed.), The Spinal
Cord, pp. 247-260. London: Churchill.
Merton PA (1954). Voluntary strength and fatigue. Journal of Physiology (London) 123,553-564.
Nicholls TR (1994). A biomechanical perspective on spinal mechanisms of co-ordinated muscle action: an
architecture principle. Acta Anatomica 151, 1-13.
Porter R & Whelan J (1981). Human Muscle Fatigue: Physiological Mechanisms (CIBA Foundation Sympo-
sium 82). London: Pitman Medical.
Prochazka A (1995). The physiological control of motoneuron activity. In: Rowell LB, Shepherd JT (eds.),
Handbook Of Physiology, Integration ofMotor, Respiratory and Metabolic Control during Exercise,
Sect. A, J. Smith (Sec. Ed.), Neural Control ofMovement, pp. 00-00. Bethesda: American Physiologi-
cal Society.
Rayment I, Holden HM, Whittaker M, Yohn CB, Lorenz M, Holmes KC & Mulligan RA (1992). Structure of
the actin-myosin complex and its implications for muscle contraction. Science 261,58-65.
Rome LC (1993). The design of the neuromuscular system. In: Sargeant AJ, Kernell D (eds.), Neuromuscular
Fatigue, pp. 129-136. Amsterdam: Royal Netherlands Academy of Arts and Sciences.
Sargeant AJ & Kernell D (1993). Neuromuscular Fatigue. Amsterdam: Royal Netherlands Academy of Arts
and Sciences.
Stuart DG, Binder MD & Enoka RM (1984). Motor unit organization: Application of the quadripartite
classification scheme to human muscles. In: Dyck PJ, Thomas PK, Lambert EH, Bunge R (eds.),
Peripheral Neuropathy, pp. 1067-1090. Philadelphia: W.B. Saunders Co.
Stuart DG & Callister RJ (1993). Afferent and spinal reflex aspects of muscle fatigue: issues and speculations.
In: Sargeant AJ, Kernell D (eds.), Neuromuscular Fatigue, pp. 169-180. Amsterdam: Royal Nether-
lands Academy of Arts and Sciences.
Thomas CK (1993). Muscle fatigue after incomplete human cervical spinal cord injury. Journal of the
American Paraplegia Society 16,87.
CONTRIBUTORS

[Link] B. Bigland-Ritchie
Department ofPhysiology Department ofPediatrics
University of Sydney Yale University
NSW2006 P.O. Box 208-064
Australia New Haven, Connecticut 06520

[Link] M.D. Binder

Prince of Wales Medical Research Institute Department ofPhysiology & Biophysics,
High Street, Randwick SJ-40
Sydney NSW 2031 University of Washington
Australia School of Medicine
Seattle, Washington 98195
M.A. Babcock
The John Rankin Laboratory of Pulmonary S.A. Binder-Macleod
Medicine Department ofPhysical Therapy
University of Wisconsin-Madison University of Delaware
Department of Preventive Medicine 315 McKinly Laboratory
Madison, Wisconsin 53705-2368 Newark, Delaware 19716

F. Bellemare E. Blomstrand
Departement d 'Anesthesie Division of Technology and Development
Hotel-Dieu de Montreal Pripps Brewery
3840 St. Urbain Bryggerivagen 10
Montreal, Quebec S-161 86 Bromma
Canada Sweden

L. Bertocci [Link]
Institute for Exercise and Environmental Departments of Clinical Neurosciences
Medicine and Medical Physiology
Presbyterian Hospital of Dallas The University of Calgary
7232 Greenville Avenue Faculty of Medicine
Dallas, Texas 75231-5129 3330 Hospital Drive N.W.
Calgary, Alberta T2N 4Nl
Canada

527
528 Contributors

B.R. Botterman R.B. T. Edwards

Department of Cell Biology and Muscle Research Centre
Neuroscience Department of Medicine
University of Texas Southwestern Medical University of Liverpool
Center P.o. Box 147
Dallas, Texas 75235 Liverpool L69 3BX
United Kingdom
S.v. Brooks
Institute of Gerontology and J.M. Elek
Bioengineering Program Department ofNeurology and Clinical
University of Michigan Neurophysiology
300 N. Ingalls Medical School of Hannover
Ann Arbor, Michigan 48109-2007 D-30623 Hannover
Germany
P.M. Clarkson
Department of Exercise Science [Link]
Boyden Building, Box 31020 Department of Biomedical Engineering
University of Massachusetts The Cleveland Clinic Foundation
Amherst, Massachusetts 01003 9500 Euclid Avenue
Cleveland, Ohio 44195-5254
J.A. Dempsey
The John Rankin Laboratory of Pulmonary J.A. Faulkner
Medicine Institute of Gerontology and
University of Wisconsin-Madison Bioengineering Program
Department of Preventive Medicine University of Michigan
Madison, Wisconsin 53705-2368 300 N. Ingalls
Ann Arbor, Michigan 48109-2007
R. Dengler
Department of Neurology and Clinical [Link]
Neurophysiology School ofPhysiology and Pharmacology
Medical School of Hannover University of New South Wales
D-30623 Hannover Kensington, NSW 2033
Germany Australia

J. Dushanova A.J. Fuglevand

Institute ofPhysiology The John B. Pierce Laboratory
Bulgarian Academy of Sciences 290 Congress Ave.
Acad. G. Bonchev Str., BId. 23 New Haven, Connecticut 06519-1403
1113 Sofia
Bulgaria S.C. Gandevia
Prince of Wales Medical Research Institute
K.A.P. Edman High Street, Randwick
Department ofPharmacology Sydney NSW 2031
University of Lund Australia
Solvegatan 10
S-223 62 Lund S.J. Garland
Sweden Department ofPhysical Therapy
The University of Western Ontario
Elborn College
London Ontario N6C IHI
Canada
Contributors 529

H. Gibson J.A. Kent-Braun

Muscle Research Centre UCSF Medical Center
Department of Medicine Veterans Administration
University of Liverpool 3150 Clement Street (11M)
P.O. Box 147 San Francisco, California 94121-1676
Liverpool L69 3BX
United Kingdom D. Kernell
Department ofMedical Physiology
T. Gordon University of Groningen
Department ofPharmacology Bloemsingel 10
Division of Neurosciences 9712 KZ Groningen
525 Heritage Medical Research Centre The Netherlands
University of Alberta
Edmonton T6G 2S2 M.J. Kushmerick
Canada NMR Research Laboratory
Department of Radiology, SB-05
K-E. Hagbarth University of Washington
Department of Clinical Neurophysiology Seattle, Washington 98195
University Hospital
S-751 85 Uppsala [Link]
Sweden School of Zoology
La Trobe University
A.R. Hargens Bundoora
Life Science Division Victoria 3083
NASA Ames Research Center Australia
Moffett Field, California 94035-1000
J. Lannergren
H. Hoppeler Department ofPhysiology and
Department ofAnatomy Pharmacology
University of Bern Karolinska Institutet
CH-3000 Bern S-171 77 Stockholm
Switzerland Sweden

D.A. Jones S.L. Lindstedt

School of Sport and Exercise Sciences Department of Biological Sciences
The University of Birmingham Northern Arizona University
Edgbaston, Birmingham B 15 2TT Flagstaff, Arizona 86011-5640
United Kingdom
v.G. Macefield
L.A. Jones Prince of Wales Medical Research Institute
Department ofMechanical Engineering High Street, Randwick
77 Massachusetts Ave Sydney 2031
Massachusetts Institute of Technology Australia
Cambridge, Massachusetts 02139-4307
A.J. McComas
M.P. Kaufman Department ofBiomedical Sciences
Division of Cardiovascular Medicine McMaster Health Science Centre
Departments of Internal Medicine and 1200 Main Street West
Human Physiology Hamilton
University of California Ontario L8N 3Z5
Davis, California 95616-8636 Canada
530 Contributors

D.K. McKenzie D. Popivanov

Department of Respiratory Medicine Institute ofPhysiology
Prince of Wales Hospital Bulgarian Academy of Sciences
High Street, Randwick Acad. G. Bonchev Str., Bid. 23
Sydney NSW 2031 1113 Sofia
Australia Bulgaria

T.S. Miles R.K. Powers

Department of Physiology Department of Physiology & Biophysics,
University of Adelaide SJ-40
Adelaide SA 5005 University of Washington
Australia School of Medicine
Seattle, Washington 98195
R.G. Miller
Department of Neurology Y.S. Prakash
California Pacific Medical Center Departments ofAnesthesiology &
3698 California St., Rm 545 Physiology and Biophysics
San Francisco, California 94118-1702 Mayo Clinic and Foundation
Rochester, Minnesota 55905
A. Mineva
Institute ofPhysiology c.L. Rice
Bulgarian Academy of Sciences School of Physical & Occupational
Acad. G. Bonchev Str., BId. 23 Therapy and Dept. ofAnatomy
1113 Sofia McGill University
Bulgaria Montreal, Quebec
Canada
D.J. Newham
Physiotherapy Group A.J. Sargeant
Biomedical Sciences Division Department of Muscle and Exercise
King's College Physiology
London, England SE5 9RS Faculty of Human Movement Sciences
United Kingdom Vrije University
Van der Boechorstraat 9
E. Newsholme 1081 BT Amsterdam
Department of Biochemistry The Netherlands
University of Oxford
South Parks Rd A. Sawczuk
Oxford, OXI 3QV Department of Oral Biology
United Kingdom University of Washington
Seattle, Washington 98195
M.A. Nordstrom
Department of Physiology O.M. Sejersted
University of Adelaide Institute for Experimental Medical
Adelaide SA 5005 Research
Australia The University of Oslo
Ullevaal Hospital
N-0407 Oslo
Norway
Contributors 531

K.R. Sharma V. Toescu

Department of Neurology Muscle Research Centre
University of Miami Department of Medicine
Miami, Florida 33136 University of Liverpool
P.O. Box 147
G.c. Sieck Liverpool L69 3BX
Departments ofAnesthesiology & United Kingdom
Physiology and Biophysics
Mayo Clinic and Foundation J. V. Trontelj
Rochester, Minnesota 55905 University Institute of Clinical
Neurophysiology
G. Sjogaard University Medical Center of Ljubljana
National Institute of Occupational Health 61105 Ljubljana, Zaloska 7
Lers0 Parkalle 105 Slovenia
DK-2100 Copenhagen 0
Denmark N.K. Vollestad
Department ofPhysiology
E. Stalberg National Institute of Occupational Health
Department of Clinical Neurophysiology Box 8149 Dep
University Hospital N-0033 Oslo
Uppsala Norway
Sweden
M.L. Walsh
D.G. Stephenson Department ofPediatrics
School of Zoology Yale University
La Trobe University P.O. Box 208-064
Bundoora New Haven, Connecticut 06520
Victoria 3083
Australia M.W. Weiner
Magnetic Resonance Unit
G.M.M. Stephenson VA Medical Center
Department of Chemistry and Biology Departments of Medicine, Radiology and
Victoria University of Technology Psychiatry
Footscray, Victoria 3011 University of California
Australia San Francisco, California 94121

D.G. Stuart H. Westerblad

Department ofPhysiology Department of Physiology and
University of Arizona Pharmacology
College of Medicine Karolinska Institutet
Tucson, Arizona 85724-5051 S-171 77 Stockholm
Sweden
C.K. Thomas
The Miami Project to Cure Paralysis U. Windhorst
University of Miami School of Medicine Departments of Clinical Neurosciences
1600 NW 10th Avenue, R-48 and Medical Physiology
Miami, Florida 33136-1015 The University of Calgary
Faculty of Medicine
3330 Hospital Drive N.W.
Calgary, Alberta T2N 4Nl
Canada
INDEX

Acetylcholine Adenosinetriphosphate (ATP), 50, 65, 178-179,

receptors, 94-95, 113, 507 186-189, 197-199,363-364,430-436,
release, 92-94 458-459,508---509
sensitivity, 94--95 Aging
Acidosis, extracellular, 397, 406 decline in power, 475-477
Acidosis, intracellular decline in specific force, 471-472
changes during fatigue, 38, 57-66,18&-190, decline in strength, 477-478
199-202,363 fatigability,474-475
consequences, 38,49-51,61-65,95-96 muscle damage, 473-474
regulation, 58-60
Action potential muscle fiber denervation, 472-473
axon muscle fiberJoss, 471
branch point failure, 89-91 training effects, 478
motoneuron; see also Motoneuron Airway narrowing, 395, 402, 407-409
conductance changes, 124--128 Amino acids, branched chain
current-frequency relationship, 125-126, effects on mental performance, 316
138-139 role in fatigue, 315-319
force-frequency relationship, 13&-137,229- Amyotrophic lateral sclerosis (ALS), 148, 165-
231,355 166,500-501
repetitive firing, 125-130, 138---139 Arachidonic acid, 276
spike-frequency adaptation, 126-130, 139, Axon; see also Microneurography, human; Intra-
246-247 neural stimulation, motor axons
muscle fiber branch-point failure, 89-91, 110
effect of depolarization, 46-47, 104--108
in patients, 505-509 Biochemical basis of fatigue; see also Magnetic
propagation velocity, 115--116 resonance imaging/spectroscopy; Fatigue,
singlefiberEMGstudies, 101-106, 109-118 failure sites.
t-tubule, 40, 46-49, 76; see also Excitation-con- dependence on exercise intensity, 186-188
traction coupling historical perspective, 487-488
whole muscle, (M-wave) role of hydrogen ions, inorganic phosphate, 37-
in fatigue, 15,74,87-91, 101-102,203-205, 38, 197-202
366-367,406 studies in patients, 203-208, 487-488, 508-509
in patients, 505-507 Biochemical capacitor
Activation, contractile machinery chemical energy concept, 179
failure in fatigue, 39-41, 63 Blood flow, skeletal muscle; see also Failure sites,
Adaptation, neuromuscular; see also Plasticity; fatigue
Training diaphragm, 403
cross vs. self-reinnervation, 443-445 muscle pump, 346
disuse effects, 445-448 relation to intramuscular pressure, 346-348,
fast-to-slow conversion, 432-437 365-366
slow-to-fast conversion, 437 respiratory muscle, 397-398, 403
Adenosinediphosphate (ADP), 63, 186, 197-198, Starling resistor, 347
435 vs. potassium washout, 74

533
534 Index

Bradykinin, 276, 461--462 Diaphragm muscle

Brenda Bigland-Ritchie's scientific contributions, central fatigue, 289
1-8,11-21 factors involved in fatigue, 289, 375-376, 394-
399,403--404
Caffeine muscle fiber types, 84-85
effects in myofibrillar fatigue, 30-32 neuromuscular transmission, 85-86
Calcium; see also Excitation-contraction cou- phrenic nerve stimulation, 396-397, 405--406
pling Disorders
effects on neuromuscular transmission, 93 central nervous system
Calcium binding, 53 amyotrophic lateral sclerosis (ALS), 148,
165-166,500-501
Calcium channels (receptors) chronic fatigue syndrome, 203, 485, 502-
dihydropyridine (DHP), 47--48 505
ryanodine, 47-50 hemiplegia, 170
Calcium release, within muscle fiber, 39--41, 47- multiple sclerosis, 501-502
53, 63, 365; see also Fatigue, failure sites, post-polio syndrome, 499-500
excitation-contraction coupling spinal cord injury, 148,504
Calmodulin, 51 spinal muscular atrophy, 165-166,505-506
Calcium sensitivity peripheral nervous system
myofibrillar, 63-64 denervation effects, 505-507
Camitine deficiency, 509 familial periodic paralysis, 507-508
Catch-like property, muscle fibers hereditary motor and sensory neuropathy,
effect on force, 137,232-237,355 165-166
historical perspective, 232 Lambert-Eaton syndrome, 115,507
significance in fatigue, 234-237 McArdle's disease, 490, 508-509
I3C MRS: see Magnetic resonance spectroscopy mitochondrial disorders, 241, 488, 509
Central drive myasthenia gravis, 113-114, 483, 507, 523
from motor cortex, 281-289, 410, 522 myotonic disorders, 116-117,507-508
interaction with segmental reflexes, 244 neuromuscular junction disorders, 486--487,
role of motivation, 373 507
Central fatigue Disuse; see also Muscle disuse
decline in motor unit firing rates, 287-289 experimental models, 445--447
experimental treatment, 315-319 muscle fiber changes, 445--448
historical perspective, 13--14,281-282,485
5-hydroxytryptamine levels in brain, 316-317
neural mechanisms, 289 Eccentric (lengthening) contraction; see also Mus-
in patients, 496--497 cle contraction type
readiness potential, 291, 296 ATP splitting, 458
respiratory, 404--407 early studies, 3, 12,459
role of branched-chain amino acids, 316-319 fatigability, 463
role of perception, 305-311 force generation, 462--464
task dependence, 373-376, 405--407 muscle damage, 460--462
variability, 289 muscle temperature, 459--460
Chemical potential of ATP oxygen consumption, 459
energy for muscle power, 181 proprioception, 465--466
Chronic fatigue syndrome, 203, 485, 502-504 serum levels of muscle proteins, 460--461
Compartment syndromes soreness, 461--462
intramuscular swelling, 347-348 stiffness, 464--465
Conduction velocity, muscle fiber water uptake, 461
changes in fatigue, 95-96, 115 Elastic energy, 328
in myotonia, 126-127 Electromyography (EMG); see also Methods of
normal values, 115 measurement
Contractile slowing in fatigue, 32-36, 64-65, 139, fatigue induced alterations, 16-17, 101-106,
150-151,201,228-231,246-247,355, 205,272-273,371-373
368-369 in patients, 486--487, 498
Contracture, 30, 39, 104-106 jaw muscle, 419
Creatine kinase, 179,460--461 power spectral analysis, 419
Current-frequency relationship, motoneuron, 125- single fiber, 101-106, 109-119
126, 138-139; see also Motoneuron vs. intramuscular pressure, 345-346
Index 535

Endplate, motor; see also Neuromuscular junction Fatigue (cant.)

acetylcholine receptors, 94-95, 113, 507 investigation of patients, 481-492, 495-509
endplate potential, 87, 96, III low-frequency fatigue, 50, 205-208, 230, 354,
miniature endplate potential, 92-94 365,463
Energy balance magnetic resonance spectroscopy studies, 176-
concept, 178, 197-202 178,186-190,195-208,211-221,499
in patients, 203, 508--509 recovery, 198--202, 489-491
role of ATP and PCr, 178--180, 186-190, 197- special muscles
202 diaphragm, 86, 367, 396-397, 403-404
whole-body exercise, 190-191 jaw, 416-423
Excitation-contraction coupling; see also T-system phonatory, 423
acidosis effects, 61--64 symptoms in patients, 495-496
caffeine effects, 30-32 task dependence, 188--192,205-208,236-237,
failure, 15,39-41,61--64,205-208,364-365, 323-335,339---348,351-358,361-376,
488--489,517 405-407
interstitial K+ effects, 74-75 Feedback regulation
normal sequence of events, 45-53 ATP synthesis, 179---180
relation to membrane potential, 105-106 force, 244-246, 260-263, 273-276
role ofCa2+, 39, 49---53, 61 Feedforward regulation
single fiber studies, 30-31, 45-53 ATP synthesis, 179---180
skinned fiber studies, 45-53 Force-fatigability relationship, 357-358
Exercise; see Physical activity Force-frequency relationship, 5, 17-18, 130, 136-
137,153-155,287-288,327,355-357,
Familial periodic paralysis, 507-508 370
Fatigue Force potentiation
in aging, 474-475 motor units, 136, 154-155, 232, 242, 355
biochemical changes 37-38, 57, 59, 69, 74,175- muscle, 235
181,186-190,195-208,211-221,315- Force-stiffness relationship
319,363-364,487-488,508--509 myofibrillar, 41
contractile slowing: see Contractile slowing Force-velocity relationship; see also Velocity of
definitions, 29, 83, 135, 186, 199,242,260- shortening
261,282-283,323-335,361-362,418-- muscle, 403, 458--459
419,491-492 muscle fiber, 3
differences in susceptibility of different muscle in myofibrillar fatigue, 32-38
fiber and motor unit types, 84-86, 96-97,
242-244,324-326,351-354,420-423, Glycogen
431-432 breakdown in fatigue, 65, 191-192,201,214-
failure sites 215,409,432
axon, 89---92, 11 0 breakdown in type I vs. type II fibers, 84-85,
CNS: see Central fatigue 192,432
excitation-contraction coupling, 15, 39-41, Golgi tendon organ; see also Group Ib afferents
61--64,76,205-208,364-365,488--489 in fatigue, 251-252, 275
muscle plasmalemma (sarcolemma), 95-96, as force transducer, 248, 272
101-108,507-508 Group Ia afferents; see also Muscle spindle
myofibril, 29---39, 63 axon properties, 244-248, 272
neuromuscular junction, 15,83-97, 110-115, in fatigue, 252, 263-264, 271-276
366-368,486-487,507,519 feedback mechanism, 245-248, 260-263, 273-
force 274
motor unit, 88--89, 135-145, 149, 154-155, follow-up servo theory, 260
170,227-237,242-244,324,351-358, in locomotion, 253-254
420-423 reciprocal inhibitory action, 252-253
muscle, 83, 195-208,235-236,283-285, servo assistance theory, 260-263
307-311,330-331,342-343,363-370, Group Ib afferents; see also Golgi tendon organ
404-409,418--420,436-448,462-464, axon properties, 244-248, 272
473-478,498 different synaptic effects, 251-252
muscle fiber, 39-41, 60--63, 325-327 in fatigue, 248, 275
myofibrillar, 30-39, 63--65 force feedback, 246, 251-252
high-frequency fatigue, 356 inhibitory intemeurons, 249
historical perspective, 481-492 locomotion, 253-254
536 Index

Group III -IV afferents Jitter (in SFEMG), 110-118

axon properties, 272
effects on fusimotor neurons, 266, 289 39K MRS: see Magnetic resonance spectroscopy
endings, 272 Kronecker's laws of fatigue, 482
in fatigue, 76--77, 248-249, 265-266, 275, 372
Lactic acid, 57, 187-190, 249, 275, 487; see also
IH MRI: see [Link] resonance imaging Acidosis, intracellular
Heart rate, at Vo max, 387, 394 Lambert-Eaton syndrome, 115,507
Hereditary neuro'pathies Locomotion
motor and sensory, 165-166 pattern generating circuit, 253-254
High-frequency fatigue, 356 Low frequency fatigue, 50, 205-208, 230, 354,
Hydrogen ion concentration (pH); see also Acidosis 365,406,463
changes in fatigue, 58-60,188-189
effects on Ca2+ release, 50-51, 61 Magnesium
effects on Ca2 + sensitivity of myofibrils, 63 calcium release, 49-50
effects on Ca2 + uptake, 64 Magnetic resonance imaging (MRI)
effects on glycogen breakdown, 65 IH mechanisms, 218-220, 462
effects on single fiber force, 37-38, 61-63 IH methodology, 218-221
investigation by magnetic resonance imag- cine, 220-221
ing/spectroscopy, 198-202,218-220 Tl and T2 relaxation, 218-219, 518
regulation, 58-59 Magnetic resonance spectroscopy (MRS)
relationship to muscle force, 57-66, 95-96, 199- advantages vs. limitations, 186--187, 196--197,
202 212
stimulation of group III-IV afferents, 248-249, l3C MRS, 212-215
265-266, 275-276 in central fatigue, 202-205
studies in patients, 507 combined with EMG and force, 199-208
Hysteresis, in force-frequency relationship, 355 damaged muscles, 462
isotopomer analysis, 212
Inorganic phosphate (Pi) 39K MRS, 215-218
role in fatigue, 51, 176--184, 186--192, 197-207, quantum multiple filler, 217
363,462,500 shift reagent, 216
Inositol 1,4,5- triphosphate (IP 3), 48 TmDOTP, 216
Interneurons, spinal 23Na MRS, 215-218
Renshaw cells, 249-251, 275 neuromuscular disorders, 205-208, 505-507,
role in fatigue, 241-255, 275-276 586--587
Interpolated twitch: see Methods of stimulation in pathophysiology, 201-208, 499
Intramuscular fluid pressure (IMP) 31pNMR, 176--177, 186--189, 195-208,211-
capillary filtration, 340 212,363,462,503,517
cyclic variations, 334 Masticatory muscles
determining factors, 340-341 fatigue, 419-423
effect of contraction, 343-344 fiber types, 416-417
Laplace's law, 340-341 motor unit types, 417-418
lymph formation, 340 McArdle's disease, 490, 508-509
methods of measurement, 341-342 Mechanical efficiency
myalgia, 348 human muscle fiber differences, 175-176, 333-
relation to muscle blood flow, 346--347, 365-366 335
Starling forces, 341 Membrane potential; see also Muscle Fiber
task dependency of, 333-334 activation oft-tubular voltage sensor, 47-48,
vs. EMG, 345-346 104
Intraneural stimulation, motor axon,s relation to developed force, 104-105
force-frequency relationship, 153 resting
general considerations, 152, 156--157 muscle fiber, 74-75
in humans, 147, 161 relation to potassium, 104
value in fatigue studies, 154-155 Metabolism
Isotopomer analysis; see also Magnetic resonance effects of acidosis, 65-66
imaging muscle energy, 175-184, 185-192, 197-199,
glutamate, 212 212-213,432
non-steady state, 212 Methods of analysis
steady state, 212 chaotic dynamics, 296
Index 537

Methods of analysis (cont.) Microneurography, human; see also Methods of

correlation dimension, 297-298 measurement
non-linear filtering, 297 motor axons, 20, 152-153,262-263,288--289,
non-linear prediction, 299 369
point D2 dimension, 297 sensory axons, 263-264, 288
power spectral analysis, 419
signal cancellation, 101 Microstimulation; see also Methods of measure-
singular spectrum analysis, 297 ment; microneurography
spike-triggered averaging (STA) intramuscular, 161-170
applications, 148--149, 161 intraneural, 152-157
limitations, 149-151, 161-162 Mitochondria
Methods of measurement disorders, 509
biochemical microvascular oxygen delivery, 388--390
brain, 315-319 oxygen sink, 389-390
muscle, 69, 175-181, 185-192, 195-208, Motoneuron
211-221,324-326 alpha (skeletomotor)
muscle fiber, 45,57, 175-181,325-326 action potential, 124-127
blood flow, 365-366, 394, 403404 afierhyperpolarization, 124-125, 138--139,
electromyography (EMG) 247-248
motor unit, 16-18,87--89, 101-106, 147- discharge during fatigue, 130, 139-140,246-
151,287-288,353,420 247,262-263,288--289,354-357,370-372
muscle, 16, 101 effect of muscle vibration, 262
muscle fiber, 101-108, 109 intrinsic properties, 123-128, 138--139,247-
single-fiber (SFEMG), 109-118 248
extracellular recording matching to motor-unit properties, 123-124,
motor axon, 81 140-142,228--229,242-244
muscle nerve, 83 postsynaptic EPSPs and IPSPs, 241-255
nerve axon, 83 presynaptic inhibition, 252
sensory axon, 248--249, 271-276 reciprocal Ia inhibition, 252-253
force: see Fatigue, force recruitment order, 123-125, 135,242-244,
intracellular recording 327,358
motoneuron, 123-130, 135-143,241-255 reflex excitation and inhibition, 241-255,
muscle fiber, 57, 71-73, 95, 104-105,462 265-266,272-276,288--289
intramuscular fluid pressure (IMP), 342-343 spike-frequency adaptation, 126-130, 139,
joint torque, 285 246-247,273
magnetic resonance spectroscopy, 69, 176-177, stimulus-current to spike-frequency transduc-
186-192,195-208,211-218,499 tion, 125-126, 138--139
microneurography volitional discharge rate, 262
human axons, 147,262-264,288,369 gamma (fusimotor)
oxygen consumption and transport, 187-192, discharge during fatigue, 263-264
385-390,393-399 Motor cortex; see also Readiness potential
oxygen uptake descending command signals, 283
mitochondria, 178, 388--390 recordings, 283, 289
power, human muscle stimulation 284, 522
maximum, vs. fiber type, 330-331 Motor unit; see also Motor unit recruitment order;
submaximal, vs. fiber type, 331-333 Motor unit type
psychophysical testing, 283-285, 305-311 EMG: see Methods of measurement, elec-
senses of effort, fatigue, force, 281-285, 305- tromyography, motor unit
311, 483-485 force: see Methods of measurement, force, mo-
sensory threshold, 306, 308 tor unit
Methods of stimulation human
artificial wisdom, 228 effects offatigue, 149, 150, 154-156, 169,
electrical 371,420-423
interpolated twitch technique, 14,283-285, twitch parameters, 151, 153, 163-167,420
373-374,406 innervation ratio, age-related decrease, 472-473
motor unit, 123, 135, 152-157, 161-170, potentiation, 154
227-237,353-354 remodeling; see also Plasticity
muscle, 69, 83, 235-236, 283-285, 432-436 age-related, 472-473
muscle fiber, 29, 57 to overuse and underuse, 432-438
538 Index

Motor unit recruitment order; see also Mo- Muscle endurance

toneurons, alpha, recruitment order dynamic vs. static exercise, 438-441
in dynamic exercise, 329-333 training effects, 438-442, 478
vs. exercise level, 148, 187-192 Muscle force: see Methods of measurement, force,
in neuromuscular disorders, 148 muscle
in normal muscle, 123-125, 135, 148,242-244 Muscle fiber
Motor unit type effects of aging, 471-478
differences in fatigability, 149, 154-155,230- EMG: see Methods of measurement, elec-
231,242-244,352-354 tromyography, muscle fiber
histochemical and physiological differences, fast-to-slow conversion, 432-434
325-326,352-358,416-417,430-432 fiber type conversion, 432-448
significance of differences, 323-335, 416-417 force; see also Methods of measurement, force,
Multiple sclerosis, 501-502 muscle fiber
Muscle during fatigue, 61
biopsy technique, 186 maximum calcium-activated, 63
compound action potential, 74, 87-89, 101-102, relaxation rate during fatigue, 64
203-205,366-367,406,507-508 shortening velocity during fatigue, 65
experimental overload, 442-443 membrane potential
glycogen depletion, 84 resting, 75
group interactions vs. task, 374-376 vs. developed force, 104-105
microstimulation, 161-170 role of extracellular potassium, 74, 102-
plasticity, 181,429-448,472-473 104
Muscle afferents; see also Group Ia afferents; role of sodium/potassium pump, 73
Group Ib afferents; Group III-IV afferents metabolism, fast vs. slow types, 187-188,430-
human studies, 259-267, 272-274, 289 432
non-human studies, 241-255, 274-276 sarcolemma excitability, 95
morphology, 272 SFEMG, 109-118
reflexes, 241-255, 260-263,271-276 shortening velocity during fatigue, 65
Muscle atrophy slow-to-fast adaptation, 437-438
aging effects, 477-478 type
underuse effects, 445-448 bioenergetic differences, 175-178
Muscle contraction type contractile properties and histochemistry,
anisometric, 236-237 325-326
concentric (shortening), 328-329 diaphragm, 84-85
eccentric (lengthening), 3, 328, 458-460 usage-dependent control of expression, 326
isometric, 328 Muscle fiber action potential; see also Sar-
stretch shortening, 329-330 colemma
Muscle damage depolarizationirepolarization disturbances, 117-
aged muscles, 473-474 118
creatine kinase, 460-461 fatigue-related changes, 74-75, 101-106,203-
force generation and power output, 462-464 205
histology and ultrastructure, 460 fiber-type and species comparisons, 104-105
magnetic resonance imaging/spectroscopy, 462 propagation velocity, 95, 115-118
neuromuscular dysfunction, 465-466 Muscle length
range of motion and stiffness, 464-465 effect on fatigue, 403
Muscle denervation effect on intramuscular pressure, 343-344
aged muscles, 472-473 vs. motoneuron firing rate, 141
effects on fatigue, 443-445, 474-475, 505- relationship to tension, 403
506 sense of force, 306
Muscle development Muscle power
neuromuscular transmission, 91, 94, 96 decline in aging, 475-477
Muscle disuse in eccentric contractions, 328, 464, 475-477
damage resulting from, 448 in fatigue, 32, 370
effects on fatigue, 85-86, 445-448, 497 vs. fiber type, 330-333
effects on fiber types, 445-448 maximum, 330-331, 475
hindlimb suspension and space flight, 445-446 submaximal, 331-333
limb immobilization, 446-447 sustained,331-333,475-477
pharmacological paralysis, 447 velocity of shortening, 324-325
spinal cord transection, 86, 445 Muscle soreness, 461-462, 473-474
Index 539

Muscle spindle Oxygen consumption, 383-390, 393--399

activity during muscle contractions, 248, 26G- adaptive variation. 385
264 allometric variation, 384-385
amplification of skeletomotor output, 262-263 in exercise, 187-192,383--384,459
discharge-rate during fatigue, 248, 263--264, vs. exercise level, 187-192,393-397
274-275,289 Oxygen diffusion
thixotropic contractile after-effects, 266--267 lung capillaries, 385-386, 394-395
Muscle temperature muscle capillaries, 388-389
in eccentric contractions, 459--460 Oxygen transport vs. exercise level, 384
effect on contractile properties, 137,324 Oxygen uptake
effect on shortening velocity, 335 by lungs, 385-388
in fatigue, 189-190 by mitochondria, 388-390
vs. motoneuron firing rate, 141
Muscle unit; see also Motor unit 31p NMR: see Magnetic resonance spectroscopy
matching to motoneuron properties, 135-143 Pathobiology of muscle fatigue; see also Disor-
potentiation and fatigue, 135-136 ders, central nervous system; Disorders,
spike-frequency to force transduction, 136--137 peripheral nervous system
Muscle weakness historical perspective, 481-492
in aging, 477-478 patient evaluation, 203--208, 491-492, 495-499
in disorders, 202-203, 495-496 Perception; see also Sense of force
and perceived effort and force, 31 G-311 dissociation between effort and force, 309-310
Muscle wisdom sense of effort, 309-310
historical perspective, 228-229 historical perspective, 483-484
mechanisms, 77, 130, 142,246--249,261-262, sense of fatigue, historical perspective, 483
272-273,372 sense of force, 307-311
phenomenon, 45, 130, 142,228-229,246--247, Phosphocreatine, 65,176--177,186--187,197-199,
261,272-273,352,372,520 363-364,462,500
Myasthenia gravis, 113,483,507,523 Phrenic nerve
Myofibrillar fatigue recording, 283
force-stiffness relation, 36 stimulation, 2G-21, 283-284, 396--398, 405-406
force-velocity relation, 32-36 Physical activity
intracellular acidification, 37-38 deficits due to increases in, 457-466
myoplasmic calcium concentration, 39, 52 muscle overload, 442-443
power output, 41 training for endurance, 333, 438-441, 478
studies, 29--41 isometric, 441
Myosin, heavy chain, isoforms, 324-326 Plasma fatty acids
Myotonic disorders, 116--118,507-508 effect on free tryptophan, 317-319
Plasticity
23Na MRS: see Magnetic resonance spectroscopy acute, 333-335
Negative work, 3 capillary growth, 437
Neuromuscular junction chronic stimulation, 326, 432-438
differences between fiber types, 95-97 disuse effects, 445-448
effects of disuse, 96 endurance training, 438-441
remodeling, 96--97, 472-473 energy balance as factor, 181
Neuromuscular transmission fiber type conversion, 432-448
calcium effects, 92-93 growth factors involved, 437
jitter, llG-1l8 in muscle reinnervation, 443-445
Neuromuscular transmission failure muscle fiber, 326
in different fiber types, 85--87 of neuromuscular junction, 96, 472-473
in disuse, 86 resistance training, 441-442
mechanisms involved, 94-97 slow-to-fast adaptation, 437-438
postsynaptic sites, 94-97 Post-polio syndrome, 499-500
presynaptic sites, 89-93, 11G-l11 Potassium
Nuclear magnetic resonance (NMR): see Magnetic channels, 46, 75, 93, 435
resonance spectroscopy effects on cardiovascular and respiratory sys-
tems,76
Optimization effects on motor control, 77
stimulus pattern for force output, 227-237 effects on muscle reflexes, 77
Oxidative phosphorylation, 177, 186 equilibrium potential, muscle fiber, 69
540 Index

Potassium (cont.) Sense of effort, 135,309-311,373

fluxes, 70 dissociation from sense of force, 309-310
interstitial concentration, 71 in maximal effort, 286-287
loss from muscle, 69 role of5H-T, 316-318
role in contracture, 104 Sense offatigue, 483, 495
role in muscle-fiber action potential, 104 Sense of force
Presynaptic inhibition during muscle fatigue, 307-309
possible role in fatigue, 252 measurement, 306-307
Presynaptic terminal power function, 307
in fatigue, 92-94 in weakness, 310-311
Psychophysical methods; see also Methods Sensory thresholds
contralateral limb-matching, 306 during fatigue, 308
ratio-scaling, 307 Serotonin (5-HT)
Pulmonary afferents 405 role in central fatigue, 315--319
Pulmonary system Single fiber EMG (SFEMG)
airflow limitation, 395 during ischemia, 112
fatigue of, 385-386, 393--399, 401-411 in Lambert-Eaton myasthenic syndrome, 115
lack ofmallea,bility, 394-395 in myasthenia gravis, 113
limitation to V 0 max, 394-395 in myotonia, 116-118
overbuilt concept, 394-395 in normal muscle, 110-111
Size principle, 123
Readiness potential Sodium/potassium pump; see also Muscle fiber
application of chaos theory, 297-303 role in fatigue, 73
effect of load, 296 Spike-frequency adaptation, motoneurons
in fatigue, 296 characteristics, 126-127, 139,246-247
methodology, 295-303 mechanisms, 127-128
respiratory, 410 relation to fatigue, 130, 142,247
Reciprocal group Ia inhibition Spike triggered-averaging: see Methods of analysis
possible role in fatigue, 252-253 Spinal cord injury, 150-151,501-502
Recovery from fatigue, 198--202 Spinal muscular atrophy 167
historical perspective, 485-491 Stiffness
Recurrent inhibition, role in fatigue, 249-251 of muscles after exercise, 464-465
Respiratory muscles, 395--399, 401-411; see also ofmyofibrils in fatigue, 36, 40
Pulmonary system Stroke volume, 387, 394
central fatigue, 290, 404-411 Stroop color and word test, 318
failure, 396-399, 404-409 Synaptic vesicles
in hypoxia, 375 quantal content, 92
Resting membrane potential role of calcium, 92
axon
K+-mediated effects, 90 Task-dependence of fatigue; see also Fatigue, fail-
motoneuron ure sites, task dependence
change to threshold potential, 124 central, 281
muscle vs. failure sites, 351-358, 361-376
contribution ofNa+,K+-pump, 75 indicated by intramuscular pressure (IMP), 333--334
following stimulation, 74-75, 95-96, 102 metabolic factors, 185--192
K+-induced depolarization, 74-75 of respiratory muscle, 396-399, 406
in patients, 507-508 Tension-frequency relationship; see also Force-fre-
Ryanodine channels (receptors): see Calcium chan- quency relationship
nels motor units, 18, 130, 136-137, 153--155,335--337
muscle, 288
Safety factor; see also Transmission safety Tetrodotoxin (TTX), 86, 447
muscle fiber action potential, 104-105 Training
Sarcolemma endurance, 438-441
action potential, 101-106 resistance (strength), 441-442
excitability, 83 Transmission failure
SarcoplasmiC reticulum assessment, 83
calcium release channels, 49 neuromuscular junction, 83
luminal (calcium), 52 postsynaptic sites, 94
Satellite cells, 442-443 presynaptic sites, 89
Index 541

Transmission safety, action potential; see also Velocity of shortening

Safety margin effect of temperature, 335
axon, 110 in fatigue, 32-38, 45, 61, 65, 324-329, 370,
disorders, 110 475--476,516
muscle fiber, 102 Voluntary activation
neuromuscular junction, 11 ~111 motor unit, 18, 109--110, 148-151, 161-162,
Transverse tubular system: see T-system 229,261-265,287-289,37~371,42~
Triadin,48 423,520
Tryptophan, free; see also Plasma fatty acids movement(s), 3--4, 317-318, 327, 33~335,
in central fatigue, 315-319 374,384,393-398,409--411
influence of plasma fatty acids, 317-319 muscle(s) 16--17, 187-191, 199--208,281-287,
T-system; see also Excitation-contraction coupling 305-312,343-344,366--369,409--411,
action potential, 40, 46, 76, 364-365 418--420,423,462--464,475--477,484,
calcium channels, 46--50 498,521
characteristics, 46--47
operation in fatigue, 40, 46--47, 61, 76
Twitch interpolation, 283-285, 406--407 Water uptake by muscle, 71, 218-220, 461