Accepted Manuscript

Treatment of avoidance behavior as an adjunct to exposure therapy: Insights from

modern learning theory

Michael Treanor, PhD, Tom J. Barry

PII: S0005-7967(17)30087-6
DOI: 10.1016/j.brat.2017.04.009
Reference: BRT 3129

To appear in: Behaviour Research and Therapy

Received Date: 29 September 2016

Revised Date: 11 April 2017
Accepted Date: 17 April 2017

Please cite this article as: Treanor, M., Barry, T.J., Treatment of avoidance behavior as an adjunct to
exposure therapy: Insights from modern learning theory, Behaviour Research and Therapy (2017), doi:
10.1016/j.brat.2017.04.009.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 1

PT
Treatment of Avoidance Behavior as an Adjunct to Exposure Therapy: Insights from

Modern Learning Theory

RI
Michael Treanor1*

SC
And

Tom J Barry2,3

U
AN
1
University of California, Los Angeles
2
Institute of Psychiatry, Kings College London
3
Experimental Psychopathology Lab, Department of Psychology, The University of Hong
M

Kong
D

*
Correspondence should be addressed to
TE

Michael Treanor, Ph.D.

Department of Psychology
EP

University of California, Los Angeles

1285 Franz Hall
Box 951563
Los Angeles, CA 90095
C

mtreanor@psych.ucla.edu
AC
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 2

Abstract

Pathological avoidance of benign stimuli is a hallmark of anxiety and related disorders,

and exposure-based treatments have often encouraged the removal of avoidance, or safety

behaviors, due to their negative effects on extinction learning. Unfortunately, empirical

PT
evidence suggests that avoidance behaviors can persist following treatment, and the mere

RI
availability of avoidance behavior can be sufficient to renew fear following successful

extinction learning. The present paper critically examines the function of avoidance

SC
behavior through the lens of modern learning theory, and speculates on novel behavioral

and pharmacological strategies for targeting avoidance as an adjunct to current evidence-

based treatments.
U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 3

Avoidance has long held a central role in theories regarding the genesis and

maintenance of anxiety disorders. For example, Mowrer (1951) conceptualized avoidance

as maintained through negative reinforcement resulting from anxiety reduction. More

recently, avoidance has been conceptualized as being driven by expectation that a

PT
stimulus will lead to an aversive outcome (Declercq, De Houwer, & Baeyens, 2008). In

RI
both instances, avoidance becomes pathological when performed in response to relatively

benign stimuli.

SC
Although avoidance has been important in theories of anxiety, translational

research has largely focused on other Pavlovian processes, such as extinction learning, as

U
the principal mechanism of treatment for anxiety disorders (e.g., exposure therapy). The
AN
implicit assumption has been that avoidant behavior would decrease as individuals

learned that a threatening stimulus (conditional stimulus or CS) no longer predicted an

M

aversive outcome (unconditional stimulus or US). That is, following extinction training,
D

and the repeated presentation of the CS in the absence of the US, there would no longer
TE

be any need to avoid the CS. However, empirical evidence suggests that avoidance

behavior can persist following extinction (Rodriguez-Romaguera, Greenberg,

EP

Rasmussen, & Quirk, 2016; Solomon, Kamin, & Wynne, 1953), and the availability of

avoidant behavior can renew fear even following successful extinction learning. For
C

example, Vervliet and Indekeu (2015) conditioned avoidance behavior (a button press
AC

prevented a shock during a CS presentation) and then conducted extinction training

where the avoidance behavior was not available. Self-reported fear and physiological

arousal to the CS decreased during the extinction phase, however, simply making the

avoidant response available at a later test phase when the CS was presented again caused
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 4

fear to return to the CS. Similar results have been obtained in rodents (Bravo-Rivera,

Roman-Ortiz, Montesinos-Cartagena, & Quirk, 2015).

This presents obstacles to evidence-based interventions based on extinction such

as exposure therapy. In exposure-based treatment clients are often encouraged to refrain

PT
from avoidant behavior (e.g., use of anxiolytic medication, compulsive behaviors, having

RI
a “safe” person). However, the above evidence suggests that avoidance behavior may

persist, and the mere availability of an avoidant response may be sufficient to renew fear

SC
following treatment. This may represent one reason patients relapse following exposure

therapy (Ginsburg et al., 2014).

U
The reason for the deleterious impact of avoidant behavior availability following
AN
successful extinction or exposure remains unclear. One possibility is that removing

avoidant behavior during extinction represents a context shift such that it differs from
M

both the original context in which fear was acquired as well as other contexts that might
D

be encountered after extinction/exposure. As such, when the avoidance response is

TE

available again, this represents another context shift from extinction, and fear is then

renewed in the same way that it might if extinction had taken place in a different physical
EP

context/environment (Vansteenwegen et al., 2005; Vervliet & Indekeu, 2015).

Regardless, these findings suggest that it may be important to examine the treatment of
C

avoidance behavior as an adjunct to exposure-based procedures in order to mitigate

AC

renewal of symptoms (Vervliet & Indekeu, 2015).

The present paper critically examines the treatment of avoidant behavior through

the lens of modern learning theory. Through examination of the various functions

avoidant behavior may serve in associative learning processes, as well as its

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 5

neurobiological substrates, we aim to highlight novel behavioral and pharmacological

interventions that may serve as useful adjuncts to traditional evidence-based strategies for

anxiety and related disorders. In addition, given the dearth of evidence elucidating the

mechanisms responsible for the return of fear following treatment as a result of the

PT
availability of avoidance behavior, we conclude with concrete recommendations for

RI
future research.

Avoidant Behavior as a Conditional Inhibitor

SC
Extinction of conditional inhibition

Extinction learning is one of the presumed mechanisms of exposure therapy

U
(Craske et al., 2008; Scheveneels, Boddez, Vervliet, & Hermans, 2016) and operates via
AN
error correction mechanisms, such that the associative strength of a CS is updated when

the US does not occur. During learning, the greater discrepancy between what is
M

predicted and what actually occurs, the larger the amount of associative change that takes
D

place (Rescorla & Wagner, 1972). Conditional stimuli that predict the occurrence of a US
TE

are known as “conditional excitors” whereas stimuli that directly predict the absence of

the US are “conditional inhibitors”. During extinction training, in which a conditional

EP

excitor is repeatedly presented in the absence of the US, the concurrent presence of

conditional inhibitors decreases the expectation that a US will occur, resulting in less
C

expectancy violation, and therefore negatively impacts extinction learning (Lovibond,

AC

Chen, Mitchell, & Weidemann, 2013; Lovibond, Mitchell, Manard, Brady, & Menzies,

2009; Rescorla, 1969).

Avoidant behaviors, or “safety behaviors”, have often been discussed in terms of

conditional inhibition (Krypotos, Effting, Kindt, & Beckers, 2015). For example, the use
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 6

of benzodiazepines in panic disorder, washing one’s hands in obsessive-compulsive

disorder, and a combat veteran sitting with his back to a wall in a restaurant are all

examples of avoidant behaviors that may function as conditional inhibitors as they are

directly associated with the decreased likelihood of the US occurring (See Figure 1 for a

PT
graphical representation of the relationship between a CS+, conditional inhibitor, and a

RI
US). Importantly, despite functioning as a conditional inhibitor, the availability of

avoidance behavior may still become a contextual feature and lead to context renewal

SC
(Vervliet & Indekeu, 2015). That is, when the avoidance behavior is available following

treatment, this may represent a context shift from exposure procedures when avoidance

U
was prohibited, and results in a return of fear. The implication would be that allowing
AN
some avoidance behavior during exposure may be beneficial to reduce subsequent

context renwewal, although the deleterious impact of avoidance behavior (e.g.,

M

conditional inhibitors) on extinction learning represents a significant problem. Thus,

D

conditional inhibition has to be reduced, or the negative effects of conditional inhibitors

TE

on extinction learning needs to mitigated, prior to allowing avoidance behaviors that

function as inhibitors during exposure therapy. Below, we discuss specific treatment

EP

approaches for targeting conditional inhibitors as an adjunct to exposure therapy.

The traditional paradigm for developing conditional inhibition is to pair a neutral

C

stimulus (B) with an excitatory stimulus (A) without reinforcement (e.g., A+ then AB-).
AC

The resulting decrease in associative strength gradually transforms the previously neutral

stimulus into an inhibitor. For example, engaging in compulsive behavior (neutral

stimulus) when one has obsessive thoughts (conditional stimulus) gradually transforms

the compulsive behavior into a conditional inhibitor when the US doesn’t occur.
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 7

However, dominant learning models suggest that presenting a conditional inhibitor by

itself should result in a gradual loss of inhibition, and may offer one potential strategy for

targeting avoidance behavior. Rescorla & Wagner (1972) conceptualized change in

associative strength as a function of the total amount of learning a US can support (λ)

PT
minus the sum of the associative strength of all the stimuli present on a given trial (ΣV).

RI
Let us assume a negative associative strength of a conditional inhibitor of -.5. Presenting

it alone, in the absence of another CS or US should result in a net positive amount of

SC
associative change (λ - ΣV becomes 0 – [-.5]) that will gradually eliminate inhibition

(Zimmer-Hart & Rescorla, 1974). For example, an individual with obsessive-compulsive

U
disorder may be asked to wash her hands compulsively in the absence of touching a
AN
contaminated surface while someone with panic disorder may be asked to take a

benzodiazepine at times when he is not anxious. Although this is consistent with

M

dominant learning models, numerous animal studies have failed to find any loss of
D

inhibition after repeatedly presenting a conditional inhibitor in isolation (e.g., DeVito &
TE

Fowler, 1987).

However, in a study of human contingency learning, Melchers, Wolff and Lachnit

EP

(2006) argued that one could produce extinction of conditional inhibition depending on

the nature of the US. The authors argue that traditional Pavlovian procedures use
C

unconditional stimuli that only vary unidirectionally. For example, one is either shocked
AC

or not shocked in conditioning and extinction experiments. However, the Rescorla-

Wagner model’s assumption that inhibition is the opposite of excitation would necessitate

that the US can take on values less than zero. When the US can only vary in one

direction, a conditional inhibitor predicts the non-occurrence of the US and there is no

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 8

discrepancy, or extinction learning, when it is presented alone without the US. However,

when the US can take on both positive and negative values, then presenting a conditional

inhibitor in isolation can still lead to expectancy violation.

In the Melchers et al. (2006) study, participants were divided into two groups

PT
tasked with determining whether a fictional individual’s hormone levels would rise (US)

RI
based upon consumption of certain foods (CS). In one group, hormone levels could only

rise or remain the same (unidirectional group), whereas in the other group hormone levels

SC
could rise, remain the same, or decrease (bidirectional group). Using standard paradigms

for developing inhibition the authors demonstrated that you could reduce conditional

U
inhibition through non-reinforced presentations of the inhibitor but only in the group in
AN
which the US was allowed to vary bidirectionally (Melchers et al., 2006). Subsequent

studies have replicated and extended this result (Baetu & Baker, 2010; Lotz & Lachnit,
M

2009).
D

If extinction of conditional inhibition is possible with a bidirectional US, then the

TE

question becomes whether clinical disorders meet this criterion. In social anxiety

disorder, exposures are often tailored towards interpersonal interactions. The feared
EP

consequence or US may be rejection, but the US could be conceptualized as lying on a

continuum ranging from approval, to a neutral response, to rejection. Thus, social anxiety
C

disorder may be one instance in which a perceived US can vary bidirectionally, and
AC

therefore avoidant behaviors that function as conditional inhibitors may be amenable to

unreinforced exposure. However, other disorders, such as panic disorder, may entail

unconditional stimuli that are best characterized as unidirectional. A person may only

suffer a heart attack or not as a result of a rapid heartbeat. A task for future research is to
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 9

determine whether or not unconditional stimuli in anxiety disorders are best characterized

as unidirectional or bidirectional, and the degree to which conditional inhibitors in these

disorders are amenable to extinction learning via presentation of the inhibitor in the

absence of the excitatory CSs or US. In addition, it may be important to examine the bi-

PT
directional nature of unconditional stimuli across disorders (e.g., eating disorders) to

RI
further elucidate potential translational applications. Although extinction of inhibition

represents an exciting possibility, considerable additional research is needed to further

SC
explore its translational applicability.

Aversive learning and counter-conditioning

U
Vervliet and Indekeu (2015) suggest that the availability of avoidance behavior
AN
following extinction may result in fear renewal partly because it is a context shift from

extinction (where avoidant behavior was not available). They argued that presenting the
M

avoidant behavior occasionally during extinction, and therefore making it a feature of the
D

extinction context, might facilitate generalization of extinction learning by making it

TE

more similar to other contexts one encounters outside of extinction/exposure. However,

presenting a conditional inhibitor during extinction, even sparingly, might still negatively
EP

impact extinction learning on a given trial.

Conditional inhibitors mitigate extinction learning because of their negative

C

associative “charge” that reduces expectancy violation. It follows that if this negative
AC

charge can be reversed, such that the inhibitor is now itself associated with an aversive

event, then this may reduce its detrimental effects on extinction. Indeed, initial results in

rodents demonstrated that pairing the inhibitor with a reinforcer reduced conditional
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 10

inhibition (Holland, 1984; Zimmer-Hart & Rescorla, 1974), although some inhibitory

strength may remain (Pearce &Wilson, 1991).

Unfortunately, this may have limited clinical utility. It is certainly not ethical to

present the US (e.g., a trauma) alongside a conditional inhibitor in most anxiety

PT
disorders. However, in social anxiety disorder the US (e.g., rejection) can occur without

RI
detrimental results to the patient. For example in occasional reinforced extinction the CS

is intermittently paired with the US to enhance extinction learning (Bouton, 2004).

SC
Clinically, this most often takes the form of “shame attacks” in social anxiety disorder

where the individual engages in behavior that has a high likelihood of rejection (e.g.,

U
asking strangers if the earth revolves around the sun) during an exposure exercise.
AN
Allowing the individual to engage in avoidant behavior, such as hiding signs of anxiety,

during occasional reinforced extinction may represent one strategy for altering the
M

inhibitory strength of avoidance behaviors that function as conditional inhibitors.

D

Although this strategy may be useful for social anxiety disorder, it still requires the
TE

presentation of an aversive US, and it is unlikely to be useful for other anxiety disorders.

Timing issues and attentional redirection

EP

If allowing avoidance behavior during extinction is necessary to reduce return of

fear than it is important to consider the optimal timing of avoidance behavior during
C

extinction, as well as strategies to mitigate the deleterious effects of conditional inhibitors

AC

on expectancy violation. Rescorla & Wagner (1972) suggest that the largest changes in

associative strength will occur in the early phases of extinction learning, as the CS+ is

still a strong predictor of the US. During initial training, there is a large discrepancy

between the predictive strength of the CS and the non-occurrence of the US. However, as
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 11

extinction proceeds this discrepancy is reduced as the CS-noUS relationship becomes

stronger.

Inasmuch as there are smaller changes in associative strength in the later phases

of extinction, then the presence of a conditional inhibitor will have less of a negative

PT
effect on extinction learning. This might represent the optimal time to combine the

RI
availability of avoidance behavior with extinction learning. The occasional availability of

avoidant behavior may allow it to become a feature of the extinction context, thereby

SC
serving as a retrieval cue for extinction if it is available at a later time, while the

infrequent presence of conditional inhibitors, combined with their use only during the

U
later phases of extinction, will mitigate any negative effects on expectancy violation.
AN
Clinically, this may entail the availability of avoidance behaviors that function as

conditional inhibitors in the last several sessions of exposure therapy, and then only
M

during a select few trials, as a relapse prevention measure.

D

In addition to error-correction, associative change is enhanced by CS salience.

TE

Dominant learning models suggest that stimuli “compete” for changes in associative

learning, such that the more salient stimuli acquire the largest amount (Mackintosh, 1975;
EP

Pearce & Hall, 1980, Rescorla & Wagner, 1972). Thus, in trials with compound stimuli,

manipulations that enhance attention to one stimulus (e.g., A) at the expense of the other
C

(e.g., B) may allow the target stimulus (A) to acquire the greater amount of associative
AC

change. When presenting a conditional inhibitor alongside a conditional excitor,

enhancing the attentional salience of the excitor may allow it to accrue the bulk of

associative change.
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 12

In addition, the impact of a conditional inhibitor on extinction learning may be

further reduced through attentional manipulations. Learning models often employ a

summed error term (ΣV) that determines the overall amount of associative change that

occurs on a given trial. For example, if a conditional excitor and inhibitor have the same

PT
amount of associative strength then they will cancel each other out and result in no

RI
learning during extinction. However, the more salient a stimulus is, the more its current

associative strength is likely to contribute to the summed error term. Reducing the

SC
salience of a conditional inhibitor may reduce the degree to which its inhibitory strength

contributes to this summed error term, and therefore could mitigate the negative effects of

U
conditional inhibitors during extinction. Clinically, this may take the form of explicit
AN
instructions to focus on the CS+ at the expense of the inhibitor. For example, in exposure

therapy for panic disorder the patient may be provided with explicit instructions to attend
M

to the sensations of a rapid heartbeat (CS+) while ignoring the anxiolytic medication in
D

her purse (conditional inhibitor). However, additional research is necessary to explore

TE

these possibilities, along with examination of feasible treatment strategies for

manipulating attention (e.g., attentional bias modification, explicit instructions).

EP

Avoidance Behavior as Occasion Setting

Although avoidance behavior may often function as a conditional inhibitor, it may

C

modify CS-US relationships in other manners as well. Modulatory stimuli, or occasion

AC

setters, are stimuli that are not directly associated with the US, but “set the occasion” for

whether or not the CS will lead to the US (Holland, 1989). For example, in panic

disorder, an individual may fear being alone with a rapid heartbeat. Although being alone

is not directly related to the US (e.g., being alone is not predictive of a heart attack), the
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 13

individual may fear that a rapid heartbeat is more likely to lead to a heart attack if they

are alone. Avoidance behavior has often been discussed in terms of negative occasion

setting (Declerq & De Hower, 2008; De Hower, Crombez, & Baeyens, 2005; Kryptotos

et al., 2015; See Figure 2 for a graphical representation of the relationship between a

PT
CS+, occasion setter, and a US). Occasion setting is more likely when the behavior or

RI
stimulus precedes the CS or is a less salient feature (e.g., the context; Holland, 1986). In

anxiety disorders, typical avoidance behaviors that may function as negative occasion

SC
setters include cellular phones or “safe” individuals in panic disorder and preparing

conversation topics ahead of time in social anxiety disorder.

U
As with conditional inhibition, one strategy may be to present the occasion setter
AN
alone, thereby extinguishing its modulatory properties or any additional association it

might have with the US. Unfortunately, several experimental investigations in animals
M

and humans have demonstrated that this strategy fails to result in extinction of occasion
D

setting (Holland, 1989; Rescorla, 1986; van Vooren, Franssen, Beckers, Hermans, &
TE

Baeyens, 2012). However, reversing the contingencies of occasion setting, such that a

negative occasion setter that predicted a CS would not lead to a US now predicts the CS
EP

will lead to the US, does successfully extinguish occasion setting (Rescorla, 1986).

Unfortunately, similar to counter-conditioning conditional inhibition, reversing the

C

contingencies of a negative occasion setter may have limited clinical applicability beyond
AC

social anxiety disorder.

Reducing the ability of avoidant behavior to renew fear may require occasionally

presenting the avoidant behavior during extinction, thereby making it a feature of the

extinction context, although this may simultaneously reduce expectancy violation and
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 14

mitigate extinction learning. We previously suggested several behavioral strategies for

reducing these deleterious effects when the avoidant behavior functions as a conditional

inhibitor (e.g., attentional modulation, allowing the avoidance behavior only during the

later phases of extinction, etc.). Indeed, similar strategies should be examined in

PT
reducing the impact of avoidance behaviors that function as negative occasion setters, as

RI
the extant evidence suggests that occasion setters can similarly impact error-correction

and conditional responding (Morel & Holland, 1993).

SC
Pharmacological interventions for Conditional Inhibitors and Occasion Setters

There has been increased interest in pharmacological agents that may impact

U
associative learning processes as adjuncts to traditional behavioral interventions (Quirk &
AN
Mueller, 2008). However, we are unaware of any research examining pharmacological

treatment of avoidance behavior in the context of extinction and exposure procedures.

M

Thus, the following section examines potential molecular pathways that may serve as
D

useful targets for pharmacological agents in the treatment of conditional inhibition and
TE

occasion setting.

As mentioned previously, presentation of a conditional inhibitor in isolation (e.g.,

EP

extinction of inhibition) may represent one behavioral strategy for reducing the impact of

avoidance behavior on renewal of fear when the US varies bi-directionally.

C

Neurobiological models of traditional extinction implicate neuroplasticity and de novo

AC

protein synthesis in the consolidation of the extinction memory (Quirk & Mueller, 2008).

For example, a wealth of research has demonstrated the importance of the N-Methyl-D-

aspartate (NMDA) receptor in extinction learning (Myers & Davis, 2007). Indeed,

NMDA receptor agonists, such as d-cycloserine (DCS) have been found to enhance
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 15

extinction learning in both animals and humans (Davis, Ressler, Rothbaum, &

Richardson, 2006), and have shown promise as adjuncts to exposure-based treatments

(Guastella et al., 2008).

Importantly, NMDA receptor activity is not confined to fear extinction, and is an

PT
important component of a variety of memories including conditional inhibition (Foilb,

RI
Flyer-Adams, Maier, & Christianson, 2016). If NMDA receptor activity underlies both

conditional inhibition and extinction learning then it is possible that DCS may be useful

SC
during extinction of inhibition. Extinction of inhibition requires presentation of the

inhibitor in isolation. For example, an individual with post-traumatic stress disorder may

U
repeatedly check his household locks while an individual with obsessive-compulsive
AN
disorder may wash her hands repeatedly. A key to extinction of inhibition is presenting

the conditional inhibitor in the absence of both the CS+(stimulus that predicts the US)
M

and US. Thus, in the examples above the patient would only engage in the avoidance
D

behavior when they are not “triggered” or confronted with a CS+. If pharmacological
TE

agents, such as DCS, also enhance extinction of inhibition, then the amount of time

dedicated to exposure to the conditional inhibitor in isolation can be reduced, while

EP

simultaneously strengthening the consolidation of extinction of inhibition. However,

there may also be risk in attempting to enhance the consolidation of extinction of

C

inhibition, as any failure in extinction may paradoxically result in strengthening

AC

conditional inhibition.

As discussed previously, reducing the impact of avoidance behavior on extinction

generalization may occasionally require presenting the conditional inhibitor in compound

with the CS+ during exposure therapy. However, presenting a conditional inhibitor also
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 16

reduces expectancy violation during a given trial and mitigates extinction learning.

Pharmacological agents that enhance the excitatory strength of the CS+ may reduce the

deleterious impact of avoidance behavior on extinction by magnifying expectancy

violation despite the presence of a conditional inhibitor during an extinction trial.

PT
The neurobiological substrates of error correction mechanisms have been linked

RI
to dopamine signaling, such that increased dopaminergic activity is associated with

surprise or expectancy violation (e.g., Steinberg, Keiflin, Boivin, Witten, Deisseroth &

SC
Janak, 2013). Although originally associated with reward, there is increased evidence that

dopamine signaling plays a role in aversive conditioning and extinction as well (Haaker

U
et al., 2013; Hikind & Maroun, 2008; Holtzman-Assif, Laurent, & Westbrook, 2010;
AN
Lissek, Glaubitz, Wolf, & Tegenthoff, 2015; Mueller, Bravo-Rivera, & Quirk, 2010;

Yuan Li & McNally, 2014).

M

During normal extinction, the non-occurrence of the US should produce

D

prediction error for the CS+ but not a conditional inhibitor (Schultz, 2007). Indeed,
TE

dopamine neurons demonstrate reactivity during US omission to a CS+, but not during

the presentation of a conditional inhibitor or conditional inhibitor combined with a

EP

conditional excitor (Tobler, Dickinson, & Schultz, 2003). The latter case represents a lack

of prediction error when the CS+ is presented in compound with a conditional inhibitor.
C

Thus, increasing dopaminergic activity during extinction with a CS+/conditional inhibitor

AC

compound may enhance the molecular substrates of expectancy violation, mitigating any

negative effects of conditional inhibition on extinction learning. Unfortunately, we are

unaware of any studies that have directly examined this possibility. However,

amphetamine sensitization (which increases dopamine levels) reduces the effect of a

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 17

conditional inhibitor on a conditional excitor (Shiflett, Riccie, & DiMatteo, 2013 but see

Harmer & Phillips, 1999), and suggests that dopamine agonists might offer protection

against the negative effects of avoidance behavior during exposure therapy.

In addition to dopamine, opioid receptors play a role in error correction mechanisms.

PT
Opioid release modulates Pavlovian learning such that excitatory conditioning

RI
corresponds with reduced opioid release whereas extinction learning is facilitated by

increased opioid activity (Myers & Davis, 2007; Yuan Li & McNally, 2014). However,

SC
as with dopamine agonists, a key question is how modulating opioid activity affects

extinction learning in combination with a conditional inhibitor. As discussed previously,

U
during presentation of an excitatory/inhibitory compound, the learning history of each
AN
stimulus affects responding to the compound (e.g., inhibition can reduce expectancy

violation). Blocking is a related cue competition phenomenon where a well-established

M

conditional excitor (A+) prevents conditioning to a new stimulus (B) when they are
D

presented in compound (AB+). This occurs because the occurrence of the US is already
TE

well predicted by stimulus A, so no new learning accrues to stimulus B. However,

modulating opioid activity reduces blocking, and allows conditioning to stimulus B

EP

(Iordanova, McNally, & Westbrook, 2006). This may be tantamount to a well-

conditioned inhibitor (that strongly predicts the non-occurrence of the US) reducing
C

extinction learning when combined with a conditional excitor, and suggests that targeting
AC

opioid receptors is a promising intervention. Indeed, recent evidence suggests that

modulating opioid activity reduces the impact of a conditional inhibitor on a conditional

excitor in rats (Laurent, Wong, & Balleine, 2015).

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 18

Unfortunately, pharmacological agents may have limited utility when targeting

occasion setting. For example, inasmuch as presenting an occasion setter alone does not

appear to result in extinction learning (Holland, 1989; Rescorla, 1986), pharmacological

enhancement of memory consolidation (as discussed above) will not be helpful. In

PT
addition, hippocampal lesions and pharmacological disruption of hippocampal activity

RI
reduce the acquisition of occasion setting (Meyer, Putney, & Bucci, 2015; Yoon,

Graham, & Kim, 2011), but it is unclear how this would be leveraged in clinical

SC
situations, as the avoidant behavior (i.e., negative occasion setting) would have already

been acquired.

U
However, if, as Vervliet and Indekeu (2015) suggest, the availability of avoidance
AN
behavior following treatment can renew fear because it represents a context shift from

extinction, then pharmacological strategies that may reduce the contextual gating of
M

extinction learning may be helpful. The contextual nature of extinction learning has been
D

linked to increased hippocampal activity during extinction (Hermann, Stark, Milad, &
TE

Merz, 2016; Holland & Bouton, 1999). In an elegant study, rats received systemic

injections of the anticholinergic scopolamine. The hippocampus is rich in cholinergic

EP

receptors (Yi et al., 2015), and therefore “impairing” hippocampal activity may reduce

the ability of the hippocampus to bind extinction learning to a particular contexts. Indeed,
C

rats treated with scopolamine generalized extinction learning to a novel context, whereas
AC

vehicle treated rats demonstrated increased fear consistent with context renewal

(Zelikowsky, et al., 2013). This may suggest that pharmacological agents that disrupt

hippocampal activity during extinction may allow learning to generalize to new contexts,

and may represent one strategy to mitigate any contextual specificity when avoidance
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 19

behavior is prevented during extinction or exposure procedures. Of course, one potential

concern with combining extinction and pharmacological agents is that the drug may

create an “internal context” that precipitates context renewal when the individual is tested

at a subsequent time point in the absence of the drug (Bouton, 2004).

PT
Summary

RI
Despite the efficacy of extinction-based procedures (i.e., exposure therapy) in the

treatment of anxiety disorders, pathological avoidance may still persist, and empirical

SC
evidence suggests that the mere availability of avoidance behavior can renew fear. The

present paper attempted to elucidate novel behavioral and pharmacological interventions

U
for avoidance behavior derived from learning theory and neurobiology. A few key
AN
findings emerged. First, consistent with previous discussions (Krypotos et al., 2015)

avoidance behavior may serve various functions including conditional inhibition and
M

negative occasion setting. Accurately determining whether a given avoidant behavior is

D

functioning as a conditional inhibitor or negative occasion setter may be important, as

TE

each may require different behavioral and pharmacological strategies. Unfortunately,

typical laboratory-based paradigms for assessing inhibition or occasion setting (e.g., tests
EP

of summation or retardation) will be difficult to carry out with actual avoidance behavior

in clinical settings. Translational interventions may have to rely on detailed functional

C

analyses in order to determine the function of an avoidant behavior. For example, if the
AC

behavior directly predicts the non-occurrence of the US, rather than simply modulating

the likelihood that a CS will lead to a US, then it is most likely a conditional inhibitor.

Second, the type of intervention may differ depending on how avoidant behavior

increases risk for renewal of fear. If removing avoidance behavior during extinction
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 20

represents a context shift (Vervliet & Indekeu, 2015), then allowing avoidance behavior

during extinction may be necessary. However, this will require behavioral or

pharmacological methods to reduce any negative effects of avoidance behavior on

expectancy violation.

PT
Each of these possibilities requires further empirical investigation. Below, we offer

RI
several concrete recommendations for future research.

1. Determine the degree to which a bidirectional US facilitates extinction of

SC
inhibition, and the extent to which anxiety disorders correspond to this criterion.

Although several studies in human contingency learning have demonstrated this

U
possibility, it is important to replicate these results in human fear conditioning
AN
with more externally valid types of unconditional stimuli. For example, standard

laboratory conditioning and extinction paradigms can be combined with relevant

M

unconditional stimuli that either vary bidirectionally or unidirectionally (e.g.,

D

breathing occlusion for panic patients or insults for individuals with social anxiety
TE

disorder; Lissek et al., 2008). Employing a conditional inhibitor in these

paradigms will allow examination of extinction of inhibition depending upon the

EP

nature of the US.

2. Examine the impact of attentional manipulations (e.g., explicit or implicit training

C

in attending to the CS+) on extinction alongside a conditional inhibitor and

AC

subsequent return of fear in both healthy controls and individuals with anxiety

disorders.

3. Examine effects of pharmacological agents, such as dopamine agonists and opioid

agonists/antagonists, on extinction of a CS+ in combination with a conditional

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 21

inhibitor. This may allow the avoidant behavior to become a feature of the

extinction context, thereby reducing return of fear when it is made available

during a subsequent test, while simultaneously reducing any negative effect on

expectancy violation during extinction.

PT
In examining the treatment of avoidance behavior alongside exposure procedures, it

RI
will be necessary to employ a more nuanced view of avoidance behavior derived from

modern learning theory. For example, the extent to which a given avoidance behavior

SC
may negatively affect extinction learning may be dependent upon whether it is a

conditional inhibitor or occasion setter, as well as the overall ratio of inhibition to

U
excitation (i.e., stronger inhibitors will have a greater impact on extinction).
AN
Unfortunately, many studies have failed to take into account this degree of complexity,

which may partially explain mixed results regarding whether or not avoidance is
M

detrimental to exposure processes (Meulders, Van Daele, Volders, & Vlaeyen, 2016). It
D

is also important to note that while learning theory may provide a parsimonious
TE

explanation of numerous key mechanisms in the etiology, maintenance, and treatment of

avoidance and anxiety disorders, it may represent only a portion of the complexity
EP

inherent in psychopathology. However, given the centrality of learning theory to anxiety

disorders, the present article attempted to provide a more nuanced analysis of avoidance
C

behavior in order to determine translational treatment approaches that may be useful as

AC

adjuncts to current exposure-based procedures.

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 22

References

Baetu, I. & Baker, A. G. (2010). Extinction and blocking of conditioned inhibition in

human causal learning. Learning & Behavior, 38, 394-407.

doi:10.3758/LB.38.4.394

PT
Bouton, M. E. (2004). Context and behavioral processes in extinction. Learning &

RI
Memory, 11, 485-494. doi: 10.1101/lm.78804

Bravo-Rivera, C., Roman-Ortiz, C., Montesinos-Cartagena, M., and Quirk, G. J. (2015).

SC
Persistent active avoidance correlates with activity in prelimbic cortex and ventral

striatum. Front. Behav. Neurosci. 9:184. doi: 10.3389/fnbeh.2015.00184

U
Craske, M. G., Kircanski, K., Zelikowsky, M., Mystkowski, J., Chowdhury, N., & Baker,
AN
A. (2008). Optimizing inhibitory learning during exposure therapy. Behaviour

Research and Therapy, 46, 5-27. doi: 10.1016/j.brat.2007.10.003

M

Davis, M., Ressler, K., Rothbaum, B. O., & Richardson, R. (2006). Effects of D-
D

cycloserine on extinction: translation from preclinical to clinical work. Biological

TE

psychiatry, 60(4), 369-375. doi.org/10.1016/j.biopsych.2006.03.084

Declercq M., De Houwer J. (2008). Evidence for the interchangeability of an avoidance

EP

behavior and a negative occasion setter. Learning & Behavior, 36, 290–300.

10.3758/LB.36.4.290
C

Declercq, M., De Houwer, J., & Baeyens, F. (2008). Evidence for an expectancy-based
AC

theory of avoidance behaviour. The Quarterly Journal of Experimental

Psychology, 61, 1803-1812. doi: 10.1080/17470210701851214

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 23

DeVito, P. L., & Fowler, H. (1987). Enhancement of conditioned inhibition via an

extinction treatment. Animal Learning & Behavior, 15(4), 448-454.

doi:10.3758/BF03205055

Foilb, A. R., Flyer-Adams, J. G., Maier, S. F., & Christianson, J. P. (2016). Posterior

PT
insular cortex is necessary for conditioned inhibition of fear. Neurobiology of

RI
Learning and Memory, 134, 317-327. doi.org/10.1016/j.nlm.2016.08.004

Ginsburg, G. S., Becker, E. M., Keeton, C. P., Sakolsky, D., Piacentini, J., Albano, A.

SC
M…Kendall, P. C. (2014). Naturalistic follow-up of youths treated for pediatric

anxiety disorders. JAMA Psychiatry, 71, 310-318.

U
doi:10.1001/jamapsychiatry.2013.4186
AN
Guastella, A. J., Richardson, R., Lovibond, P. F., Rapee, R. M., Gaston, J. E., Mitchell,

P., & Dadds, M. R. (2008). A randomized controlled trial of D-cycloserine

M

enhancement of exposure therapy for social anxiety disorder. Biological

D

psychiatry, 63(6), 544-549.

TE

Haaker, J., Gaburro, S., Sah, A., Gartmann, N., Lonsdorf, T. B., Meier, K., ... & Kalisch,

R. (2013). Single dose of L-dopa makes extinction memories context-independent

EP

and prevents the return of fear. Proceedings of the National Academy of Sciences,

110(26), E2428-E2436. doi/10.1073/pnas.1303061110

C

Hikind, N., & Maroun, M. (2008). Microinfusion of the D1 receptor antagonist,

AC

SCH23390 into the IL but not the BLA impairs consolidation of extinction of

auditory fear conditioning. Neurobiology of learning and memory, 90(1), 217-

222. doi.org/10.1016/j.nlm.2008.03.003
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 24

Harmer, C. J., & Phillips, G. D. (1999). Enhanced conditioned inhibition following

repeated pretreatment with d-amphetamine. Psychopharmacology, 142(2), 120-

131. doi:10.1007/s002130050870

Hermann, A., Stark, R., Milad, M. R., & Merz, C. J. (2016). Renewal of conditioned fear

PT
in a novel context is associated with hippocampal activation and connectivity.

RI
Social cognitive and affective neuroscience, nsw047. doi: 10.1093/scan/nsw047

Holland, P. C. (1984). Differential effects of reinforcement of an inhibitory feature after

SC
serial and simultaneous feature negative discrimination training. Journal of

experimental psychology. Animal behavior processes, 10, 461-475.

U
Holland, P. C. (1986). Temporal determinants of occasion setting in feature-positive
AN
discriminations. Animal Learning & Behavior, 14(2), 111-120.

doi:10.3758/BF03200045
M

Holland, P. C. (1989). Feature extinction enhances transfer of occasion setting. Animal

D

Learning & Behavior, 17(3), 269-279. doi:10.3758/BF03209799

TE

Holland, P. C., & Bouton, M. E. (1999). Hippocampus and context in classical

conditioning. Current opinion in neurobiology, 9(2), 195-202.

EP

doi.org/10.1016/S0959-4388(99)80027-0

Holtzman-Assif, O., Laurent, V., & Westbrook, R. F. (2010). Blockade of dopamine

C

activity in the nucleus accumbens impairs learning extinction of conditioned fear.

AC

Learning & Memory, 17(2), 71-75. doi: 10.1101/lm.1668310

Iordanova, M. D., McNally, G. P., & Westbrook, R. F. (2006). Opioid receptors in the

nucleus accumbens regulate attentional learning in the blocking paradigm. The

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 25

Journal of neuroscience, 26(15), 4036-4045. doi: 10.1523/JNEUROSCI.4679-

05.2006

Krypotos, A., Effting, M., Kindt, M., and Beckers, T. (2015). Avoidance learning: a

review of theoretical models and recent developments. Front. Behav. Neurosci.

PT
9:189. doi: 10.3389/fnbeh.2015.00189

RI
Li, S. S. Y., & McNally, G. P. (2014). The conditions that promote fear learning:

prediction error and Pavlovian fear conditioning. Neurobiology of learning and

SC
memory, 108, 14-21. doi.org/10.1016/j.nlm.2013.05.002

Laurent, V., Wong, F. L., & Balleine, B. W. (2015). δ Opioid receptors in the

U
accumbens shell mediate the influence of both excitatory and inhibitory
AN
predictions on choice. British journal of pharmacology, 172(2), 562-570. DOI:
M

10.1111/bph.12731

Lissek, S., Glaubitz, B., Wolf, O. T., & Tegenthoff, M. (2015). The DA antagonist
D

tiapride impairs context-related extinction learning in a novel context without

TE

affecting renewal. Frontiers in behavioral neuroscience, 9.

doi: 10.3389/fnbeh.2015.00238
EP

Lissek, S., Levenson, J., Biggs, A. L., Johnson, L. L., Ameli, R., Pine, D. S., & Grillon,

C. (2008). Elevated fear conditioning to socially relevant unconditioned stimuli in

C

social anxiety disorder. The American Journal of Psychiatry, 165, 124-132.

AC

doi.org/10.1176/appi.ajp.2007.06091513

Lovibond, P.F., Chen, S.X., Mitchell, C.J. & Weidemann, G. (2013). Competition

between an avoidance response and a safety signal: Evidence for a single learning

system. Biological Psychology, 92, 9-16

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 26

Lovibond, P.F., Mitchell, C.J., Minard, E., Brady, A., Menzies, R.G. (2009). Safety

behaviours preserve threat beliefs: Protection from extinction of human fear

conditioning by an avoidance response. Behaviour Research and Therapy 47, pp.

716-720.

PT
Mackintosh, N. J. (1975). A theory of attention: Variations in the associability of stimuli

RI
with reinforcement. Psychological Review, 82, 276-298.

Melchers, K. G., Lachnit, H., & Shanks, D. R. (2004). Within-compound associations in

SC
retrospective revaluation and in direct learning: A challenge for comparator

theory. Quarterly Journal of Experimental Psychology, 57B, 25–53.

U
Morell, J. R., & Holland, P. C. (1993). Summation and transfer of occasion setting.
AN
Animal Learning & Behavior, 21, 145-153.

Mowrer O. H. (1951). Two-factor learning theory: summary and comment. Psychological

M

Review, 58, 350–354. doi.org/10.1037/h0058956

D

Mueller, D., Bravo-Rivera, C., & Quirk, G. J. (2010). Infralimbic D2 receptors are
TE

necessary for fear extinction and extinction-related tone responses. Biological

psychiatry, 68(11), 1055-1060. doi.org/10.1016/j.biopsych.2010.08.014

EP

Myers, K. M., & Davis, M. (2007). Mechanisms of fear extinction. Molecular psychiatry,

12(2), 120-150. doi:10.1038/sj.mp.4001939

C

Meyer, H. C., Putney, R. B., & Bucci, D. J. (2015). Inhibitory learning is modulated by
AC

nicotinic acetylcholine receptors. Neuropharmacology, 89, 360-367.

doi.org/10.1016/j.neuropharm.2014.10.025

Meulders, A., Van Daele, T., Volders, S., & Vlaeyen, J. W. (2016). The use of safety-

seeking behavior in exposure-based treatments for fear and anxiety: Benefit or

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 27

burden? A meta-analytic review. Clinical psychology review, 45, 144-156.

doi.org/10.1016/j.cpr.2016.02.002

Pearce, J. M., & Hall, G. (1980). A model for Pavlovian learning: Variations in the

effectiveness of conditioned but not of unconditioned stimuli. Psychological

PT
Review, 87, 532-552.

RI
Pearce, J. M., & Wilson, P. N. (1991). Failure of excitatory conditioning to extinguish the

influence of a conditioned inhibitor. Journal of experimental psychology. Animal

SC
Behavior Processes, 17, 519-529.

Quirk, G. J., & Mueller, D. (2008). Neural mechanisms of extinction learning and

U
retrieval. Neuropsychopharmacology, 33, 56-72. doi:10.1038/sj.npp.1301555
AN
Rescorla, R. A. (1986). Extinction of facilitation. Journal of Experimental Psychology:

Animal Behavior Processes, 12(1), 16. doi.org/10.1037/0097-7403.12.1.16

M

Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian conditioning: Variations

D

in the effectiveness of reinforcement and nonreinforcement. A.H. Black, W.F.

TE

Prokasy (Eds.), Classical conditioning II: Current research and theory, Appleton-

Century-Crofts, New York (1972), pp. 64–99

EP

Rodriguez-Romaguera, J., Greenberg, B. D., Rasmussen, S. A., & Quirk, G. J. (2016). An

avoidance-based rodent model of exposure with response prevention therapy for

C

obsessive-compulsive disorder. Biological Psychiatry.

AC

doi.org/10.1016/j.biopsych.2016.02.012

Scheveneels, S., Boddez, Y., Vervliet, B., & Hermans, D. (2016). The validity of

laboratory-based treatment research: Bridging the gap between fear extinction and

exposure treatment. Behaviour Research and Therapy, 86, 87-94.

 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 28

Schultz, W. (2007). Behavioral dopamine signals. Trends in neurosciences, 30(5), 203-

210. doi.org/10.1016/j.tins.2007.03.007.

Shiflett, M. W., Riccie, M., & DiMatteo, R. (2013). The effects of amphetamine

sensitization on conditioned inhibition during a Pavlovian–instrumental transfer

PT
task in rats. Psychopharmacology, 230(1), 137-147. doi:10.1007/s00213-013-

RI
3144-3

Solomon, R. L., Kamin, L. J., and Wynne, L. C. (1953). Traumatic avoidance learning:

SC
the outcomes of several extinction procedures with dogs. J. Abnorm. Soc.

Psychol. 48, 291–302. doi: 10.1037/h0058943

U
Steinberg, E. E., Keiflin, R., Boivin, J. R., Witten, I. B., Deisseroth, K., & Janak, P. H.
AN
(2013). A causal link between prediction errors, dopamine neurons and learning.

Nature neuroscience, 16(7), 966-973. doi:10.1038/nn.3413

M

Vansteenwegen, D., Hermans, D., Vervliet, B., Francken, G., Beckers, T., Baeyens, F., &
D

Eelen, P. (2005). Return of fear in a human differential conditioning paradigm

TE

caused by a return to the original acquisition context. Behaviour Research and

Therapy, 43, 323-336. doi.org/10.1016/j.brat.2004.01.001

EP

van Vooren, P. R., Franssen, M., Beckers, T., Hermans, D., & Baeyens, F. (2012).

Narrowing down the conditions for extinction of Pavlovian feature-positive

C

discriminations in humans. Learning & behavior, 40(4), 393-404.

AC

doi:10.3758/s13420-011-0060-4

Vervliet, B., & Indekeu, E. (2015). Low-cost avoidance behaviors are resistant to fear

extinction in humans. Frontiers in Behavioral Neuroscience, 9, 351. doi:

10.3389/fnbeh.2015.00351
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 29

Yoon, T., Graham, L. K., & Kim, J. J. (2011). Hippocampal lesion effects on occasion

setting by contextual and discrete stimuli. Neurobiology of learning and memory,

95(2), 176-184. doi.org/10.1016/j.nlm.2010.07.001

PT
Yi, F., Catudio-Garrett, E., Gábriel, R., Wilhelm, M., Erdelyi, F., Szabo, G., ... &

RI
Lawrence, J. (2015). Hippocampal “cholinergic interneurons” visualized with the

choline acetyltransferase promoter: anatomical distribution, intrinsic membrane

SC
properties, neurochemical characteristics, and capacity for cholinergic

modulation. Frontiers in synaptic neuroscience, 7, 4.

doi.org/10.3389/fnsyn.2015.00004
U
AN
Zelikowsky, M., Hast, T. A., Bennett, R. Z., Merjanian, M., Nocera, N. A., Ponnusamy,

R., & Fanselow, M. S. (2013). Cholinergic blockade frees fear extinction from its
M

contextual dependency. Biological psychiatry, 73(4), 345-352.

D

doi.org/10.1016/j.biopsych.2012.08.006
TE

Zimmer-Hart, C. L., & Rescorla, R. A. (1974). Extinction of Pavlovian conditioned

inhibition. Journal of Comparative and Physiological Psychology, 86, 837-845.

C EP
AC
 ACCEPTED MANUSCRIPT
Treating Avoidance Behavior 30

Conditional
Inhibitor
US

PT
CS+

RI
Figure 1: Relationship of a conditional inhibitor to a US. Dashed lines represent direct

SC
inhibitor associations, whereas solid lines represent direct excitatory associations.

U
AN
Negative
Occasion
Setter
M

CS+ US
D
TE

Figure 2: Modulatory impact of a negative occasion setter on CS/US relationship.

C EP
AC
 ACCEPTED MANUSCRIPT

Highlights

• Avoidance behavior can persist following extinction or exposure

• The availability of avoidance behavior can renew fear
• Learning theory provides an understanding of avoidance behavior that may
suggest novel behavioral and pharmacological interventions

PT
RI
U SC
AN
M
D
TE
C EP
AC