Brain Research Bulletin 67 (2005) 343–354

NeuroEconomics: An overview from an economic perspective

P. Kenning ∗ , H. Plassmann
Department of General Management, University of Münster, Münster, Germany

Available online 11 August 2005

Abstract

Until now, economic theory has not systematically integrated the influence of emotions on decision-making. Since evidence from neuro-
science suggests that decision-making as hypothesized in economic theory depends on prior emotional processing, interdisciplinary research
under the label of “neuroeconomics” arose. The key idea of this approach is to employ recent neuroscientific methods in order to analyze
economically relevant brain processes. This paper aims to provide an overview of the current state of neuroeconomic research by giving a
brief description of the concept of neuroeconomics, outlining methods commonly used and describing current studies in this new research
area. Finally, some future prospects and limitations are discussed.
© 2005 Elsevier Inc. All rights reserved.

Keywords: Decision-making; Neuroscience; Economic theory

1. Introduction For instance, the integration of psychology led to the devel-

opment of behavioral economics [15,59], out of which sev-
Economics traditionally conceptualized a world popu- eral new models on decision-making evolved. A well-known
lated by rational, self-interest guided, unemotional maximiz- example is depicted in Fig. 1. It contains a dual-process model
ers. Despite the undisputable success of economic models that distinguishes between the two systems “intuition” and
based on the homo oeconomicus concept, their limitations “reasoning”.
are nowadays hard to ignore. Experimental and behavioral This model was introduced to explain seemingly contra-
economics have repeatedly revealed deviations from this dictory results of experimental studies of judgments under
classical theory. Subjects show non-opportunistic or recip- uncertainty. Despite their unquestionable explanatory value,
rocal behavior and other “anomalies” and “paradoxes” that “intuition” and “rationality” are theoretical constructs. The
are not explicable with the traditional concept [17]. Conse- existence of these underlying cognitive systems was pos-
quently, psychological ideas were formalized and translated tulated from observation and analysis of behavior, which
into testable predictions, thus leading to extended models of in turn is used to explain behavior. Criticism from tradi-
economic behavior. Rather than ignoring non-rational behav- tional economics concerns this circular reasoning as well
ior, concepts like bounded rationality, bounded willpower as the fact that – in contrast to market shares, unemploy-
or bounded self-interest arose and it was realized that eco- ment rates and economic growth – theoretical constructs
nomic behavior is frequently influenced by emotions and are neither observable nor objectively measurable. The only
subconscious processes [5,39]. Since these “anomalies” were way to solve this dilemma was to use tools or methods
replicated in field experiments [44] focusing on the homo that enable researchers to investigate behavior in a new,
oeconomicus does not seem to be a promising option for more objective way. Since neuroscience provides these tools,
advanced economic research [16,70,76]. researchers in several disciplines started to use neuroscien-
Consequently, in the early 1970s extended economic mod- tific tools in order to observe brain activities that underlie
els were developed using inputs from other scientific fields. behavior.
However, the approach to analyze behavior from a
∗ Corresponding author. Tel.: +49 251 83 25021; fax: +49 251 83 22808. neurobiological perspective offers different possibilities.
E-mail address: Peter.Kenning@wiwi.uni-muenster.de (P. Kenning). Our ability to explain behavior scientifically with any kind of

0361-9230/$ – see front matter © 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.brainresbull.2005.07.006
344 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354

2. Overview of selected neuroimaging techniques

currently used in neuroeconomics

In principle all kinds of neuroscientific tools can be used

to investigate economic decision-making. Table 1 summa-
rizes the prevalent methods currently employed in neuroe-
conomic research. They can be roughly grouped into two
categories according to the underlying mechanisms: proce-
dures for measuring electromagnetic activity of the brain
and those sensitive to changes of cerebral blood flow or
metabolism.

2.1. Electromagnetic recordings

2.1.1. Electroencephalography (EEG)

EEG measures voltage fluctuations on the scalp. The
underlying ion currents occur rather remote from the elec-
Fig. 1. Dual process model taken from [36]. trodes (across skin, skull and meninges) in surface-near
cortex areas and result from changes in membrane con-
ductivity elicited by synaptic activity and intrinsic mem-
analysis depends on the complexity of the respective system brane processes [46]. An electrode on the skin virtually
on one hand, and the available scientific tools on the other “sees” the summed potentials generated by a large number of
hand. neurons.
It is a futile attempt to try and explain the function of a While EEG with a temporal resolution of milliseconds and
human brain based on modeling individual activities of all below can easily detect the time course of neural activity,
estimated 1011 neurons and their interactions. Models must spatial resolution is mainly limited by the so-called inverse
reduce this complexity by defining larger structures as func- problem [35]. As an infinite number of source configurations
tional units. With presently available methods, neuroscience can generate identical potentials on the skin, estimated solu-
allows to investigate brain structures that can be looked upon tions of the inverse problem, i.e. source localization, therefore
as such functional units. require appropriate a priori assumptions about sources and
Attempting to link both disciplines – economics and neu- volume conduction to yield physiologically meaningful data
roscience – it becomes evident that they have to learn a lot [4,51].
from each other [5,71]. The protagonists of this interpretation
defined neuroeconomics as the application of neuroscientific
methods to analyze and understand economically relevant 2.1.2. Magnetoencephalography (MEG)
behavior [16]. This definition represents the current main- MEG is sensitive to changes of magnetic fields that are
stream understanding of neuroeconomics. Compared to the induced by the electrical brain activity. The temporal res-
model of the homo oeconomicus it implies a totally different olution can be compared to that of the EEG, so that this
idea of man. From a neuroscientific perspective his counter- modality can, e.g. resolve the temporal sequence of dif-
part is the homo neurobiologicus, whose behavior and social ferent cortical activities involved in decision-making [9].
and economic nature are the result of neurobiology. It is the However, in contrast to the EEG, MEG is also able to
latter that largely determines his thinking and feeling, decid- depict activity in deeper brain structures [2,8,9]. The inverse
ing and acting, as well as his buying and selling, i.e. his problem basically applies to MEG as well, so that source
economic life [5]. In this understanding neuroeconomics can localization depends on valid assumptions, too. Integrat-
contribute to create models of economy that are based on a ing the different brain imaging techniques could further
realistic description of human behavior and the comprehen- improve currently existing models of source localization
sion of the driving forces of this behavior. [46].
Appreciating possible contributions of neuroscience to
economic modeling requires a basic understanding of the Table 1
applied neuroscientific methods such as functional imag- Frequently used neuroimaging techniques
ing of the brain. Section 2 therefore provides an intro- Changes in electric activity Changes in cerebral blood
duction to these methods for economists. Results from flow/metabolism
current neuroeconomic research are outlined in Section Electroencephalography (EEG) Positron emission tomography
3. Possible implications for economic theory will be dis- (PET)
Magnetoencephalography (MEG) Functional magnetic resonance
cussed and future research prospects will be envisioned in
imaging (fMRI)
Section 4.
 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354 345

2.2. Methods measuring metabolic or hemodynamic Temporal and spatial resolution of fMRI depends on scan-
responses to neural activity ning technology on the one hand and on the underlying
physiology of the detected signal intensity changes on the
2.2.1. Positron emission tomography (PET) other hand.
Positrons, the antiparticles of electrons, are emitted by cer- With current scanner technology structural images
tain radio-nuclides. These nuclides have the same chemical are usually obtained with a resolution of about
properties as their non-radioactive isotopes and can replace 1 mm × 1 mm × 1 mm voxels (the equivalent of a pixel
the latter in biologically relevant molecules. After injection or (picture element) in a volume), while fMRI voxels typically
inhalation of tiny amounts of these modified molecules, e.g. have edge lengths of about 3–5 mm. As a rule temporal
modified glucose (FDG) or neurotransmitters, their spatial resolution of fMRI is between 1 and 3 s. For evaluation one
distribution can be detected by a PET-scanner. This device is must take into account that the cerebral blood flow (“CBF”)
sensitive to radiation resulting from the annihilation of emit- response to a brain activation is delayed by about 3–6 s.
ted positrons when they collide with ubiquitously present The unquestionable attractiveness of the method results
electrons. From the detected distribution information con- from a balance between temporal and spatial resolution on
cerning metabolism or brain perfusion can be derived and the one hand, allowing whole brain scans in less than 3 s,
visualized in tomograms. Spatial resolution is quite high and on its non-invasiveness on the other hand, permitting
(about 3–6 mm), but temporal resolution is low (several min- repeated measurements in healthy volunteers ad libitum. In
utes to fractions of an hour). As radioactive tracers are used, addition, the choice of scanning parameters allows increas-
the application to healthy test persons is restricted. ing one parameter at the expense of the other. Recent fMRI
approaches show that for some neural systems the temporal
2.2.2. Functional magnetic resonance tomography resolution can be improved down to milliseconds [56] and
(fMRI) spatial resolution can be increased to the level of cortical
fMRI is currently the most frequently used functional columns as basic functional units of the cortex [37].
brain imaging technique. Magnetic resonance scanners pro- In summary, all currently available neuroimaging tech-
duce sets of cross sections – tomograms – of the brain, exploit- niques have advantages and disadvantages. While the where
ing weak but measurable resonance signals that are emitted of brain activity is more easily assessed by fMRI or PET, the
by tissue water in a very strong magnetic field after excitation question of when – e.g. the discrimination between parallel
with a high frequency electromagnetic pulse. The acquired and sequential processing – can be more precisely answered
resonance signals can be attributed to their respective spa- by EEG or MEG. Thus, the method (or even a combination
tial origin, and cross sectional images can be calculated. The of several techniques) should be chosen carefully, depending
signal intensity, coded as gray value of a picture element, on the economic question to be answered.
depends on water content and certain magnetic properties of
the local tissue. In general, structural MR imaging is used to 2.3. An outline of neuroeconomic studies
depict brain morphology with good contrast and high reso-
lution. 2.3.1. Study design
The ability to visualize brain function by MRI depends on At the beginning of a neuroeconomic project a meaningful
the fact that increased neural activity of a brain region is fol- problem or hypothesis must be formulated. Already this first
lowed by a hemodynamic response. Albeit the mechanisms of step is sometimes difficult to accomplish, because scientists
this so-called neurovascular coupling are still not fully under- of both involved fields – i.e. economists and neuroscientists
stood, the increased perfusion of activated brain tissue is the – need to cross the borders between different approaches and
basis of the so-called Blood Oxygenation Level Dependent terminologies to be able to communicate and understand each
(BOLD)-effect: hemoglobin, the oxygen carrying molecule other’s methods and scientific problems.
in blood, has different magnetic properties depending on its After developing a hypothesis the applicability of neuro-
oxygenation state. While oxy-hemoglobin is diamagnetic, science, i.e. its ability to answer the question specified should
deoxy-hemoglobin is paramagnetic, i.e. it locally distorts the be ascertained. Subsequently, the most suitable method or
magnetic field, leading to a local signal loss. In activated brain combination of methods needs to be determined. Translat-
tissue the increased oxygen consumption is overcompensated ing the scientific question into a sound stimulation paradigm
by the blood flow response. Thus, during activation deoxy- or study design, as devoid of possible confounders as possi-
hemoglobin is partly replaced by oxy-hemoglobin, leading ble, constitutes another critical issue. Various mistakes can be
to less distortion of the local magnetic field, i.e. to increased made right at the beginning of a study that may later obstruct
signal intensity. data analysis (see R. Savoy’s article in this issue).
In contrast to PET-scans, no absolute levels of brain activ-
ity can be measured. The well-known color-coded statistical 2.3.2. Comparing brains
parametric activation maps (SPMs) are generated from elabo- When neuroeconomic studies collect data from groups
rate statistical analyses of fMRI time series comparing signal of individuals, there are certain steps to be taken to enable
intensity during different activation states. data analysis across several individuals. Trivial as this might
346 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354

sound: like fingerprints there are no two brains identical to knowledge for quite some time. Before the emergence of non-
each other. In order to be able to compare activations across invasive neuroimaging techniques patients with brain lesions
individual subjects, the brains are usually spatially normal- caused by trauma or disease were studied. Since then it is
ized to a template brain, i.e. they are transformed so that they known that for instance the destruction of parts of the occip-
are similar in overall size and spatial orientation. Generally, ital lobe leads to cortical blindness, and aphasia can be a
the aim of this transformation is to bring homologous brain consequence of lesions in the temporal (Wernicke’s area) or
areas into the closest possible alignment. frontal lobe (Broca’s area). More complex functional losses
In this context the Talairach stereotactic coordinate system result from lesions of the ventromedial prefrontal cortex, as
is used. Talairach and Tournoux [74] introduced three impor- in the well-known case of Phineas Gage. When this part of
tant innovations: a coordinate system to identify a particular his frontal lobe was destroyed in an accident 1848 his per-
brain location relative to anatomical landmarks; a spatial sonality and social behavior were severely altered, whereas
transformation to match one brain to another; an atlas describ- his intellectual capabilities largely remained intact [73].
ing a standard brain, with anatomical and cytoarchitectonic Experiments with intra-operative direct brain stimulation
labels. The coordinate system is based on the identification of can be given as an example of other historic sources of knowl-
the line connecting the anterior commissure (AC) and poste- edge about localizing brain functions [58].
rior commissure (PC)—two relatively invariant fiber bundles Another, older, approach to describe brain localizations
connecting the two hemispheres (see Fig. 2). This line, the is derived from microscopic anatomy. The foundations were
AC–PC line, defines the y-axis of the brain coordinate sys- laid by Brodmann almost 100 years ago [13]. According to
tem. The origin is set at the AC. The z-axis is orthogonal to variations of cytoarchitecture in stained brain sections he
the AC–PC-line in foot–head direction and passes through divided the human cortex into 52 distinct areas, the Brod-
the interhemispheric fissure. The x-axis is orthogonal to both mann areas (BAs, Fig. 3).
the other axes and points from AC to the right. Any point in The convention of relating activation foci to these cytoar-
the brain can be identified relative to these axes, which define chitectonically defined areas was adopted early in func-
the Talairach coordinate system. tional imaging and results of activation localizations are
A lot of analysis software like, e.g. SPM currently uses often ascribed to BAs in neuroimaging studies (for exam-
templates created by the Montreal Neurological Institute ple, see [24]). However, their microanatomical (rather than
(MNI), based on the average of many normal MR scans. functional) derivation has to be kept in mind. Although for
Although similar, the Talairach and the MNI templates are some areas (e.g. the motor cortex, BA 4) there seems to be
not identical, and care has to be taken to assign localizations a good correlation between areas delineated by cytoarchitec-
given in MNI coordinates correctly to, e.g. cytoarchitectoni- ture and their function, this relation is less certain for other
cally defined brain areas like the Brodmann areas (see below). areas (for details, see [12]). A closer correlation may possibly
Due to the great variability of brain anatomy and function this be revealed by recent microanatomical approaches, utiliz-
problem is anything but trivial. A detailed description of this ing, e.g. receptor autoradiography to analyze the regional and
issue can, e.g., be found in [12]. laminar distribution of various neurotransmitter receptors in
order to delineate distinct brain areas [82].
2.3.3. Data interpretation Currently our knowledge about the localization of
The fact that certain brain functions, like speech or vision, even more complex brain functions (e.g. decision-making
are processed in dedicated brain areas has been common or emotions) is expanding at an enormous pace, which
is mainly due to the development of non-invasive brain
imaging research techniques. The sources of knowledge
mentioned above resemble only a fraction of the various
available methods, which as a whole provide the necessary
background for the interpretation of neuroeconomic studies.
It should be noted, however, that even the most refined
neuroscientific tools currently available are rather coarse
compared to the complexity of our central nervous system
and that we are far from an in-depth understanding of the
brain. For a rather compact review of the history and future
of Human Brain Mapping see, e.g., Ref. [65].

3. Current topics in neuroeconomic research

3.1. Preferences

Preferences play a substantial role in economic theory

Fig. 2. The AC–PC-line. [70,72] since they have a significant influence on economic
 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354 347

Fig. 3. Lateral and medial aspect of the human brain. The Brodmann areas (BAs) are represented by the various symbols.

decision-making (e.g. brand preferences). Moreover, pref- sions between different brands of sensorily nearly undis-
erence judgments, i.e. selecting one of several brands in a tinguishable consumer goods. As one result Deppe et al.
department store, occur frequently in every-day life. found reduced activation in the dorsolateral prefrontal, poste-
Until recently preference – as a theoretical term – could rior parietal and occipital cortices and the left premotor area
not be observed directly and had to be assessed for instance (Brodmann areas 9, 46, 7/19 and 6) only when the target
by questionnaires or observation of behavior. By now neural brand was the subjects’ favorite one. Simultaneously activity
correlates of preference were detected in several neuroimag- was increased in the inferior precuneus and posterior cin-
ing studies. For instance, Deppe et al. [24] could show that gulate (BA 7), right superior frontal gyrus (BA 10), right
in a simulated buying decision task between sensorily sim- supramarginal gyrus (BA 40) and most pronounced in the
ilar fast moving consumer goods only a subject’s preferred ventromedial prefrontal cortex (“VMPFC”, BA 10).
brand elicited a distinct mode of decision-making. In their There are some more studies that indicate that the VMPFC
fMRI-study, 22 subjects were asked to make binary deci- is a central neural substrate in preference judgment [49,57].
348 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354

In future it might be possible to reveal differences in corti- One of the players (the proposer) suggests how to split the
cal processing of seemingly similar preferences or show that money. The responder may accept or reject the offer. If he
ostensibly different preferences have a similar neural corre- accepts, both get the share agreed to, if he refuses the offer,
late [79]. For instance there might be preferences that are none of them receives any money. Obviously, a selfish homo
predominantly influenced by emotions or other more delib- oeconomicus will offer his opponent the smallest share pos-
erative ones. Based on the fact that highly emotional events sible because game theory suggests that the responder will
are often remembered clearly, vividly and with great detail, accept it. Yet in experiments conducted by game theorists the
one can hypothesize that preferences built on emotions are most frequent outcome is a fair (i.e. 50:50) share [17]. More-
more robust than preferences built on deliberation. Sharot et over, several studies report that in about 50% of all games
al. support this hypothesis [68]. responders who are offered some 20% off the total amount
A better understanding of preferences might be useful for choose to reject the offer although this means missing out
more concise predictions of market reactions (e.g. to adver- altogether [15]. Why do people do this? With fMRI Sanfey
tising) or estimations of elasticity of demand [16]. et al. were able to prove that perceived unfairness correlates
with activations in particular regions of the brain, i.e. the ante-
3.2. Utility and the rewarding system rior insula and the dorsolateral, prefrontal cortex (“DLPFC”)
[64]. What does this mean to economists?
Measuring the utility derived from a good objectively
1. Anterior insula activation is consistently seen in neu-
presents a problem for researchers, because utility is always
roimaging studies focusing on pain and distress [25],
subjective and depends on the circumstances. The neural
hunger and thirst [75] and autonomic arousal [19]. The
bases of the (expected and experienced) utility constitute a
insula has also been dealt with in surveys on emotion,
further facet of neuroeconomic research (for a delineation
in particular involvement in the evaluation and represen-
of the concept of utility and preference see [27]). However,
tation of specific negative emotional states [14]. Conse-
neuroeconomic methods might contribute to solve these prob-
quently, the activation in the anterior insula is a reflec-
lems, because in several studies where people gain some
tion of the responders’ negative emotional response to an
useful good (i.e. money, juice or other incentives) by judg-
unfair offer [64].
ments, activation can be observed in the so-called “reward
2. In contrast to the insula region, the DLPFC is usually
areas” of the brain ([29]; see also the article of Peterson et al.
linked to cognitive processes such as goal maintenance
in this issue). Therefore, the “feeling” of utility may correlate
and executive control [52]. Thus, the DLPFC activation
with the activation in the rewarding systems of the brain.
observed by Sanfey et al. may indicate that despite the
It is possible that activity in this rewarding system also
unfair offer, people still try to accomplish the goal of accu-
influences investor behavior, e.g. the general disposition
mulating as much money as possible. As indicated by the
to sell winners too early (see Ref. [69] and [55]). Invest-
higher rejection rates, an unfair offer is more difficult to
ing money and gaining returns might correlate with activa-
accept. Hence higher cognitive effort may be required to
tion in the rewarding systems of the brain, e.g. the ventral
overcome the strong emotional tendency to reject the offer
striatum [53] or more general the orbitofrontal prefrontal
[64].
cortex–amygdala–Nucleus Accumbens (NAc)-circuit [50].
Knutson and Peterson [40] find that monetary payoffs induce Conceivably, people in the ultimatum game experience
some activation in the Nucleus Accumbens. The NAc is conflicting emotions when confronted with unfair offers. Up
densely innervated by dopaminergic fibers originating from to a certain level, the DLPFC controls the action. Therefore,
neurons in the midbrain [53]. As it is possible that the greater one accepts the offer although it might not be fair. But if
release of dopamine after an unexpected reward leads to the offer decreases, the activation in the insula gets more
accepting a risk more easily [5] it is also possible that dysfunc- intense up to the point where emotion dominates and the
tion in the OFC-amygdala-NAc reward circuit might explain offer is rejected. Possibly, people behave this way in order to
extreme risk-seeking behavior in some cases. avoid negative feelings, i.e. the feeling of having accepted an
unfair offer. Neuroscientists suggest that this kind of behavior
3.3. The social brain: fairness, altruism and trust may be rooted in evolution [5]. This notion corresponds to
some statements in economic literature [7,54,78]. Moreover,
While preferences and utility are due to research on recent studies about evolution theory support this supposition
individual decision-making, neuroeconomics also covers the and demonstrate that in groups of people both fairness and
field of social behavior (e.g. cooperation). As outlined before, defection inevitably occur [26]. An fMRI-study conducted
in classical theoretical approaches the assumption of self- by McCabe et al. supports this assumption. Based on the evi-
interest is crucial to modeling economic decision-making. dence that the ability for mental state attribution involves the
However, if this assumption is juxtaposed to the data acquired prefrontal cortex, McCabe and colleagues hypothesized that
from studies in experimental economics, various problems this area is involved in integrating theory-of-mind-processing
arise (see e.g. [10,34,54]). In the ultimatum game, two play- and cooperative actions. Studying lesions implies that the
ers are offered a chance to win a certain amount of money. PFC is engaged in social cooperation [63]. Therefore, we
 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354 349

can conclude that there is a relation between the activity in correlates of trust using event-related hyperscan-fMRI (“h-
the PFC and the ability to cooperate with other humans. fMRI” which means that two volunteers are measured parallel
In a similar context the existence of altruistic punishment in two scanners), King-Casas et al. made a simple modifica-
arises. “Altruistic punishment” means that individuals punish tion to a single-exchange trust game to make the model more
others, although the punishment is costly for them and yields realistic by changing the originally single-round format to a
no material gain [31]. In an experimental economic study multi-round format in which the same two individuals (one
Fehr and Gächter hypothesized that free riding causes strong designated the “investor”, and the other the “trustee”) played
negative emotions. These emotions trigger (altruistic) pun- 10 consecutive rounds. Their findings suggest that especially
ishment [31]. In a complementary PET-study, de Quervain et the caudate nucleus is involved in trust-building and reci-
al. were able to prove that people derive “satisfaction” from procity in economic exchange [38]). The caudate nucleus
punishing norm violations and that the activation in the dorsal is commonly active when learning about relations between
striatum reflects the anticipated satisfaction from punishing stimuli and responses [66].
defectors [22]. In the light of these results one might say that
altruistic people who punish others for transgressing a rule 3.4. Insights on dynamic concepts: learning, memory
tend to reward themselves for their behavior. For this kind and knowledge
of behavior brain areas such as the striatum and the medial
prefrontal cortex have been shown to play an important role In western economies innovations are considered impor-
[39]. As evidenced by several MRI-studies on the prisoner’s tant for economic growth. Therefore, economic policy is
dilemma, cooperation is associated with activations in brain aiming to establish an innovative, knowledge-creating soci-
areas that have also been linked to reward processing [61]. ety [41]. One major concept in the context of knowledge
Integrating these results into studies on organizational behav- management is learning. So far, economic researchers know
ior to help explain the interplay of intrinsic and extrinsic little about the neural foundations of learning [76]. There-
reward, e.g. in the context of the motivation crowding effect fore, a theory of learning may be closely connected to the
[32], might be interesting. field of neuroeconomics [80]. A good starting point for this
The construct of trust might be based on concepts similar theory might be the prediction error hypothesis [50,60]. The
to altruism [20]. Trust is generally considered a precondition core concept of this hypothesis is the prediction error δ(Ti ).
to achieve economic growth, individual income, democratic It is defined as followed:
stability, government performance and, in general, cooper-
δ(Ti ) = r(Ti ) + γV (T(i+1) ) − V (Ti )
ation [3]. Since cooperation is a basic element of modern
societies, investigating trust is high on the agenda of eco- where r is the occurrence of rewards at a current instance Ti in
nomic research. In classical theoretical approaches trust is a sequence of time steps Ti [i = 1, 2, . . .], V(Ti ) an estimated
defined as a rational calculation. For instance, a specific value as a function of Ti and γ is a scale factor for future
model explains the existence of trust by means of three rewards. When δ is positive, then the estimation V(Ti ) was
variables: (1) p, the likelihood of making a profit (a gain). too pessimistic and vice versa. It is suggested that dopaminer-
This equals the probability that the recipient of trust actually gic neurons generate a prediction error signal correlates with
behaves in a trustworthy manner; (2) L (for losses), the poten- δ(Ti ) [5,60]. Several studies have been performed to test this
tial loss that occurs if trust is betrayed and (3) G (gain), the hypothesis with fMRI [50]. By now, fMRI experiments have
potential gain if the recipient of trust in fact proves to be trust- identified a set of reward related brain structures including
worthy [18]. The normative hypothesis of this model is that the OFC, the amygdala, the ventral striatum and the medial
an actor will always be trusted if the following assumption prefrontal cortex.
applies: If one defines knowledge as individual problem-solving
capability, it is evident that memory and learning are the
p × G > (1 − p) × L
fundamental bases of knowledge. For instance, if in an experi-
This simple model seems convincing but it leads to some ment consumers have to judge whether the price for a certain
kind of “trust mathematicum”, computed trust [77]. In prac- product is acceptable or not, they have to be aware of the
tice, a number of questions remain unanswered: how would common price. In practice it is well-known that people are
the ‘giver of trust’ determine the exact value of p in an not able to name prices for a wide range of products in all
unfamiliar situation, for example, at the beginning of a new kinds of business sectors [30]. Therefore, it seems possible
business relationship? If all actors by definition behave oppor- that price has only little influence on buying behavior. As any-
tunistically, where does trust evolve? And why does trust exist body can imagine, in most cases the opposite is true. How
between anonymous traders on the Internet [6]? It seems far can this be explained? Price knowledge for some products
more realistic that trust built on the basis of simple and emo- could be assumed to become an automatic process which is
tional heuristics and will be reciprocated within certain limits. typically hard to articulate as price-decision may be uncon-
By now, there are first studies that investigate the influence scious [62]. Or, as Bechara and Damasio pointed out [5]:
of hormones like oxytocin [81] or cortisol (see the article of “Thus knowledge without emotional signaling leads to the
Takahashi in this issue) on trust. In order to measure neural dissociation between what one knows or says and how one
350 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354

Table 2
Overview of first neuroeconomic studies (in chronological order)
Author Theoretical background Problem Method Results
Breiter et al. [11] Behavioral decision theory, Neural responses to expectancy fMRI Activation changes in the
prospect theory and experience of monetary gains sublenticular extended amygdala
and losses (SLEA) and orbital gyrus were
triggered by expected values of the
prospects. Responses to experience
of rewards increased monotonically
with monetary value in the nucleus
accumbens, SLEA, and thalamus.
Responses to prospects and outcomes
were generally, but not always, seen
in the same regions. Overlaps with
activation changes seen previously in
response to tactile stimuli, gustatory
stimuli, and euphoria-inducing drugs
were found.
McCabe et al. [47] Behavioral decision theory, Neural correlates of cooperative fMRI Within the group of cooperative
game theory, particularly behavior subjects the PFC showed activation
trust and willingness to changes when subjects are playing a
cooperate human than when they are playing a
computer. Within the group of
non-cooperators, no significant
activation changes in the PFC
between computer and human
conditions were found.
Erk et al. [29] Behavioral decision theory, Neural correlates of social fMRI Products symbolizing wealth and
social interactions rewards status lead to increased activity in
reward-related brain areas.
Smith et al. [71] Behavioral decision theory, Neural correlates of attitudes PET Participants turned out to be risk
game theory, in particular about monetary gains or losses averse in gains and risk-seeking in
ambiguity, risk, gains and and risk or ambiguity losses; and ambiguity-seeking in
losses neither gains nor losses. Interactions
between attitudes and beliefs trigged
neural activation changes in
dorsomedial and ventromedial brain
areas.
Sanfey et al. [64] Behavioral decision theory, Neural correlates of fMRI Unfair offers lead to activity changes
game theory, in particular decision-making processes in brain areas related to both emotion
ultimatum game during the Ultimatum Game and cognition. Increased activity in
anterior insula for rejected unfair
offers suggests an important role for
emotions in decision-making.
Ambler et al. [2] Behavioral decision theory Neural correlates of product MEG Brain activations in product choice
choices differed from those for height
discrimination and a positive
relationship between brand
familiarity and choice time was
found. Neural activation during
choice task involved brain areas
responsible for silent vocalization.
Decision processes took
approximately 1 s and can be seen as
two halves. The first period seems to
involve gender-specific problem
recognition processes. The second
half concerned the choice itself (no
gender differences).
Knutson and Peterson [40] Behavioral decision theory, Neural correlates of monetary fMRI Increasing monetary gains activates a
expected utility rewards, review of several studies subcortical region of the ventral
striatum in a magnitude-proportional
manner. This ventral striatal
activation is not evident during
anticipation of losses. Actual gain
outcomes instead activate a region of
 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354 351

Table 2 (Continued )
Author Theoretical background Problem Method Results
the medial prefrontal cortex. During
anticipation of gain, ventral striatal
activation is accompanied by feelings
characterized by increasing arousal
and positive valence.
de Quervain et al. [22] Behavioral decision theory, Neural bases of “altruistic PET Sanctions against defectors activate
altruism, cooperation punishment” reward processing brain regions.
McClure et al. [49] Behavioral decision theory, Neural correlates of preference fMRI For the anonymous task, a consistent
preferences for culturally familiar drinks neural response in the ventromedial
prefrontal cortex is reported that
correlates with subjects’ behavioral
preferences for these beverages. In
the brand-cued experiment, brand
knowledge for one of the drinks had
a dramatic influence on expressed
behavioral preferences and on the
measured brain responses.
McClure et al. [48] Behavioral decision theory, Neural correlates of immediate fMRI Two separate systems were found to
temporal preferences and delayed monetary rewards be involved. Parts of the limbic
system are activated by decisions
involving immediate rewards.
Activity changes in the lateral
prefrontal cortex and posterior
parietal cortex were triggered by
inter-temporal choices. The relative
engagement of the two systems is
directly associated with subjects’
choices, with greater relative
fronto-parietal activity when subjects
choose longer term options.
Deppe et al. [24] Behavioral decision theory, Influence of implicit brand fMRI Only the presence of a subject’s
preference decisions of information on individual favorite brand leads to a distinctive
consumers economic decisions mode of decision-making, activating
regions responsible for integrating
emotions.
King-Casas et al. [38] Behavioral decision theory, Neural correlates of trust Hyperscan-fMRI The authors suggest that the head of
game theory, trust game reciprocity and reputation in a the caudate nucleus processes
multi-round trust game information about the fairness of a
social partner’s decision and the
intention to repay with trust.
Abler et al. [1] Behavioral decision theory Neural correlates of omission fMRI The authors found a neural correlates
relative to receipt of reward of frustration in form of decreased
(frustration) activation in the ventral striatum and
increased activation in the anterior
insula and the right medial prefrontal
cortex.
Deppe et al. [23] Behavioral decision theory, Neural correlates of framing fMRI The credibility of ambiguous news
credibility judgments of effects and prejudgements headlines is biased by the magazine
news headlines in the brand, in which the news headline is
context of different published. These framing effects
magazine frames correlate with activation changes in
the medial prefrontal cortex.

decides to act”, e.g. at the point of sale. Therefore, explicit 3.5. Towards a neural theory of economic
price knowledge might follow some kind of inverted U-shape decision-making
as function of purchase-frequency. It is low for both, prod-
ucts whose prices we really do not know and products the Although the debate about the role of emotion, cognition,
prices of which we only know by intuition or, gut feeling. memory and information processing in decision-making has
In between it is high. Neuroeconomic research may provide grown over the years, scientists have yet to reach a consen-
some insights on this hypothesis by investigating the neural sus on these terms (see e.g. [5]). This holds particularly true
correlates of price knowledge. for the role of emotions in decision-making [5,43]. To the
352 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354

best of our knowledge there are just a few articles published studies carried out in this field so far (other synopsis can be
in major economic journals, which address the concept of found in [16] and [80]).
emotions explicitly [5,28,43,45]. Perhaps economists have
ignored the role of emotions in decision-making because of
its “checkered history”[5]. However, when defining emotions
4. Conclusion and remarks
as a collection of changes in body and brain conditions [5,21],
neuroeconomic studies seem to be a promising approach to
Although still a young discipline, neuroeconomics has
gain insights in the role of emotions in economic decisions.
made some interesting contributions to economic theory.
According to the literature in neuroscience, emotions play an
While first studies can be characterized as explorative
important role in the survival of man [5,23]. Consequently,
research, researchers now focus on concepts crucial to mod-
it seems to be quite logical, that many of our decision are
ern economic theory such as fairness, trust, altruism, mem-
influenced (not just biased in a negative sense) by emotions
ory, learning and knowledge. At present, investigations of
[33,80]. By localizing brain activity during specific decision
decision-making processes, which have already been per-
and relating these functions neuro-anatomically, methods of
formed for decades, in both disciplines separately, are of
functional brain imaging are able to visualize different disso-
particular importance. In order to gain a deeper understanding
ciated neural networks which are assumed to be responsible
of the ‘economic man’, neuroeconomics broadens the con-
for emotions in decision-making. It will be a challenging
cepts of behavioral economics by means of neuroscientific
task for neuroeconomics to identify emotions significant for
tools. These tools enable economic research on brain pro-
economic decision-making and the context in which they
cesses, which in economic research have so far been looked
occur.
upon as black box.
Introducing the somatic marker hypothesis Bechara and
Neuroeconomics deals with measurements of the actual
Damasio suggest that (economic) decision-making is influ-
behavior of man and related brain functions. It does not pro-
enced by somatic states that can be stimulated by so-called
pose ideal or optimized models of behavior or systems. Espe-
“primary” and “secondary inducers” [5]. “Primary inducers”
cially, as in the field of neuroeconomics several disciplines
are defined as learned states that cause pleasurable or aver-
cooperate, care must be taken to prevent misinterpretation of
sive states. “Secondary inducers” are entities generated by the
experimental results.
recall of a personal or hypothetical emotional event. Accord-
All measurement techniques employed to investigate brain
ing to this hypothesis several brain structures are involved in
functions in principle give physical information about spe-
triggering somatic states in decision-making. The amygdala,
cific physical measures. For most techniques it is generally
which is seen “as a critical substrate in the neural system nec-
accepted that these measures allow statements about brain
essary for triggering somatic states from primary inducers”,
functions, however, the measures remain indirect, and state-
plays a prominent role. In contrast, the VMPFC is crucial to
ments that are valid in a certain setting may not be valid at
triggering somatic states from secondary inducers. Unlike the
all in a different one. Brain function is still a field with many
amygdala response, which is sudden and habituates quickly,
open questions, and while there are functional relations that
responses of the VMPFC are deliberately slow and longer
are understood in general, the more detailed processes are
lasting.
still unclear.
Although it has not been tested in economic decision-
The fact should always be remembered that in a strict
making yet, some researchers found evidence in favor of
sense most experiments are only able to give evidence for a
the somatic marker hypothesis. In the aforementioned study,
correlation between fulfilling a specific task and brain activity
Deppe et al. report some significantly increased activity in
of specific areas. This must not be misunderstood as a proof
the VMPFC when subjects had to decide on their “first-
for an actual causal relation.
choice-brand” [24]. Another important finding is reported in
Similar to psychology, models can be built that allow to
this issue by Deppe et al. In an event-related fMRI study,
relate (normal or pathologic) behavior of man to individual
the authors show that individual activity changes in the
brain activities. However, it is always mandatory to state the
VMPFC correlate with the degree of susceptibility of an
limits, outside which the model is not valid, and one has
individual to a judgment bias. Furthermore, the article of
to be aware of the fact that also new findings may render a
Volz et al. published in this issue provides evidence that
previously accepted model invalid. All statements about brain
a part of PFC, the posterior frontomedian cortex (BA 8),
functions that control economic relevant behavior rely on:
correlates with uncertainty in decision-making. These find-
ings are consistent with the role of the VMPFC as sug-
gested by the somatic marker hypothesis and provide new • That the measured quantities give true values, not only
insights into how people address the problem of (economic) noise or results of systematic errors.
decision-making. • That the spatial and temporal assignment of measured
Until now there are only a few groups worldwide that quantities to brain structures is correct.
explicitly conduct “neuroeconomic” research on economic • That the model relating the measured quantities to brain
decicion-making. Table 2 shows a selection of the empirical activity is still valid in the actual measurement.
 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354 353

• That models and statements about typical functions of spe- Economics from University of Cologne, Department of Economics,
cific brain structures are correct and valid also in the actual 2004.
case. [7] G.E. Bolton, A. Ockenfels, ERC: a theory of equity, reciprocity, and
competition, Am. Econ. Rev. 90 (2000) 166–193.
• That the tested person actually reacts to the stimulus that [8] S. Braeutigam, S.P. Rose, S.J. Swithenby, T. Ambler, The dis-
shall be investigated and is not distracted by other activities tributed neuronal systems supporting choice-making in real-life situa-
(there is always a lot of confounding factors). tions: differences between men and women when choosing groceries
detected using magnetoencephalography, Eur. J. Neurosci. 20 (2004)
In the light of the relative imprecise knowledge about emo- 293–302.
tion related brain functions one has to resist the temptation to [9] S. Braeutigam, J.F. Stins, S.P. Rose, S.J. Swithenby, T. Ambler,
Magnetoencephalographic signals identify stages in real-life decision
generalize a result found in a specific setting. Even if it would
processes, Neural Plast. 8 (2001) 241–254.
fit perfectly to a proposed model, generalization might be [10] H. Brandstätter, W. Güth, Personality in dictator and ultimatum
premature. Other settings and other tasks may employ totally games, Cent. Eur. J. Operat. Res. 10 (2002) 191–216.
different reaction systems. [11] H.C. Breiter, I. Aharon, D. Kahneman, A. Dale, P. Shizgal, Func-
A critical review of own results and previous results is tional Imaging of Neural Responses to Expectancy and Experience
of Monetary Gains and Losses, 2001.
always necessary. Especially if own interpretations rely on
[12] M. Brett, I.S. Johnsrude, A.M. Owen, The problem of functional
previously published work, one has also to critically assess localization in the human brain, Nat. Rev. Neurosci. 3 (2002)
whether the interpretation of the older results is still valid in 243–249.
the light of later research. [13] K. Brodmann, Vergleichende Lokalisationslehre der Großhirnrinde
Keeping all this in mind neuroeconomic researchers may in ihren Prinzipien dargestellt auf Grund des Zellenbaues, Barth,
Leipzig, 1909.
reach their goal: To provide a descriptive decision-making
[14] A.J. Calder, A.D. Lawrence, A. Young, Neuropsychology of fear and
theory, which is not restricted to economic theory and more loathing, Nat. Rev. Neurosci. 2 (2001) 352–363.
realistic than that of the homo oeconomicus. [15] C.F. Camerer, Behavioral Game Theory Experiments in Strategic
Interaction, Princeton University Press, 2003.
[16] C.F. Camerer, G. Loewenstein, D. Prelec, Neuroeconomics: How
Neuroscience can Inform Economics, Working Paper, UCLA Depart-
Acknowledgements ment of Economics, Levine’s Bibliography, 2003.
[17] C.F. Camerer, R.H. Thaler, Anomalies: ultimatums, dictators and
We would like to thank Dr. Wolfram Schwindt, Dr. Har- manners, J. Econ. Perspect. 9 (1995) 209–219.
ald Kugel, Dr. Hagen Schiffbauer, and Dr. Michael Deppe [18] J.S. Coleman, Foundations of Social Theory, Cambridge, MA, 1990.
[19] H.D. Critchley, R. Elliot, C.J. Mathias, R.J. Dolan, Neural activity
for their tremendous help with this article. Furthermore,
relating to generation and representation of galvanic skin conduc-
we would like to thank Kristina Böcker, Brian Bloch, and tance responses: a functional magnetic resonance imaging study, J.
Prof. Sanjaya S. Gaur for their support and the Wilhelm- Neurosci. 20 (2000) 3033–3040.
Lorch Stiftung for funding the neuroeconomic research con- [20] R. Croson, N. Buchan, Gender and culture: international experimen-
ducted at our department. We are greatly indebted to Prof. tal evidence from trust games, Am. Econ. Rev. 89 (1999) 386–392.
[21] A.R. Damasio, The somatic marker hypothesis and the possible func-
Dr. Dieter Ahlert, Prof. Dr. Walter Heindel, and Prof. Dr.
tions of the prefrontal cortex, Philos. Trans. R. Soc. B Biol. 351
Erich Ringelstein for their continuous support. Last but (1996) 1413–1420.
not least, we would like to thank Brigitte Scho and Jörg [22] D.J.-F. de Quervain, U. Fischbacher, V. Treyer, M. Schellhammer,
Niessing. U. Schnyder, A. Buck, E. Fehr, The neural basis of altruistic pun-
ishment, Science 305 (2004) 1254–1258.
[23] M. Deppe, W. Schwindt, J. Krämer, H. Kugel, H. Plassmann, P. Ken-
ning, E.B. Ringelstein, Evidence for a neural correlate of a framing
References effect: Bias-specific activity in the ventromedial prefrontal cortex
during credibility judgments, Brain Res. Bull. (2005).
[1] B. Abler, H. Walter, S. Erk, Neural correlates of frustration, Neu- [24] M. Deppe, W. Schwindt, H. Kugel, H. Plassmann, P. Kenning, Non-
roreport 16 (2005) 669–672. linear responses within the medial prefrontal cortex reveal when
[2] T. Ambler, S. Braeutigam, J. Stins, S.P. Rose, S.J. Swithenby, specific implicit information influences economic decision making,
Salience and choice: neural correlates of shopping decisions, Psy- J. Neuroimaging 15 (2005) 171–182.
chol. Market. 21 (2004) 247–266. [25] S.W.G. Derbyshire, A.K.P. Jones, F. Gyulai, S. Clark, D. Townsend,
[3] R. Axelrod, W.D. Hamilton, The evolution of cooperation, Science L.L. Firestone, Pain processing during three levels of noxious stimu-
211 (1981) 1390–1396. lation produces differential patterns of central activity, Pain 73 (1997)
[4] F. Babiloni, F. Cincotti, C. Babiloni, F. Carducci, D. Mattia, L. 431–445.
Astolfi, A. Basilisco, P.M. Rossini, L. Ding, Y. Ni, J. Cheng, K. [26] M. Doebeli, C. Hauert, T. Killingback, The evolutionary origin of
Christine, J. Sweeney, B. He, Estimation of the cortical functional cooperators and defectors, Science 306 (2004) 859–862.
connectivity with the multimodal integration of high-resolution EEG [27] S.A. Drakopoulos, Two levels of hedonistic influence on microeco-
and fMRI data by directed transfer function, NeuroImage 24 (2005) nomic theory, Scott. J. Polit. Econ. 37 (1990) 360–379.
118–131. [28] J. Elster, Emotions in economic theory, J. Econ. Lit. XXXVI (1998)
[5] A. Bechara, A.R. Damasio, The somatic marker hypothesis: a neural 47–74.
theory of economic decision making, Games Econ. Behav. 25 (2005) [29] S. Erk, M. Spitzer, A.P. Wunderlich, L. Galley, H. Walter, Cultural
336–372. objects modulate reward circuitry, Neuroreport 13 (2002) 2499–2503.
[6] G. E. Bolton, E. Katok, A. Ockenfels, Trust Among Internet Traders: [30] H. Evanschitzky, P. Kenning, V. Vogel, Consumer price knowledge in
A Behavioral Economics Approach, No. 5, Working Paper Series in the German retail market, J. Prod. Brand Manag. 13 (2004) 390–406.
354 P. Kenning, H. Plassmann / Brain Research Bulletin 67 (2005) 343–354

[31] E. Fehr, S. Gächter, Altruistic punishment in humans, Nature 415 MRI at neural time scale down to milliseconds, Proc. Natl. Acad.
(2002) 137–140. Sci. U.S.A. 97 (2000) 11026–11031.
[32] B.S. Frey, R. Jegen, Motivation crowding theory, J. Econ. Surv. 15 [57] M.P. Paulus, L.R. Frank, Ventromedial prefrontal cortex activation is
(2001) 589–611. critical for preference judgments, Neuroreport 14 (2003) 1311–1315.
[33] P.W. Glimcher, A. Rustichini, Neuroeconomics: the consilience of [58] W. Penfield, E. Boldrey, Motor and sensory representation in the
brain and decision, Science 306 (2004) 447–451. cerebral cortex of man studied by electrical stimulation, Brain 60
[34] W. Güth, R. Schmittberger, B. Schwarze, An experimental analysis (1937) 389–443.
of ultimatum bargaining, J. Econ. Behav. Organ. 3 (1982) 367–388. [59] M. Rabin, A perspective on psychology and economics, Eur. Econ.
[35] H.L.F. Helmholtz, Ueber einige Gesetze der Vertheilung elektrischer Rev. 46 (2002) 657–686.
Strome in korperlichen Leitern mit Anwendung auf die thierisch- [60] A.D. Redish, Addiction as a computational process gone awry, Sci-
elektrischen Versuche, Ann. Phys. Chem. 89 (1853) 211–233. ence 306 (2004) 1944–1947.
[36] D. Kahneman, Maps of Bounded Rationality: A Perspective on Intu- [61] J.K. Rilling, T.R. Zeh, G.S. Berns, C.D. Kilts, A neural basis for
itive Judgment and Choice, Nobel Prize Lecture, Stockholm, 2002. social cooperation, Neuron 35 (2002) 395–405.
[37] S.G. Kim, T.Q. Duong, Mapping cortical columnar structures using [62] A.G. Sanfey, J.D. Cohen, Is knowing always feeling? PNAS 101
fMRI, Physiol. Behav. 77 (2002) 641–6444. (2004) 16709–16710.
[38] B. King-Casas, D. Tomlin, C. Anen, C.F. Camerer, S.R. Quartz, P.R. [63] A.G. Sanfey, R. Hastie, M.K. Colvin, J. Grafman, Phineas gauged:
Montague, Getting to know you: reputation and trust in a two-person decision making and the human prefrontal cortex, Neuropsychologia
economic exchange, Science 308 (2005) 78–83. 41 (2003) 1218–1229.
[39] B. Knutson, Sweet revenge? Science 305 (2004) 1246–1247. [64] A.G. Sanfey, J.K. Rilling, J.A. Aronson, L.E. Nystrom, J.D. Cohen,
[40] B. Knutson, R. Peterson, Neurally reconstructing expected utility, The neural basis of economic decision-making in the ultimatum
Games Econ. Behav. 25 (2005) 305–315. game, Science 300 (2003) 1755–1758.
[41] E. Kremp, J. Mairesse, Knowledge Management, Innovation, and [65] R.L. Savoy, History and future directions of human brain mapping
Productivity: A Firm Level Exploration Based on French Manu- and functional neuroimaging, Acta Psychol. 107 (2001) 9–42.
facturing CIS3 Data, National Bureau of Economic Research, Inc., [66] C.A. Seger, C.M. Concotta, The roles of the caudate nucleus in
NBER Working Papers: 10237, 2004. human classification learning, J. Neurosci. 25 (2005) 2941–2951.
[43] J.E. LeDoux, Emotion circuits in the brain, Ann. Rev. Neurosci. 23 [68] T.D. Sharot, R. Mauricio, Phelps, A. Elizabeth, How emotion
(2000) 155–184. enhance the feeling of remembering, Nat. Neurosci. 7 (2004)
[44] S. Lichtenstein, P. Slovic, Response-induced reversals of preference 1376–1380.
in gambling: a extended replication in Las Vegas, J. Exp. Psychol. [69] H.S. Shefrin, M. Statman, The disposition to sell winners too early
101 (1973) 16–20. and ride losers too long: theory and evidence, J. Finance 40 (1985)
[45] G. Loewenstein, Emotions in economic theory and economic behav- 777–791.
ior, Am. Econ. Rev. 90 (2000) 426–433. [70] P. Slovic, The construction of preference, Am. Psychol. 50 (1995)
[46] F. Lopes da Silva, Functional localization of brain sources using 364–371.
EEG and/or MEG data: volume conductor and source models, Magn. [71] K. Smith, J. Dickhaut, K. McCabe, J.V. Pardo, Neuronal substrates
Reson. Imaging 22 (2004) 1533–1538. for choice under ambiguity, risk, gains, and losses, Manag. Sci. 48
[47] K. McCabe, D. Houser, L. Ryan, V. Smith, T. Trouard, A functional (2002) 711–718.
imaging study of cooperation in two-person reciprocal exchange, [72] J.H. Sobel, Money pumps, Philos. Sci. 68 (2001) 242–258.
Proc. Natl. Acad. Sci. U.S.A. 98 (2001) 11832–11835. [73] A.T. Steegmann, Dr. Harlow’s famous case: the “impossible” acci-
[48] S.M. McClure, D.I. Laibson, G. Loewenstein, J.D. Cohen, Sepa- dent of Phineas P. Gage, Surgery 52 (1962) 952–958.
rate neural systems value immediate and delayed monetary rewards, [74] J. Talairach, P. Tournoux, Co-Planar Stereotaxic Atlas of the Human
Science 306 (2004) 503–507. Brain, Thieme Medical Publishers, Inc., New York, 1988.
[49] S.M. McClure, J. Li, D. Tomlin, K.S. Cypert, L. Montague, P.R. [75] P.A. Tataranni, J.-F. Gautier, K. Chen, A. Uecker, D. Bandy, A.D.
Montague, Neural correlates of behavioral preference for culturally Salbe, R.E. Pratley, M. Lawson, E.M. Reiman, E. Ravussin, Neu-
familiar drinks, Neuron 44 (2004) 379–387. roanatomical correlates of hunger and satiation in humans using
[50] S.M. McClure, M. York, P.R. Montague, The neural substrates of positron emission tomography, Med. Sci. 96 (1999) 4569–4574.
reward processing in humans: the modern role of fMRI, Neurosci- [76] R.H. Thaler, From homo economicus to homo sapiens, J. Econ.
entist 10 (2004) 260–268. Perspect. 14 (2000) 133–142.
[51] C.M. Michel, M.M. Murray, G. Lanz, S. Gonzalez, L. Spinelli, R. [77] F.A. von Hayek, The pretence of knowledge, Am. Econ. Rev. 79
Grave de Peralta, EEG source imaging, Clin. Neurophysiol. 115 (1989) 3–8.
(2004) 2195–2222. [78] E. Xiao, Daniel Houser, Emotion expression in human punishment
[52] E.K. Miller, J.D. Cohen, An integrative theory of prefrontal cortex behavior, PNAS 102 (2005) 7398–7401.
function, Ann. Rev. Neurosci. 24 (2001) 167–202. [79] R.B. Zajonc, Feeling and thinking: preferences need no inferences,
[53] P.R. Montague, S.E. Hyman, J.D. Cohen, Computational roles for Am. Psychol. 35 (1980) 155–175.
dopamine in behavioural control, Nature 431 (2004) 760–767. [80] P. Zak, Neuroeconomics, Philos. Trans. R. Soc. B Biol. 359 (2004)
[54] M.A. Nowak, K.M. Page, K. Sigmund, Fairness versus reason in the 1737–1748.
ultimatum game, Science 289 (2000) 1773–1775. [81] P.J. Zak, R. Kurzban, W.T. Matzner, The neurobiology of trust, Ann.
[55] T. Odean, Are investors reluctant to realize their losses? J. Finance N.Y. Acad. Sci. 1032 (2004) 224–227.
53 (1998) 1775–1799. [82] K. Zilles, N. Palomero-Gallagher, A. Schleicher, Transmitter recep-
[56] S. Ogawa, T.-M. Lee, R. Stepnoski, W. Chei, X.-H. Zhu, K. Ugurbil, tors and functional anatomy of the cerebral cortex, J. Anat. 205
An approach to probe some neural systems interaction by functional (2004) 417–432.