University of Zurich
Zurich Open Repository and Archive
Winterthurerstr. 190
CH-8057 Zurich
http://www.zora.uzh.ch
Year: 2010
Embedding reward signals into perception and cognition
Pessoa, L; Engelmann, J B
http://dx.doi.org/10.3389/fnins.2010.00017.
Postprint available at:
http://www.zora.uzh.ch
Posted at the Zurich Open Repository and Archive, University of Zurich.
http://www.zora.uzh.ch
Originally published at:
Pessoa, L; Engelmann, J B (2010). Embedding reward signals into perception and cognition. Frontiers in
Neuroscience, 4(17):?-?.
Embedding reward signals into perception and cognition
Abstract
Despite considerable interest in the neural basis of valuation, the question of how valuation affects
cognitive processing has received relatively less attention. Here, we review evidence from recent
behavioral and neuroimaging studies supporting the notion that motivation can enhance perceptual and
executive control processes to achieve more efficient goal-directed behavior. Specifically, in the context
of cognitive tasks offering monetary gains, improved behavioral performance has been repeatedly
observed in conjunction with elevated neural activations in task-relevant perceptual, cognitive and
reward-related regions. We address the neural basis of motivation-cognition interactions by suggesting
various modes of communication between relevant neural networks: (1) global hub regions may
integrate information from multiple inputs providing a communicative link between specialized
networks; (2) point-to-point interactions allow for more specific cross-network communication; and (3)
diffuse neuromodulatory systems can relay motivational signals to cortex and enhance signal
processing. Together, these modes of communication allow information regarding motivational
significance to reach relevant brain regions and shape behavior.
FOCUSED REVIEW
published: 15 September 2010
doi: 10.3389/fnins.2010.00017
Embedding reward signals into perception
and cognition
Luiz Pessoa1* and Jan B. Engelmann2
1
2
Edited by:
Anna C. Nobre, University of Oxford, UK
Reviewed by:
Tobias Egner, Duke University, USA
Anna C. Nobre, University of Oxford, UK
Department of Psychological and Brain Sciences, Indiana University, Bloomington, IN, USA
Institute for Empirical Research in Economics, University of Zurich, Zurich, Switzerland
Despite considerable interest in the neural basis of valuation, the question of how valuation
affects cognitive processing has received relatively less attention. Here, we review evidence from
recent behavioral and neuroimaging studies supporting the notion that motivation can enhance
perceptual and executive control processes to achieve more efficient goal-directed behavior.
Specifically, in the context of cognitive tasks offering monetary gains, improved behavioral
performance has been repeatedly observed in conjunction with elevated neural activations in
task-relevant perceptual, cognitive and reward-related regions. We address the neural basis
of motivation-cognition interactions by suggesting various modes of communication between
relevant neural networks: (1) global hub regions may integrate information from multiple inputs
providing a communicative link between specialized networks; (2) point-to-point interactions allow
for more specific cross-network communication; and (3) diffuse neuromodulatory systems can
relay motivational signals to cortex and enhance signal processing. Together, these modes of
communication allow information regarding motivational significance to reach relevant brain
regions and shape behavior.
Keywords: motivation, attention, executive function, fronto-parietal, posterior cingulate cortex
*Correspondence:
INTRODUCTION
Luiz Pessoa received his B.Sc. and M.Sc.
in Computer Science in Brazil, and a
Ph.D. in Computational Neuroscience at
Boston University (1996). From 1999
until 2003, he was a Visiting Fellow at the
National Institute of Mental Health,
Bethesda, in the laboratory headed by
Leslie Ungerleider. After a few years at
Brown University (2003–2006), he joined
the Department of Psychological and
Brain Sciences at Indiana University,
Bloomington, where he is an Associate
Professor. His interests center on the
interactions between cognition and
emotion/motivation in the human brain.
lpessoa@indiana.edu
Frontiers in Neuroscience
Navigating through the world requires the constant
assessment and reassessment of the value of our
choices and actions. Making one choice may cause
the loss of resources, such as food, but allow the
attainment of others, such as mating. This harsh
economic reality requires the brain to assess the
costs and benefits of actions before committing to a
motor response. Due to the ubiquity of such valuation processes and their importance for learning,
action selection, and choice, there has been ample
scientific interest in their neurobiological basis. This
line of research has yielded results underscoring the
importance of the dopamine system and its cortical
projection sites in behavioral control (Schultz et al.,
1992), has led to the formulation of computational
models of valuation (Montague et al., 2004), and
has contributed to the development of the novel
www.frontiersin.org
field of neuroeconomics (e.g., Platt and Glimcher,
1999; Berns et al., 2001).
Despite much interest in the neurobiological basis of valuation processes, the question of
how they interact with other cognitive systems
has received relatively little attention. If valuation
processes shape behavior, they should be expected
to influence the perceptual and cognitive processes that are central to the production of behavior.
Indeed, consistent with this notion, neuroimaging studies have shown that monetary incentives
improve behavioral performance, concurrent with
enhanced hemodynamic responses in task-relevant
perceptual and cognitive regions, as well as regions
of the reward system (e.g., Pochon et al., 2002;
Small et al., 2005; Engelmann et al., 2009). Based
on such findings, it has been proposed that the
neural interaction between reward-related regions
September 2010 | Volume 4 | Article 17 | 1
Pessoa and Engelmann
Embedding reward into executive function
Dprime (d′)
Perceptual sensitivity measure
commonly employed in signal detection
theory. By taking into account both hits
(e.g., correct target identification) and
false alarms (e.g., incorrectly identifying
distracters as targets) d′ scores account
for shifts in response criterion that are
commonly observed under different
reward conditions.
with perceptual and cognitive networks that are
task-relevant improves behavioral performance to
maximize reward outcomes (Pessoa, 2009).
Executive processes, such as attention, working
memory and inhibition, constitute a set of processes that are particularly important for behavioral
planning and production. Given limited processing
capacity, how can the brain separate those stimuli
that deserve further processing from those that are
better left ignored, to efficiently guide behavior?
Traditionally, both bottom-up and top-down processes are posited as potential solutions to the “limited processing-resources dilemma” faced by the
brain’s executive system. By using top-down control,
the brain can more efficiently allocate its resources
based on current behavioral goals and prior knowledge. At the same time, processing resources should
preferentially shift to salient features of the environment. Based on behavioral evidence, both of
these processes are intimately linked to reward and
motivation, as described below. Furthermore, these
findings mesh well with previous demonstrations
that the motivational dimensions of (top-down)
goals rely on the dopamine system and its projection sites (e.g., Schultz et al., 1992). Interestingly
(bottom-up) stimulus salience is also encoded in
specific nodes of the reward system, such as the
caudate nucleus (e.g., Zink et al., 2006).
EFFECTS OF REWARD ON EXECUTIVE
FUNCTION
PARAMETRIC INCREASES IN BEHAVIORAL
PERFORMANCE AND ACTIVATION STRENGTH IN THE
FRONTO-PARIETAL ATTENTIONAL NETWORK
Here, we review evidence indicating that motivational factors guide perceptual and executive
control processes, likely by modulating both bot-
Behavior
Detection Sensitivity (dp)
2.2
2
1.8
1.6
Exp. 1
1.4
0
Exp. 2
Cue Period
B
2.4
Exp. 3
1
2/2.5
4
0.5
Absolute Incentive Value ($)
Mean Parameter Estimate (a.u.)
A
0.04
0.02
0
–0.02
Attention Net
Reward Net
Vision Net
–0.04
–0.06
Target Period
C
0.06
4
0
1
2.5
Absolute Incentive Value ($)
FIGURE 1 | Behavioral and neural effects of incentive motivation. (A) In all
experiments, the detection sensitivity measure dprime (dp) increased as a
function of absolute incentive magnitude. Red line: experiment 1 of
Engelmann and Pessoa (2007); light orange line: experiment 2 of Engelmann
and Pessoa (2007); dark red line: behavioral results of Engelmann et al. (2009).
Frontiers in Neuroscience
tom-up and top-down processes, thereby helping
to solve the limited processing-resources dilemma.
In a series of experiments, Engelmann and Pessoa
(2007) and Engelmann et al. (2009) investigated
the effects of motivation on task performance
by probing the effects of parametric changes in
incentive value on behavior during difficult spatial localization tasks. Participants were asked to
indicate the location of a target stimulus (e.g.,
degraded face) relative to that of a distracter stimulus (e.g., random noise) as quickly and accurately
as possible. Attention was manipulated by using
a central cue that predicted target location with
70% validity (such that 30% of the time the cue
indicated the incorrect target location) – in such
cases, performance during validly cued trials is
known to exceed that during invalidly cued ones.
Motivation was parametrically manipulated in a
blocked fashion by linking payoff to behavioral
performance (if performance was both accurate
and fast in a given block of trials, participants were
given the chance to win cash incentives that varied
from $0–$4, or to avoid losing money).
Our behavioral findings revealed improved
detection performance as a function of absolute
incentive value (Figure 1A). Critically, because
behavior was characterized via the detection
sensitivity measure d′, the results revealed a
“specific” effect of motivation on behavioral
performance, instead of more unspecific influences such as arousal (e.g., purely faster response
times) or response bias (e.g., more conservative
responses) – but see below for further discussion
on more general effects that may be, at least in
part, linked to arousal. The same basic pattern of
behavioral results was observed in three distinct
versions of the task that varied in difficulty level,
Mean Parameter Estimate (a.u.)
Reward system
Typically denotes a collection of
interconnected structures that signal
information related to rewards (among
other functions). Major pathways
include the mesolimbic and
mesocortical dopamine systems, both
of which originate in the ventral
tegmental area and connect to multiple
subcortical and cortical regions.
Although we continue with the
common practice of employing the
term “reward system,” it should be
noted that its function is quite diverse
(e.g., regulation of effort and resource
allocation), subject to debate, and not
restricted to reward per se.
0.19
0.15
0.11
0.07
0.03
Attention Net
Reward Net
Vision Net
0
1
2.5
4
Absolute Incentive Value ($)
Parallel increases in evoked brain responses observed in the study by
Engelmann et al. (2009) during the cue (B) and target (C) task phases in three
types of regions, namely attentional, visual and reward-related (see Figure 2
for some of the sites). Results were obtained by pooling the responses from
regions within these three networks. Net = network.
www.frontiersin.org
September 2010 | Volume 4 | Article 17 | 2
Pessoa and Engelmann
Incentive motivation
Present when an actor engages in
effortful behavior to attain a valued
goal. Incentives can be primary rewards,
such as food items, or more abstract
rewards, such as money. Often, highlyvalued incentives, such as career goals,
require complex behaviors that demand
sustained motivation over prolonged
time periods.
Embedding reward into executive function
the type of target and distracter stimuli, and cue
types (endogenous vs exogenous).
One of the versions of the behavioral task was
accompanied by fMRI scanning (Engelmann et al.,
2009), allowing us to probe the neural basis of
enhanced performance by incentive motivation.
Specifically, we sought to elucidate the workings
of “process-specific” effects of motivation on cueand target-related processing during these attentional tasks. Non-specific motivational effects due
to effort and arousal were removed by using a
hybrid task design that included: (1) event-related
(i.e., transient) components with relatively long,
jittered and optimized intertrial and interstimulus
intervals between cue and target periods; and (2)
a blocked (i.e., sustained) motivational component. Hybrid designs allow for separate estimates
of transient and sustained signals (Visscher et al.,
2003). Importantly, transient processes could be
dissociated from each other, i.e., cue- and targetrelated responses.
In parallel with the behavioral findings,
the neuroimaging results revealed parametric
increases in activation strength as a function of
absolute incentive value in three types of brain
regions (Figure 2): (i) fronto-parietal sites that are
important for the control of attention, including
FIGURE 2 | Brain regions exhibiting correlations with absolute incentive magnitude during
the cue and target task periods. Some of the attentional (blue font), visual (light green), and
valuation (orange) regions are illustrated. ACC, anterior cingulate cortex; FEF, frontal eye field; IPS,
intraparietal sulcus; pre-SMA, pre-supplementary motor area; and preSMA, pre-supplementary
motor area.
Frontiers in Neuroscience
www.frontiersin.org
frontal eye field (FEF), anterior cingulate cortex
(ACC; and other sites along the midline), intraparietal sulcus (IPS) and temporo-parietal junction
(TPJ); (ii) occipito-temporal visual cortical sites,
including sites around the calcarine fissure and
in the fusiform gyrus, a region that is sensitive to
face stimuli (which were employed in the task);
and (iii) nodes of the reward system, including
caudate and substantia nigra (SN)/midbrain.
Parametric influences of incentive motivation on
evoked responses were obtained during both the
cue (Figure 1B) and target (Figure 1C) periods.
In particular, our findings concerning rewardrelated sites are consistent with previous reports
of parametric response increases in the nucleus
accumbens (e.g., Knutson et al., 2005), caudate
nucleus (Delgado et al., 2003) and orbitofrontal cortex (e.g., O’Doherty et al., 2001). Taken
together, our observations revealed that parametric improvements in detection performance
were accompanied by systematic modulations in
three sets of brain regions that, together, support
task performance, namely attentional, visual and
reward-related regions.
CONVERGING EVIDENCE: MOTIVATIONAL EFFECTS
ON COGNITIVE AND SENSORY PROCESSING
Our findings support the notion that motivational signals act both at more “central” levels
in fronto-parietal cortex and at sensory levels.
Here, we briefly discuss other studies that support this view and, in particular, have evaluated
how motivation influences cognitive function. In
a study by Small et al. (2005), fast target detection could lead to monetary wins or avoidance
of monetary losses and, in the control condition,
did not involve monetary outcomes. Better performance during the disengagement of attention was associated with enhanced activity in the
inferior parietal lobe in the vicinity of the TPJ,
a region that has been implicated in the reorienting of attention. Importantly, this effect was
enhanced by incentive motivation during trials
in which participants could win or avoid losing
money, and were accompanied by activations in
valuation-related regions, including the orbitofrontal cortex. Of particular interest, responses in
the posterior cingulate cortex (PCC) were correlated with visual spatial expectancy (defined as the
degree to which the cue benefited performance as
evidenced by faster reaction times), an effect that
was enhanced by incentive motivation. Given the
known connectivity of this region with areas of
the brain involved in attention and motivational
processing, it was proposed that the PCC serves
as a neural interface between motivation and the
top-down control of attention.
September 2010 | Volume 4 | Article 17 | 3
Pessoa and Engelmann
Embedding reward into executive function
A subsequent study by Mohanty et al. (2008) also
investigated motivation effects on attention, this
time manipulating motivational state, namely hunger. Specifically, in the context of a covert-orienting
task with a central cue, participants detected motivationally relevant (food) or irrelevant (tools) targets
under conditions of hunger and satiety. As in the
study by Small et al., responses in sites in parietal
cortex (e.g., intraparietal sulcus, IPS) exhibited correlations with the speed of attentional shifts that
were sensitive to not just motivational state, but also
to the motivational value of the target. Similar patterns were also observed in the PCC and the orbitofrontal cortex (OFC). Furthermore, amygdala, PCC,
locus coeruleus and SN/midbrain showed sensitivity
to food-related cues when hungry, but not when
satiated, an effect that did not generalize to tools.
These findings demonstrate that motivational state
(hunger) modulates spatial attention via response
modulations across several brain regions.
Given that the findings from the above studies
are being explicitly related to those of our own
neuroimaging study, it is of relevance to ascertain
the degree of spatial concordance of the parietal
activation sites. In some cases, the concordance
was good when compared to other attentional
studies in the literature (Corbetta et al., 2000;
Hopfinger et al., 2000; Kincade et al., 2005), such
as target-related activations in the IPS (distance
between our study and relevant published reports:
∼6 mm). However, the concordance with the studies investigating attention and motivation per se
(Small et al., 2005; Mohanty et al., 2008) was less
impressive, such as ∼17 mm for the PCC, and even
∼23 mm for the TPJ. Hence, it will be important
in the future to understand the reasons behind
the sources of spatial variability.
The two studies reviewed above, in addition
to our own work, provide evidence that motivation modulates fronto-parietal regions involved in
attention. Additional evidence also supports the
modulation of sensory cortex by motivation. For
instance, Pantoja et al. (2007) investigated neuronal responses in the rat primary somatosensory
cortex (S1) during a tactile discrimination task.
Stimulus-related information encoded by S1
neuronal ensembles increased when the contingency between stimulus and response was crucial
for reward, but not when reward was freely available. In addition, stimulus-related information was
directly related to behavioral task performance.
Related neuroimaging findings in humans were
reported by Pleger et al. (2008, 2009), who used a
tactile discrimination task coupled with financial
rewards awarded for correct performance at the
end of each trial. While reward improved discrimination performance and concordantly enhanced
Frontiers in Neuroscience
www.frontiersin.org
activity in the ventral striatum, the effect of reward
on somatosensory responses was only observed in
a post-stimulus phase between stimulus offset and
reward delivery. Interestingly, the increase in somatosensory cortex responses varied parametrically as
a function of reward magnitude. In addition, the
effect of reward on somatosensory responses was
mediated by the dopaminergic system, as evidenced
via pharmacological manipulations (Pleger et al.,
2009). As observed in our own study, the contribution of motivational signals to sensory processing
extends to other sensory systems, with modulatory
signals detected at the level of the primary visual
cortex (V1) in both rats (Shuler and Bear, 2006)
and humans (Serences, 2008). Thus, it appears that
motivation not only modulates sensory processing, but that such influences are present at the
first stages of cortical processing. Naturally, such
effects likely reflect “late” contributions from other
processing stages (see next section).
Thus far, we have reviewed motivational effects
that appear to be more transient in nature; however, although relatively little is known about
sustained motivational signals, such modulations have also been observed. For instance, in
our experiment discussed above, we employed
an experimental design in which incentive was
manipulated in a blocked fashion, allowing us
to investigate sustained responses throughout
the block of trials and how they were modulated
by motivation. State-like effects were observed
in several brain regions, including sites in the
prefrontal cortex (PFC; e.g., FEF, middle frontal
gyrus), parietal cortex (e.g., IPS), in addition to
the PCC. Related findings were also reported by
Locke and Braver (2008) who reported increased
sustained fMRI activity during rewarded blocks
of a cognitive control task in a network of regions
including the right lateral PFC, right parietal cortex, and dorsal medial frontal cortex. Importantly,
in a recent study, Jimura et al. (in press) showed
that the effect of an individual’s sensitivity to
reward on working memory performance was
mediated by sustained effects of reward observed
in the right lateral PFC. These studies highlight
the importance of studying sustained effects of
motivation, which may be more closely related
to arousal processes. Indeed, future investigations seeking to unravel the contributions of both
transient and sustained responses to behavioral
performance are greatly needed.
POTENTIAL MECHANIMS OF MOTIVATIONAL
EFFECTS
Our study, together with the ones cited above
and several others, illustrates attempts at understanding how motivation influences cognitive and
September 2010 | Volume 4 | Article 17 | 4
Pessoa and Engelmann
Embedding reward into executive function
sensory processes. More generally, what are the
neural bases for these effects? At the outset, it is
instructive to consider the relationship between
motivation and cognition more abstractly. For
concreteness, we can consider attention as the cognitive task. We know that attention affects behavior,
and one possibility is that motivation has similar
effects that take place independently of attention
(Figure 3A). A second scenario would suggest that
motivation affects behavior by engaging the same
set of processes that are used by attention. In this
case, the impact of motivation on behavior could
be described as mediated by attention (Figure 3B).
This mediation could be partial only, such that
both direct (motivation → behavior) and indirect (via attention) effects take place. Finally, it is
possible to imagine situations in which attention
and motivation are more highly interactive, such
that they jointly influence behavior (Figure 3C).
In this latter case, although one may choose to
describe certain processes as “attentional” and others as “motivational”, the interactions between the
systems are sufficiently high, and their strict separation is possibly more semantic than real.
Methodologically, disentangling attention and
motivation may be quite hard, and many effects
attributed to motivation could be attentional and
vice versa (Maunsell, 2004). Although separating their potential contributions is challenging,
A Parallel
some evidence suggests that these processes may
be partially dissociable. For instance, Bendiksby
and Platt (2006) suggested that cell responses in
the lateral intraparietal area (LIP) in the monkey exhibit separate contributions from reward
and attention – in the context of a cued-saccade
reaction time task that was coupled with blocked
rewards. For example, neuronal activity was positively correlated with saccade reaction times,
which in their task was considered to reflect the
cost of attentional re-orienting, in a way that was
independent of reward size. At the same time,
modulation by reward size was independent
of which type of cue appeared in the neuronal
receptive field, with some cues being more predictive of target location than others, therefore
putatively capturing more attention. Thus, their
results are consistent with the idea that motivation and attention independently contribute to
responses in LIP – although a stronger case for
their separation would require converging evidence (e.g., in their study, other variables may
have affected saccade reaction times, and not only
attentional processes).
Both monkey electrophysiology and human
neuroimaging research suggest that the control
of selective attention relies on a distributed set of
fronto-parietal regions, including FEF in frontal
cortex and IPS in parietal cortex. These regions,
D Modes of communication
Fronto-parietal
attentional network
Attention
Behavior
Motivation
1
B Mediation
Attention
Motivation
1
2
2
Behavior
2
C Integration
Valuation network,
cortical:
OFC, ant. insula, ACC,
PCC, etc.
Attention
3
Neuromodulatory:
Midbrain: VTA, SN
2
Valuation network,
subcortical:
Caudate, putamen, NAcc,
amygdala, etc.
Behavior
Connector hub region
Motivation
Frontoparietal region
FIGURE 3 | Mechanisms of motivational effects on attention. (A–C)
Potential, abstract relationships between attention and motivation and their
effects on behavior. (D) Modes of communication between cognitive and
motivation networks illustrated for attentional-motivational interactions. (1)
Interactions rely on “hub” regions, such as the anterior cingulate cortex, which
Frontiers in Neuroscience
are part of both attentional and motivational networks (indicated via the red
outline in both the valuation-cortical and attentional networks). (2) In addition,
specific regions may link the two networks, either directly or via the thalamus.
(3) Finally, motivational signals are embedded within cognitive mechanisms via
the action of diffuse neuromodulatory systems.
www.frontiersin.org
September 2010 | Volume 4 | Article 17 | 5
Pessoa and Engelmann
Hubs
Regions of high connectivity that have a
disproportionately large impact on
regulating information flow. Hubs are
regions within neural networks that,
through their privileged connectivity
patterns, are able to integrate
information and influence the
processing of multiple connecting
regions, thus greatly influencing brain
dynamics.
Frontiers in Neuroscience
Embedding reward into executive function
which in many cases appear to work together,
are often conceptualized as “source” regions that
exert control over sensory-processing areas to
help select the information that is most relevant
at a given time. One way to interpret the data
from the attention studies described in the previous sections is to suggest that motivation acts
on cognition to maximize potential reward in a
way that relies on robust interactions between
the attentional network and other reward/valuation networks. Among others, valuation regions
include: (i) subcortically: the caudate, putamen
and nucleus accumbens in the ventral striatum,
and the amygdala; and (ii) cortically: the OFC,
anterior insula, ACC and PCC. During trials in
which reward (or punishment avoidance) is possible, valuation and attentional networks interact,
resulting in enhanced behavioral performance
that is supported by improved selection of sensory information. Critically, reward-related effects
on cognitive function are specific (e.g., increased
detection performance), as opposed to global (e.g.,
arousal). Whereas “independent” contributions
from attention and motivation (Figure 3A) are
not necessarily excluded, the above considerations
are suggestive of the “mediation” and “integration”
scenarios (Figures 3B,C) described above.
Indeed, we would like to propose that, more
generally, the “integration” needs to be more seriously considered. Accordingly, the integration of
motivational signals with those that are central to
specific executive functions, including task switching, inhibition, and information maintenance,
will rely on interactions between specific “cognitive” networks and those involved in determining
the behavioral significance of the stimulus or task
at hand. For instance, both dorsal (e.g., middle
frontal gyrus) and ventral (e.g., inferior frontal
gyrus) PFC sites, in addition to regions of parietal
cortex (e.g., IPS), are important for maintaining
and updating contextually relevant information
“in mind.” As in the case of the control of attention, we suggest that working memory-related
signals are integrated with motivational ones in
these areas. Consistent with this notion, cells in
monkey lateral PFC not only hold information
concerning an object’s shape and location, but are
also modulated by reward magnitude (Leon and
Shadlen, 1999; Watanabe and Sakagami, 2007).
Human neuroimaging studies have shown similar modulations of working memory signals in
lateral PFC by reward (e.g., Pochon et al., 2002;
Taylor et al., 2004). Furthermore, motivational
information does not act simply as an “additive”
mechanism; instead, in lateral PFC, cognitive and
motivational signals appear to be integrated (see
Jimura et al. (in press). For instance, in monkeys,
www.frontiersin.org
during the delay period of a task involving spatial information, spatial and reward information
do more than just add, as there is an increase of
the amount of transmitted information concerning target position, as quantified by information
theory (Kobayashi et al., 2002). In other words,
reward information increases the discriminability of target positions, leading to enhanced performance. In the context of our own studies, in
sharp contrast with other proposals (Kouneiher
et al., 2009), we have suggested that the effect of
motivation goes well beyond an “energizing” (i.e.,
a generalized “additive”) function and, instead,
involves enhancing and/or optimizing executive
function (see also below) – a notion supported
by the specific increases in detection sensitivity
observed in our study; see also Small et al. (2005)
and Mohanty et al. (2008).
The interaction between cognitive and motivation networks appears to take place via several modes of communication (Figure 3D). For
instance, a specific brain region may function as
a hub linking the two types of network. Recent
advances in network theory (Guimera and Nunes
Amaral, 2005) have shown that regions characterized by a high degree of connectivity, i.e., hubs
(Sporns et al., 2007), are critical in regulating the
flow and integration of information between
regions. However, whereas the number of connections of a region is important in determining
whether it will function as a hub, the structural
topology of the region is also relevant. For instance,
some regions are best characterized as “provincial”
hubs (they occupy a central position within a single
functional cluster; e.g., visual area V4, Sporns et al.,
2007), whereas others act as “connector” hubs (they
link separate region clusters). Hubs connecting
cognitive and motivation networks would comprise examples of the latter type of region.
An intriguing suggestion by Mesulam et al. is
that the PCC provides an important site for the
integration of motivational and spatial attention
information (Small et al., 2005; Mohanty et al.,
2008; see also Platt and Huettel, 2008). In agreement with this suggestion in our neuroimaging
study, as reviewed above, the PCC exhibited both
motivation and attention signals. Specifically,
not only did the PCC exhibit cue-related, targetrelated and sustained responses that increased
with absolute incentive value, but increases in
cue-related and sustained responses were correlated with individual trait measures tied to
reward sensitivity (in this case, BAS-drive scores).
Another, not mutually exclusive, possibility is that
the ACC functions as a hub region linking the
two types of network. The ACC is known to be
important for integrating inputs from multiple
September 2010 | Volume 4 | Article 17 | 6
Pessoa and Engelmann
Embedding reward into executive function
sources, including affective and motivational
inputs (Devinsky et al., 1995; Rushworth et al.,
2007), and, in this respect, works in close cooperation with the anterior insula and OFC. The ACC
has also been suggested to be involved in several
executive processes, including conflict detection,
error likelihood processing and error monitoring,
and more generally helps determine the benefits
and costs of acting. The ACC is also important
for attentional control and controlling limitedprocessing capacity (Posner and DiGirolamo,
1998; Weissman et al., 2005; Pessoa, 2009). Thus,
the ACC is a strong candidate for a hub connecting the two types of network.
In addition to interactions at specific connector
hub regions, multiple “point-to-point” interactions may occur (indicated via the purple arrows
in Figure 3D) that provide communication pathways between valuation and cognitive regions. For
instance, in monkeys, the OFC projects to the
ventral part of Brodmann area 46 on the lateral
PFC surface (Barbas and Pandya, 1989). Another
example includes the caudate nucleus, which is
connected with several regions of frontal (including lateral PFC) and parietal cortices, in part via
the thalamus (Alexander et al., 1986).
A third type of communication involves
the diffuse action of neuromodulatory signals.
Motivationally salient items engage dopaminergic
cells in the ventral tegmental area (VTA) and SN.
Widespread modulatory connections originating in
these sites reach the entire cortical surface, thereby
having the potential to rapidly influence cortical
responses. Evidence from animal studies supports
the notion that dopaminergic modulatory effects
are associated with behavioral importance, generally (Schultz et al., 1992), and improve attentional
accuracy, specifically (Granon et al., 2000; Seamans
and Yang, 2004; Pezze et al., 2007). Several studies in
humans, including ours, have also reported rewardrelated activation in dopaminergic centers (e.g.,
Bunzeck and Duzel, 2006; D’Ardenne et al., 2008)
and, more commonly, their subcortical targets (e.g.,
caudate; both the head and body of the caudate have
been reported). It is noteworthy that dopaminergic
projections to the frontal lobe are much more significant than to posterior regions and, in particular, the occipital cortex appears to only minimally
receive such projections (Oades and Halliday, 1987).
These considerations are relevant to the understanding of the impact of motivation on both executive
and sensory processes, and suggest that the impact of
dopaminergic projection systems on visual function
is likely to be relatively minor – though a complicating factor is that the effects could be strong though
indirect. If this is correct, the effect on visual function reported in the studies above may be strongly
Frontiers in Neuroscience
www.frontiersin.org
dependent on “source” regions in frontal and parietal cortex that exert top-down modulatory signals
on sensory processing.
Although neuromodulation can be viewed as
simply another aspect of the suggested network
interactions, it is worth separating it from the others for the following reasons. Because neuromodulatory signals target superficial (I–III) and deeper
(V–VI) cortical layers, but tend to avoid layer IV
(e.g., Raghanti et al., 2008), they appear to provide
less of a “driving input” and instead may function to alter information processing. For instance,
Goldman-Rakic et al. (1989) suggest that a major
function of dopamine is to control cortical excitability, thereby possibly increasing the fidelity of
signals computed within local networks (Douglas
and Martin, 2007). More specifically, the effects of
dopamine appear to enhance the neuronal signalto-noise ratio (Sawaguchi and Matsumura, 1985),
consistent with computational modeling results
of the role of dopamine in working memory function (Gruber et al., 2006). Thus, it is intriguing
to suggest that dopaminergic neuromodulation
may be a key mechanism by which motivation
sharpens attention and behavioral performance,
for instance via the enhancement of the signalto-noise ratio of relevant neurons. Therefore, the
motivational context, which may be computed in
valuation regions, may enhance the processing
efficiency in cognitive regions via a dopaminergic signal. In the context of our task, valuation
regions (e.g., OFC) signal behavioral relevance
to neuromodulatory regions (e.g., VTA), which
then enhance neuronal processing in relevant
“cognitive” areas via dopamine signals (e.g.,
fronto-parietal regions). Future multi-site cell
recordings may be able to more directly evaluate
this working hypothesis. In any case, these specific
effects on brain function are envisaged to be quite
distinct from a simple “energizing” function.
Taken together, the available evidence suggests
that motivation and cognition interact via multiple neural substrates to guide goal-directed behavior (Figure 3D). In particular, one or more of the
above modes of communication may be operative
at a given time depending on the particulars of the
task at hand. More broadly, numerous opportunities for cognitive-emotional interactions exist in
the brain, thereby allowing motivational significance to greatly shape complex behaviors.
ACKNOWLEDGMENTS
Support for this work was provided in part by
the National Institute of Mental Health (R01
MH071589 to Luiz Pessoa). We also thank the
reviewers for many valuable insights that have
helped improve the paper.
September 2010 | Volume 4 | Article 17 | 7
Pessoa and Engelmann
REFERENCES
Alexander, G. E., DeLong, M. R., and
Strick, P. L. (1986). Parallel organization of functionally segregated circuits
linking basal ganglia and cortex. Annu.
Rev. Neurosci. 9, 357–381.
Barbas, H., and Pandya, D. N. (1989).
Architecture and intrinsic connections of the prefrontal cortex in the
rhesus monkey. J. Comp. Neurol. 286,
353–375.
Bendiksby, M. S., and Platt, M. L.
(2006). Neural correlates of reward
and attention in macaque area LIP.
Neuropsychologia 44, 2411–2420.
Berns, G. S., McClure, S. M., Pagnoni, G., and
Montague, P. R. (2001). Predictability
modulates human brain response to
reward. J. Neurosci. 21, 2793–2798.
Bunzeck, N., and Duzel, E. (2006).
Absolute coding of stimulus novelty
in the human substantia nigra/VTA.
Neuron 51, 369–379.
Corbetta, M., Kincade, J. M., Ollinger,
J. M., McAvoy, M. P., and Shulman,
G. L. (2000). Voluntary orienting is
dissociated from target detection in
human posterior parietal cortex. Nat.
Neurosci. 3, 292–297.
D’Ardenne, K., McClure, S. M., Nystrom,
L. E., and Cohen, J. D. (2008). BOLD
responses reflecting dopaminergic signals in the human ventral tegmental
area. Science 319, 1264–1267.
Delgado, M. R., Locke, H. M., Stenger,
V. A., and Fiez, J. A. (2003). Dorsal
striatum responses to reward and
punishment: effects of valence and
magnitude manipulations. Cogn.
Affect. Behav. Neurosci. 3, 27–38.
Devinsky, O., Morrell, M. J., and Vogt, B.
A. (1995). Contributions of anterior
cingulate cortex to behaviour. Brain
118(Pt 1), 279–306.
Douglas R. J., and Martin K. A. (2007).
Recurrent neuronal circuits in the neocortex. Curr. Biol. 17, R496–R500.
Engelmann, J. B., Damaraju, E., Padmala,
S., and Pessoa, L. (2009). Combined
effects of attention and motivation
on visual task performance: transient and sustained motivational
effects. Front. Hum. Neurosci. 3, 4.
doi:10.3389/neuro.09.004.2009.
Engelmann, J. B., and Pessoa, L. (2007).
Motivation sharpens exogenous spatial attention. Emotion 7, 668–674.
Goldman-Rakic, P. S., Leranth, C.,
Williams, M., Mons, N., and Gefard,
M. (1989). Dopamine synaptic complex with pyramidal neurons in primate cerebral cortex. Proc. Natl. Acad.
Sci. U.S.A. 86, 9015–9019.
Granon, S., Passetti, F., Thomas, K. L.,
Dalley, J. W., Everitt, B. J., and Robbins,
T. W. (2000). Enhanced and impaired
attentional performance after infusion
of D1 dopaminergic receptor agents
into rat prefrontal cortex. J. Neurosci.
20, 1208–1215.
Frontiers in Neuroscience
Embedding reward into executive function
Gruber, A. J., Dayan, P., Gutkin, B. S.,
and Solla, S. A. (2006). Dopamine
modulation in the basal ganglia
locks the gate to working memory. J.
Comput. Neurosci. 20, 153–166.
Guimera, R., and Nunes Amaral, L. A.
(2005). Functional cartography of
complex metabolic networks. Nature
433, 895–900.
Hopfinger, J. B., Buonocore, M. H., and
Mangung, G. R. (2000). The neural
mechanisms of top-down attentional
control. Nat. Neurosci. 3, 248–291.
Jimura, K., Locke, H. S., and Braver, T. S.
(in press). Prefrontal cortex mediation
of cognitive enhancement in rewarding motivational contexts. Proc. Natl.
Acad. Sci. U.S.A., early edition.
Kincade, J. M., Abrams, R. A., Astafiev, S.
V., Shulman, G. L., and Corbetta, M.
(2005). An event-related magnetic
resonance imaging study of voluntary and stimulus-driven orienting of
attention. J. Neurosci. 25, 4593–4604.
Knutson,B.,Taylor,J.,Kaufman,M.,Peterson,
R., and Glover, G. (2005). Distributed
neural representation of expected value.
J. Neurosci. 25, 4806–4812.
Kobayashi, S., Lauwereyns, J., Koizumi,
M., Sakagami, M., and Hikosaka, O.
(2002). Influence of reward expectation on visuospatial processing in
macaque lateral prefrontal cortex. J.
Neurophysiol. 87, 1488–1498.
Kouneiher, F., Charron, S., and Koechlin,
E. (2009). Motivation and cognitive
control in the human prefrontal cortex. Nat. Neurosci. 12, 939–945.
Leon, M. I., and Shadlen, M. N. (1999).
Effect of expected reward magnitude
on the response of neurons in the
dorsolateral prefrontal cortex of the
macaque. Neuron 24, 415–425.
Locke, H. S., and Braver, T. S. (2008).
Motivational influences on cognitive
control: behavior, brain activation,
and individual differences. Cogn.
Affect. Behav. Neurosci. 8, 99–112.
Maunsell, J. H. (2004). Neuronal representations of cognitive state: reward or attention? Trends Cogn. Sci. 8, 261–265.
Mohanty, A., Gitelman, D. R., Small, D. M.,
and Mesulam, M. M. (2008). The spatial
attention network interacts with limbic
and monoaminergic systems to modulate motivation-induced attention
shifts. Cereb. Cortex 18, 2604–2613.
Montague, P. R., Hyman, S. E., and Cohen,
J. D. (2004). Computational roles for
dopamine in behavioural control.
Nature 431, 760–767.
O’Doherty, J., Kringelbach, M. L., Rolls, E.
T., Hornak, J., and Andrews, C. (2001).
Abstract reward and punishment representations in the human orbitofrontal cortex. Nat. Neurosci. 4, 95–102.
Oades, R. D., and Halliday, G. M.
(1987). Ventral tegmental (A10)
system: neurobiology. 1. Anatomy and
connectivity. Brain Res. 434, 117–165.
Pantoja, J., Ribeiro, S., Wiest, M., Soares, E.,
Gervasoni,D.,Lemos,N.A.,and Nicolelis,
M. A. (2007). Neuronal activity in the
primary somatosensory thalamocortical
loop is modulated by reward contingency during tactile discrimination. J.
Neurosci. 27, 10608–10620.
Pessoa, L. (2009). How do emotion and motivation direct executive control? Trends
Cogn. Sci. (Regul. Ed.) 13, 160–166.
Pezze, M. A., Dalley, J. W., and Robbins, T. W.
(2007). Differential roles of dopamine
D1 and D2 receptors in the nucleus
accumbens in attentional performance
on the five-choice serial reaction time task.
Neuropsychopharmacology 32, 273–283.
Platt, M. L., and Glimcher, P. W. (1999).
Neural correlates of decision variables in
parietal cortex. Nature 400, 233–238.
Platt, M. L., and Huettel, S. A. (2008).
Risky business: the neuroeconomics
of decision making under uncertainty.
Nat. Neurosci. 11, 398–403.
Pleger, B., Blankenburg, F., Ruff, C. C.,
Driver, J., and Dolan, R. J. (2008).
Reward facilitates tactile judgments and
modulates hemodynamic responses in
human primary somatosensory cortex.
J. Neurosci. 28, 8161–8168.
Pleger, B., Ruff, C. C., Blankenburg, F.,
Kloppel, S., Driver, J., and Dolan, R. J.
(2009). Influence of dopaminergically
mediated reward on somatosensory
decision-making. PLoS Biol. 7, e1000164.
doi: 10.1371/journal.pbio.1000164.
Pochon, J. B., Levy, R., Fossati, P., Lehericy,
S., Poline, J. B., Pillon, B., Le Bihan, D.,
and Dubois, B. (2002). The neural system that bridges reward and cognition
in humans: an fMRI study. Proc. Natl.
Acad. Sci. U.S.A. 99, 5669–5674.
Posner, M. I., and DiGirolamo, G. J.
(1998). “Executive attention: conflict, target detection, and cognitive
control,” in The Attentive Brain, ed. R.
Parasuraman (Cambridge, MA: MIT
Press), 401–423.
Raghanti, M. A., Stimpson, C. D.,
Marcinkiewicz, J. L., Erwin, J. M.,
Hof, P. R., and Sherwood, C. C. (2008).
Cortical dopaminergic innervation
among humans, chimpanzees and
macaque monkeys: a comparative
study. Neuroscience 155, 203–220.
Rushworth, M. F., Buckley, M. J., Behrens,
T. E., Walton, M. E., and Bannerman,
D. M. (2007). Functional organization
of the medial frontal cortex. Curr.
Opin. Neurobiol. 17, 220–227.
Sawaguchi, T., and Matsumura, M. (1985).
Laminar distribution of neurons sensitive to acetylcholine, noradrenaline
and dopamine in the dorsolateral
prefrontal cortex of monkey. Neurosci.
Res. 2, 255–273.
Schultz, W., Apicella, P., Scarnati, E., and
Ljungberg, T. (1992). Neuronal activity
in monkey ventral striatum related to
the expectation of reward. J. Neurosci.
12, 4595–4610.
www.frontiersin.org
Seamans, J. K., and Yang, C. R. (2004). The
principal features and mechanisms of
dopamine modulation in the prefrontal cortex. Prog. Neurobiol. 74, 1–58.
Serences, J. T. (2008). Value-based modulations in human visual cortex. Neuron
60, 1169–1181.
Shuler, M. G., and Bear, M. F. (2006).
Reward timing in the primary visual
cortex. Science 311, 1606–1609.
Small, D. M., Gitelman, D., Simmons, K.,
Bloise, S. M., Parrish, T., and Mesulam,
M. M. (2005). Monetary incentives
enhance processing in brain regions
mediating top-down control of attention. Cereb. Cortex 15, 1855–1865.
Sporns, O., Honey, C. J., and Kotter, R.
(2007). Identification and classification of hubs in brain networks. PLoS
ONE 2, e1049. doi:10.1371/journal.
pone.0001049.
Taylor, S. F., Welsh, R. C., Wager, T. D., Phan,
K. L., Fitzgerald, K. D., and Gehring,
W. J. (2004). A functional neuroimaging study of motivation and executive
function. Neuroimage 21, 1045–1054.
Visscher, K. M., Miezin, F. M., Kelly, J.
E., Buckner, R. L., Donaldson, D. I.,
McAvoy, M. P., Bhalodia, V. M., and
Petersen, S. E. (2003). Mixed blocked/
event-related designs separate transient and sustained activity in fMRI.
Neuroimage 19, 1694–1708.
Watanabe, M., and Sakagami, M. (2007).
Integration of cognitive and motivational context information in the primate prefrontal cortex. Cereb. Cortex
17(Suppl. 1), i101–i109.
Weissman, D. H., Gopalakrishnan, A.,
Hazlett, C. J., and Woldorff, M. G.
(2005). Dorsal anterior cingulate cortex
resolves conflict from distracting stimuli
by boosting attention toward relevant
events. Cereb. Cortex 15, 229–237.
Zink, C. F., Pagnoni, G., Chappelow, J.,
Martin-Skurski, M., and Berns, G.
S. (2006). Human striatal activation
reflects degree of stimulus saliency.
Neuroimage 29, 977–983.
Conflict of Interest Statement: The
authors declare that the research was conducted in the absence of any commercial or
financial relationships that could be construed as a potential conflict of interest.
Received: 01 December 2009; paper pending
published: 24 January 2010; accepted: 10 March
2010; published online: 15 September 2010.
Citation: Pessoa L and Engelmann JB
(2010) Embedding reward signals into
perception and cognition. Front. Neurosci.
4:17. doi: 10.3389/fnins.2010.00017
Copyright © 2010 Pessoa and Engelmann.This
is an open-access article subject to an exclusive
license agreement between the authors and the
Frontiers Research Foundation, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original
authors and source are credited.
September 2010 | Volume 4 | Article 17 | 8