Review

Dopamine and adaptive memory

Daphna Shohamy1* and R. Alison Adcock2*
1
Department of Psychology, Columbia University, New York, NY 10025, USA
2
Center for Cognitive Neuroscience, B253 Levine Science Research Center, Duke University, Box 90999, Durham, NC 27708, USA

Memory is essential to adaptive behavior because it In humans, a common framework for understanding the
allows past experience to guide choices. Emerging find- role of the hippocampus in memory formation originated
ings indicate that the neurotransmitter dopamine, which with neuropsychological studies in patients ([5]; for a
signals motivationally important events, also modulates review see [6]). This research demonstrated that the hip-
the hippocampus, a crucial brain system for long-term pocampus supports a specialized system for creating epi-
memory. Here we review recent evidence that highlights sodic memories of everyday events, and that this episodic
multiple mechanisms whereby dopamine biases memo- system is distinct and dissociable from brain systems
ry towards events that are of motivational significance. responsible for other kinds of memory (e.g. emotional
These effects take place over a variety of timescales, memory, habit learning, etc. [6]).
permitting both expectations and outcomes to influence A key function of memory, however, is presumably to
memory. Thus, dopamine ensures that memories are improve choices and actions. Indeed, research in animals
relevant and accessible for future adaptive behavior, a has always necessarily investigated memory in the service
concept we refer to as adaptive memory. Understand- of rewards and goal-directed behaviors: prototypical para-
ing adaptive memory at biological and psychological digms for testing hippocampal memories in animals are
levels helps to resolve a fundamental challenge in mem- remembering where in a maze or underneath which object
ory research: explaining what is remembered, and why. a food reward can be found. In such situations, neurons in
the hippocampus respond to the received reward, and not
Introduction only to the location or object that predict it [7]. Further,
Memory is essential to behavior, enabling organisms to while rats navigate a maze in search of rewards, anticipa-
draw on past experience to improve choices and actions. tory hippocampal responses occur at key decision points
Much research has focused on how the hippocampus builds and before goal-directed movements [8].
accurate memory for past events. Emerging findings indi- Recent work in humans has similarly begun to demon-
cate that the neurotransmitter dopamine, known to play a strate relationships between episodic memory and future
key role in motivated behavior, has a direct impact on goal-directed behavior. First, in addition to its role in
memory formation in the hippocampus. Here, we review remembering the past, the MTL also supports the ability
this emerging literature that demonstrates that interac- to imagine specific episodes in the future [9,10], with direct
tions between midbrain dopamine regions and the hippo- implications for decision making [11]. Second, because of
campus promote memory for episodes that are rewarding their relational structure, episodic memories are flexible:
and novel and build memory representations well-suited to they are constructed in a manner that allows relevant
guide later choices. By integrating findings from both hu- elements of a past event to be brought to bear as needed
man and animal research, we argue for a framework in to guide future behavior [1,12].
which dopamine helps create enriched mnemonic represen- Together, these findings emphasize a role for the hip-
tations of the environment to support adaptive behavior. pocampus that extends beyond memory for objects and
locations, to include motivated, goal-directed behavior.
The hippocampus: Creating building blocks for Below we describe evidence demonstrating that episodic
memory-guided behavior memories are modulated by the potential relevance of
After decades of research, our understanding of the brain events to later behavior and by the motivational state of
mechanisms that contribute to long-term memory for the organism, as well as evidence that the neurotransmit-
events or episodes often referred to as episodic memory ter dopamine plays a key role in this process.
has evolved significantly. Episodic memories are formed
rapidly (after even a single experience) and are rich in Brain systems for learning and motivation
contextual details. Episodic memories are also thought to Converging evidence indicates that the release of dopa-
be relational: they encode relationships between multiple mine signals motivationally important events and beha-
elements of an event [1,2]. Extensive converging evidence viors. Key findings come from a series of seminal
indicates that episodic memory depends crucially on the neurophysiology studies of dopamine-containing midbrain
hippocampus and surrounding medial temporal lobe neurons in primates receiving reward (for a review see
(MTL) cortices [1,3,4]. [13]). In these studies, a monkey receives a reward (e.g.
juice), which is predicted by a cue (e.g. a tone). Dopamine
Corresponding authors: Shohamy, D. (shohamy@psych.columbia.edu); neurons respond with a burst of activity often referred to
Adcock, R.A. (alison.adcock@duke.edu)
*
This article represents a collaborative effort based on equal contributions from
as a phasic response when the monkey unexpectedly
both authors; the listing order was determined randomly. receives a reward. Crucially, however, the response is
464 1364-6613/$ see front matter 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.tics.2010.08.002 Trends in Cognitive Sciences, October 2010, Vol. 14, No. 10
Review Trends in Cognitive Sciences Vol.14 No.10

Box 1. Integrating methods to determine the role of neuronal mechanisms underlying these processes could
dopamine in memory be related to sustained changes in dopamine release
referred to as tonic responses and not to the temporal-
FMRI studies linking midbrain or striatal activation and memory in
humans have been interpreted as suggesting an important role for ly-specific phasic bursts [21].
dopamine in memory. However, it is important to note that with A comprehensive characterization of the role of dopa-
standard imaging parameters caution is warranted when interpret- mine in behavior is still the subject of active study. In
ing BOLD signals from brainstem nuclei [51,66]. Furthermore, fMRI particular, there is continued debate about whether phasic
measures changes in BOLD across states and does not directly
measure changes in dopamine levels (for review see [67] and [68]).
responses do indeed reflect reward prediction errors
A current challenge among researchers is thus to gain direct [19,23]. What is clear, however, is that dopamine neurons
information about how changes in dopamine levels affect memory provide multiple mechanisms for signaling the occurrence
in humans. Several approaches to addressing this question have and expectation of events that are of motivational signifi-
been developed and applied to other cognitive domains, suggesting cance, and for sending these signals to a selective set of
their suitability for advancing knowledge regarding the direct role of
dopamine in long-term episodic memory. These approaches
target regions to coordinate motivation to learn about, and
include: ultimately obtain, goals. Thus, dopaminergic signals pro-
 Genetics. Individual variability in genes affecting dopamine vide a potential mechanism for making the contents of
transmission permits correlations between genetically-deter- memory motivationally relevant. A key question is wheth-
mined dopamine availability and behavioral and neural processes
er, and how, this happens.
linked to memory (e.g. [42,69]).
 Dopamine deficiency in patients. In populations who are known to
have dopamine depletion, such as Parkinsons disease (and to a Dopamine modulates hippocampal memories
lesser extent normal aging), studying memory on and off Much of the evidence for the role of dopamine in modulat-
dopaminergic medications reveals robust effects on specific ing hippocampal function comes from anatomical and
learning processes (e.g. [34,70]).
electrophysiological studies in animals. Midbrain dopa-
 PET imaging. PET has been used to relate dopamine receptor
density to cognitive function [45]. In addition, the displacement of mine neurons project directly to the hippocampus and to
radioactive ligands from dopamine receptors following a manip- the surrounding MTL cortices [24,25] (Figure 2; see Table I
ulation can be used as an index of endogenous dopamine release in Box 2). Indeed, in animals, dopamine seems to be
[71], allowing researchers to validate the use of midbrain BOLD essential for hippocampal long-term memory. Studies in
activation as a proxy [37].
 Pharmacological manipulation. In healthy individuals and pa-
animals indicate that dopamine acting at hippocampal
tients, drugs are available that increase dopamine receptor synapses is a necessary precursor not only for long-term
activation or pools of available dopamine, or counter these effects potentiation (LTP), a prime cellular model of learning and
(e.g. [72]). memory [2628], but also for the behavioral persistence of
 Pharmacological fMRI. Dopamine agonists and antagonists have long-term memories [2931]. Importantly for models of
also been used to examine changes to BOLD activation during
cognitive processes (for review see [68]).
episodic memory, dopamine-dependent facilitation of neu-
ral plasticity is evident after even a single event [32].
Finally, dopamine release in the hippocampus is itself
not simply a report of reward: when reward is entirely modulated by hippocampal activity: outputs from the hip-
expected based on prior experience, the neurons respond pocampus facilitate dopaminergic signaling in the mid-
not to the reward but to the predictive cue instead. Fur- brain, which in turn can enhance hippocampal plasticity
thermore, when reward is expected but fails to arrive, the via dopamine release (for a review see [33]).
neurons are briefly inhibited below their baseline response These findings raise questions about the behavioral
rates. Thus, the phasic responses of dopamine neurons contexts in which dopamine would modulate memory for-
seem to report the difference between observed and mation in the hippocampus, and the implications for epi-
expected reward a so-called reward prediction error sodic memory in humans. Until recently, the role of
(for a review see [13] and Figure 1). Computational models dopamine in episodic memory in humans has been rela-
have emphasized the importance of such reward prediction tively understudied. If anything, studies with humans with
errors in driving learning, and in the past several years, specific impairments of dopamine transmission, such as in
functional magnetic resonance imaging (fMRI) has Parkinsons disease, indicate that episodic memory is in-
revealed similar findings in humans engaged in a variety tact and that only incremental, feedback-driven, learning
of reward-related behaviors (for a review see [14]). Togeth- is impaired ([34,35]; for a review see [36]). However, Par-
er, these studies demonstrate that midbrain dopamine kinsons disease involves relatively selective depletion of
neurons and their striatal targets play a central role both dopamine in the dorsal striatum and thus does not provide
in responding to rewards and in learning to predict them. a good model for understanding dopamine in the hippo-
The role of dopamine in motivated behavior, however, campus.
goes beyond putative prediction errors. Recent evidence As reviewed below, the accumulation of new evidence
indicates that dopamine neurons respond not only to re- from human research indicates that midbrain dopamine
ward and its expectation, but also to novel and surprising regions do modulate episodic memory, that this happens
events, including punishment ([15]; for parallels in human via interactions with the hippocampus, and that this mod-
neuroimaging see [16] and [17]). Further, dopamine has ulation occurs under specific behavioral contexts. Much of
long been known to relate to the amount of effort exerted this evidence comes from fMRI, demonstrating blood-oxy-
towards obtaining rewards (also related to wanting and to genation-level-dependent (BOLD) activation in midbrain
incentive salience) and is implicated broadly in behavioral regions that contain dopamine neurons. Importantly, this
vigor [1822]. Notably, it has been suggested that the indirect evidence from neuroimaging is complemented by

465
(Figure_1)TD$IG][ Review [(Figure_2)TD$IG] Trends in Cognitive Sciences Vol.14 No.10

Figure 2. Inputs and outputs of the midbrain ventral tegmental area (VTA) and the
hippocampus. For clarity, only some of the projections are shown. A loop between
the hippocampus and VTA consists of direct projections from VTA to the
hippocampus, and connections from hippocampus through nucleus
accumbens (NAcc) and globus pallidus (GP) back to VTA [33]. At least one
additional route is possible, via relays in the prefrontal cortext (PFC). Midbrain
dopamine neurons in the VTA also innervate other select brain regions, all
implicated in different forms of memory, including PFC, NAcc and the amygdala
(Amg). This selective topography is notable: dopamine neurons, unlike those in
other neuromodulatory systems such as acetylcholine or norepinephrine,
innervate a select set of brain regions [96], sometimes characterized as
convergence zones [97]. Inputs to the VTA modulate dopamine neuronal
responses. Excitatory activity (green) in the hippocampus disinhibits dopamine
neurons by inhibiting (red) GP. Other relevant inputs originate from subcortical
sensory areas (e.g. superior colliculus; SC) and PFC (directly and via the
pedunculopontine tegmentum, PPTg). Additional nuclei that are the focus of
recent research include inhibitory influences that might signal aversive stimuli
Figure 1. Schematic representation of responses in midbrain dopamine neurons. (e.g., the rostromedial tegmental nucleus (RMTg), mediator of inhibitory inputs
Dopamine neurons have two characteristic response modes: They normally fire in from lateral habenula (Hb). Putative contributions of these afferents to tonic and
a tonic pattern (approx. 5 Hz) and periodically fire with short, phasic bursts phasic dopamine are reviewed in [23,33,98].
(approx. 20 Hz). It has been suggested that these response patterns could relate to
different behavioral contexts. Here, dopamine neurons respond to rewards that are
probabilistically predicted by visual cues (based on [95]). A phasic dopamine
response is elicited when an animal receives an unexpected reward (P=0.01, after
the diamond cue) or a cue that always predicts reward (P=1.0, after the triangle
cue). When a predicted reward fails to appear, the dopamine neuronal response is
polymorphisms in the dopamine transporter (DAT1) gene
briefly inhibited below baseline. When a reward is predicted by a cue part of the affect BOLD activity in the midbrain during episodic
time (P=0.5, after the circle cue) that is, when there is uncertainty about the encoding of novel stimuli [42]. These results in humans
upcoming reward the dopaminergic phasic response seems to trade-off between
the cue and the reward. Importantly, there is also a slow and sustained ramping up
are consistent with models that suggest a key role for a
of activity. This signal is well-suited to provide sustained modulation of target network between the hippocampus and the ventral teg-
regions and could reflect tonic dopamine. Indeed, other evidence has led to the mental area (VTA) in enhancing long-term memory for
suggestion that tonic dopamine responses provide candidate signals of
expectancy or motivation [21].
novel events [33].
Interestingly, work in animals and humans indicates
that, in addition to enhanced memory for novel items,
other methods that provide more direct evidence about novelty can also lead to prolonged effects by enhancing
dopamine per se (Box 1). In particular, one key finding memory for items that take place in a novel context, an
demonstrates that BOLD responses in the midbrain dur- effect mediated by midbrain dopamine regions [43,44]. In
ing reward anticipation correlate with dopamine release in humans, there is evidence to indicate that dopamine is
the striatum, ([37]; Box 1). related to novelty-seeking behaviors and that individual
Recent findings suggest the following general princi- variability in dopamine receptor density in the midbrain is
ples: related to novelty-seeking traits [45].

Novelty engages midbrain modulation of the Reward anticipation drives interactions between
hippocampus midbrain dopamine regions and the hippocampus
It has long been recognized that novelty modulates episod- Much recent evidence supports the view that processing of
ic memory. FMRI studies demonstrate greater hippocam- novelty and reward are tightly interrelated [33,4649]. The
pal activation during encoding of novel relative to familiar role of reward in driving responses in midbrain dopamine
stimuli (e.g. [38,39]), and activation in midbrain dopamine regions has been widely documented across species (e.g.
regions is also greater in response to novel events than to [13,50,51]), indicating that reward might also modulate
familiar ones [40]. A recent study used intracranial EEG interactions between midbrain dopamine regions and the
recordings from the hippocampus and the nucleus accum- hippocampus.
bens (a primary target of midbrain dopamine neurons) in Indeed, recent fMRI studies in humans demonstrate
humans to provide more direct evidence of the role of that reward modulates activation in the hippocampus and
novelty in eliciting neural responses [41]. The presentation the midbrain in multiple ways [5254]. First, activation in
of novel pictures, but not familiar ones, led to enhanced the midbrain following reward cues has been related to
EEG responses in both the hippocampus and the nucleus episodic memory for those cues (Figure 3a, [52,53]). These
accumbens. Finally, genetic imaging demonstrates that fMRI studies importantly demonstrate a link between

466
Review Trends in Cognitive Sciences Vol.14 No.10

Box 2. Convergent systems for the modulation of memory

Midbrain dopamine neurons innervate a relatively select topography tional parallels between them. As shown in Table I, levels of binding
of brain regions implicated in different forms of memory (see Figure 1 for terminals (dopamine transporter; DAT below) and receptors (D5,
in main text). This indicates that beyond the direct effects of D1-like and D2-like) in the hippocampus seem to be more similar to
dopamine on the hippocampus, interactions among a wider network prefrontal cortex than to striatum. Beyond these parallels, it is
of brain regions provide additional mechanisms by which motivation unknown whether dopamine tunes active representations in the
can affect learning and memory. hippocampus in a manner analogous to its enhancement of prefrontal
Dopaminergic projections to the striatum and to the amygdala have working-memory representations.
been strongly implicated in learning and both have been demon- Notably, in the hippocampus, D5 is the main dopamine receptor
strated to interact with the hippocampus either competitively or type. As in frontal cortex, D5 receptors in the hippocampus are
cooperatively [41,7378]. Prefrontal cortex has long been known to be located primarily on dendritic shafts and are mainly extrasynaptic. By
modulated by dopamine [7981] and could support goal representa- contrast, D5 receptors in the striatum are mainly localized to spines
tion during reward-motivated learning [54,82]. and therefore more likely to be synaptic [86]. Phasic responses have
The effects of dopamine release on multiple targets indicate that as been argued to affect synaptic but not extrasynaptic dopamine levels,
dopamine flows into the hippocampus, it also modulates other whereas increases in tonic dopamine preferentially elevate extra-
specialized systems to render goal-related and behaviorally relevant synaptic levels [87]. In prefrontal cortex, D5 receptors are in fact
events in memory. In addition, dopaminergic neuromodulation of the closely associated with well-defined extrasynaptic microdomains
hippocampus undoubtedly co-occurs and interacts with modulation specialized for volume transmission [88].
by norepinephrine and acetylcholine, each of which have been In the hippocampus, comparison of dopamine receptor distribu-
proposed to signal salient events or uncertainty [83], and to influence tion across species highlights several important points. First, the
long-term plasticity in the hippocampus [84,85]. The consequences of pattern of receptor distributions across the hippocampus varies
interactions between these neuromodulatory systems on memory are dramatically across species. These qualitative interspecies differ-
largely unknown and are an important question for future research. ences dictate caution in drawing parallels between findings in
An additional important question is how the effects of dopamine in rodents and humans. Second, particularly in primates, the localiza-
the hippocampus compare with its effects on other regions. One way tion of terminals within the hippocampus is remote from the densest
to begin addressing this question is by comparing the distribution of receptor distributions. This mismatch in distribution, similar to the
dopamine receptors across these regions. extrasynaptic localization of D5 receptors mentioned above, indi-
Similarities between prefrontal and hippocampal dopamine inner- cates that hippocampal function especially in primates relies
vation and ultrastructural receptor distributions indicate some func- heavily on tonic dopamine.

Table I. Regional distribution of dopamine terminals and receptors across species.

Species Labeled Structure Brain Region References
Hippocampus Striatum Frontal Cortex
CA1/2 CA3 DG CPu NAcc Layers I-III
Human DAT (nCi/mg) +a + ++ ++++++++++++++ ++++++++++++ NR [89]
D5 (anti-D5 Ab) * * * * * * [90]
D1-like (fmol/mg) +++ + + +++++++++ +++++++++ +++ [91]
D2-like (fmol/mg) ++ + + +++++ ++++++ j [91]
Monkey DAT (anti-DAT Ab) - - +++ NR +++++ [92]
D5 (anti-D5 Ab) ++ ++ +++ ++ ++ [86,93]
D1-like (fmol/mg) ++ + j ++++++++++++ +++++++++++ +++++ [91]
D2-like (fmol/mg) j + + +++++++++ ++++++++ j [91]
Rat DAT (fmol/mg) + + ++ +++ +++ NR [94]
D5 (anti-D5 Ab) ++ +++ + ++ ++ [90,93]
D1-like (fmol/mg) + + +++ ++++++++++++++++++ ++++++++++++++++++ ++ [91]
D2-like (fmol/mg) + + ++ +++++++++ +++++ + [91]
a
Symbols indicate relative density across regions: + significant labeling; j minimal labeling; * present (unquantified); NR = not reported. Differences across regions (DG,
dentate gyrus; CPu, caudate/putamen; NAcc, nucleus accumbens) indicate that hippocampal dopamine receptor densities are more similar to prefrontal than striatal
levels. The distributions also indicate that dopamine terminals (shaded boxes) are remote from the densest receptor distributions, which could have important
implications for understanding mechanism, particularly in primates (see text). Methodological differences preclude direct comparison of quantitative data (D5 and DAT via
immunocytochemistry; D1-like and D2-like and DAT via autoradiography). Quantitative D5 antibody data are unavailable for humans; quantitative mRNA labeling
indicates that the D1-like receptors in hippocampus are mainly D5.

reward, midbrain activation and episodic memory per- motivated encoding found that information about reward
formance. value might be directly embedded in episodic memories
Second, midbrain responses are not limited to reward [55].
cues only, but have also been demonstrated when potential Thus, both reward cues in the present and motivation to
reward is used to motivate memory encoding ([54]; obtain rewards in the future enhance activation in the
Figure 3b). This fMRI study revealed that reward-related midbrain and episodic memory. Together, these findings
motivation was associated with coupled activation in the indicate that episodic memory is not an arbitrary record of
midbrain and in the hippocampus, that this activation was events, but could be biased to preserve reward-related
elicited before the presentation of items, and that this information.
anticipatory activation predicted later episodic memory.
Because no rewards occurred during learning in this study, Midbrain-hippocampal interactions support the
these findings indicate a broad conceptualization of the integration of memory across experiences
effects of reward on learning to include motivation to The findings reviewed above have important implications
obtain rewards to be gained in the future. Interestingly, not only for which memories are preserved, but also for how
a recent study using a similar manipulation of reward- these memories are represented. Memories modulated by

467
(Figure_3)TD$IG][ Review Trends in Cognitive Sciences Vol.14 No.10

Figure 3. Behavioral paradigms that involve reward and novelty elicit activation in the midbrain and the hippocampus that relates to memory function in humans. (a)
Reward-related encoding (based on [52]): To examine whether an experience that elicited dopamine release was better remembered, this fMRI study used pictures of
objects whose category (living or non-living) indicated reward for successful performance on the next trial of a number judgment task. Based on an expected reward
outcome, items in the category that indicated a reward trial would be expected to elicit both phasic and tonic dopamine responses. (Note that every item was novel, and
thus might also elicit some increase in tonic dopamine according to [33].) Findings revealed that both midbrain and the MTL were activated following the presentation of
items from the reward-predicting category relative to the neutral category. Furthermore, reward-predicting items were better remembered and better associated with their
encoding context in source memory. (b) Motivated memory (based on [54]): To test the effects of reward anticipation on memory, this fMRI study presented participants
with a series of novel pictures to be memorized for a test the next day. Crucially, a few seconds before each picture, participants saw an alerting cue telling them that
remembering the picture later would earn them a large monetary reward versus a negligible one. This design allowed the researchers to determine whether anticipatory
brain activity putatively related to motivation that is before the to-be-learned item was even presented modulates later memories. Findings revealed that the cues for
large rewards elicited greater activation of both the midbrain and the hippocampus, increased their correlation, and predicted better memory on the test. (c) Integrative
encoding (based on [12]): In this learning and generalization paradigm people use feedback to learn a series of associations between faces and scenes. Associations are
learned individually but have overlap between them. For example people learn that Bethany prefers cityscapes to fields, and deserts to beaches. They also learn that Walter
prefers deserts to beaches. At a later probe phase, people are able to generalize this knowledge to answer a new question: Does Walter prefer cityscapes or fields? Findings
from this study revealed that activation in the hippocampus and in the midbrain during the initial learning phase correlated with later generalization. (d) Overlay of
midbrain and hippocampal activations related to adaptive memory in studies illustrated in (a) (red) (b) (blue) and (c) (yellow). These common patterns of activation raise
questions about the common processes and putative mechanisms elicited in these studies. Across studies, joint midbrain and hippocampal activation is related to the
construction of memories that are well-suited to adaptively guide later behavior. In addition, learning in all of these paradigms requires the integration of information
across elements that do not co-occur. This alludes to the need for mechanisms that allow dopamine to help integrate information across different time points, as discussed
in detail in the main text.

reward and novelty have all the hallmark features of Whether and how dopamine would facilitate integrative
episodic memory: they are rich in contextual details (such encoding is unknown. Evidence from animals indicates
as the source of the memory [52] and the potential value of that dopamine responses to novel elements in familiar
the item [55]) and are experienced with high confidence settings might aid the incorporation of new experience
[54]. into associative networks of information, resulting in gen-
Another key feature of episodic memory that seems to be eralizable knowledge or schemas [12,31,55]. One possibil-
modulated by interactions between midbrain and hippo- ity is that partial overlap across experiences generates
campus is representational flexibility [12]. This type of predictions and the hippocampus signals violation of these
flexibility is essential for generalization of knowledge, predictions. These mismatch signals could elicit midbrain
and has long been known to depend on the hippocampus dopamine responses, similar to the responses to novelty
(e.g. [56,57]). Recent evidence from humans indicates that and reward cues reviewed earlier.
hippocampal-midbrain interactions might facilitate gener- In summary, recent studies highlight a role for inter-
alization of knowledge by promoting integration of discrete actions between midbrain dopamine regions and the hip-
episodes [12]. In this fMRI study, participants engaged in pocampus in several key functions, including detecting and
feedback-driven associative learning while being scanned recording novel and behaviorally relevant events, enhanc-
(Figure 3c). Those who showed robust activation of the ing representation of events experienced during reward
midbrain together with the hippocampus during learning anticipation, and integrating associations across memo-
were more likely to later generalize that knowledge to ries. These functions should ensure that the content of
correctly and rapidly solve a never-before-seen problem. memory is relevant and that it can be deployed flexibly to
Thus, when individuals encounter new information that guide future behavior.
evokes old associations, this associative novelty can pro-
mote not only preferential encoding but also the organiza- Dopamine modulates hippocampal memories over a
tion of memory to facilitate its later use processes that range of timescales
could be catalyzed by dopamine [29,31,58]. This integra- Interestingly, the role of dopamine in supporting the for-
tive encoding could thus enable online formation of links mation of memories extends over a range of timescales:
between discrete memories, facilitating the use of memo- before, during, and after an event. As discussed in the next
ries to guide behavior in new situations. section, this broad range of timescales implies a similarly

468
Review [(Figure_4)TD$IG] Trends in Cognitive Sciences Vol.14 No.10

broad range of neurobiological mechanisms by which mid-

brain dopamine neurons can exert an effect on memory
formation in the hippocampus.
Initial evidence highlighted the time period before an
event as a window during which the presence of dopamine
is crucial. Dopamine enhancement of LTP in the hippo-
campus was obtained in vitro when dopamine was already
available at the synapse when neurons were stimulated,
implying that dopamine release before an event would
Figure 4. Hypothesized temporal characteristics of memory modulation by
enhance memory for that event [27,59]. Exploration of a dopamine. Midbrain dopamine neurons change their activity both in response to
novel environment before electrical stimulation has also salient transient events such as novelty and reward and during sustained contexts
been shown to enhance LTP and this effect depends on such as anticipation, motivation or expectancy. Both modes are hypothesized to
promote memory formation, but might involve different mechanisms. For
dopamine [43]. In humans, cues indicating potential later dopamine to facilitate memory for an event that elicited and thus preceded a
reward for remembering an upcoming event enhance mid- dopamine response (for example the sound of a bell, analogous to [52]) a
brain-hippocampal interactions before the event occurs mechanism is required that is retroactive in time. One proposed mechanism
targets recently active synapses [99]. Enhanced memory for events that occur
[54]. Similarly, exposure to novel images can proactively during sustained dopamine activation, hypothesized to correspond to motivation
enhance memory formation for familiar images presented or expectancy (analogous to [12,54]), is consistent with a generalized decrease in
later [44]. thresholds for lasting plasticity. Such a generalized mechanism could also occur
following phasic dopamine responses to salient events; however, phasic
There is also evidence for retroactive effects of midbrain dopamine responses could be more spatially restricted to synapses and
dopamine regions on memory. In humans, the presentation therefore show more selectivity. An additional window of dopamine contribution
to memory longevity, not shown here, has been described hours after the encoded
of behaviorally relevant stimuli (novel or rewarding) elicits
events [30].
activation in midbrain dopamine regions, with better long-
term memory for those events [40,52]. Importantly, in these
studies, midbrain activation was presumably stimulated by dopamine neurons, thus increasing the number of tonically
the events themselves, rather than by events preceding active cells, but does not produce phasic responses in
them. individual neurons [63]. As detailed below, this key finding
Finally, dopamine can have effects on hippocampal could form the basis of two putative mechanisms.
plasticity and memory on a longer timescale after events
are experienced, during a phase often referred to as con- One possibility is that tonic responses increase
solidation. Recent findings in animals demonstrate that hippocampal dopamine indirectly via facilitation of
dopamine supports the persistence of associative memory phasic bursts
over timescales as long as 24 hours [31]. Investigations of It has been proposed that enhanced tonic dopamine reflects
candidate mechanisms of active consolidation for exam- an increased number of neurons that are disinhibited by
ple, hippocampal replay during rest after learning also hippocampal activity. Disinhibition increases spontaneous
show modulation by reward [60,61]. Finally, a recent study activity but also makes it easier for an individual neuron to
of dopaminergic effects on avoidance learning indicates a burst [63]. Enhanced phasic responses in the midbrain
second crucial window for dopamine availability that could then impact dopamine release and memory encoding
occurs 12 hours after learning and implies not only pro- in the hippocampus [33,47]. This mechanism is clearly
tracted effects of an earlier release, but also a role for the consistent with observed effects of dopamine on encoding
presence of dopamine during future reactivation and active of individual cues. It could also possibly explain observed
consolidation [30]. anticipatory and sustained effects of dopamine on encod-
Collectively, then, evidence suggests that the range of ing: sustained contexts (such as exploration of novel envir-
timescales over which dopamine is important to durable onments or reward anticipation), by increasing tonic
memory formation by the hippocampus is broad, as illus- responses, would increase the likelihood that dopamine
trated in Figure 4: it begins before experience and con- neurons would burst in response to a specific event within
tinues into a temporal window of hours or days. that context.
Importantly, the richness of these relationships indicates With the emphasis on a functional role for phasic bursts,
that although dopamine could indeed be a biological teach- this view makes several predictions. First, it predicts that
ing signal, to be delivered at the biologically right moment memory encoding should be enhanced specifically for indi-
(Nobel Laureate Konrad Lorenz, in [62]), the right time for vidual time-constrained events (e.g. the appearance of a
teaching signals in the hippocampal memory system seems single stimulus) rather than for the sustained episodes or
broadly drawn. contexts within which they occur. Second, it predicts that
selective disruption of either phasic responding or tonic
Putative neurobiological mechanisms: Tonic vs. phasic responding should abolish the facilitatory effects of dopa-
dopamine mine on hippocampal encoding. Notably, however, this
The precise mechanisms whereby dopamine modulates prediction is somewhat inconsistent with results from a
memory formation in the hippocampus are not yet known. recent study in rodents [64]. This study used a novel method
Nevertheless, existing findings point to a potentially im- to selectively disable phasic responses, while having no
portant role for tonic responses in dopamine neurons. In impact on tonic responses. Disabling phasic responses led
particular, animal research indicates that hippocampal to marked and selective disruption of reward learning and
activity originating in the subiculum disinhibits midbrain instrumental conditioning, whereas long-term spatial mem-

469
Review Trends in Cognitive Sciences Vol.14 No.10

ory a hallmark of hippocampal function remained intact accounted for by a single mechanism (Figure 4). Further,
(as did motivation and working memory). multiple other brain systems and neurotransmitters are
probably involved (Box 2).
Another possibility is that tonic dopamine directly A final important question to be resolved is how dopa-
modulates hippocampal encoding mine released at or around the time of encoding affects the
An increase in tonic dopamine activity from increased life of the memory trace long after it has been laid down.
hippocampal input to VTA could, in and of itself, lead Generalized enhancement of memory for events encoun-
directly to changes in tonic dopamine release in the hippo- tered following tonic (and perhaps phasic) dopaminergic
campus. This mechanism would provide a parsimonious activation could be explained via volumetric effects of
account of the observed effects of anticipatory and sus- dopamine release that lower thresholds for LTP in all
tained dopamine on encoding. This mechanism could also active synapses [2628,43]. Explaining how dopamine
possibly explain observed effects of dopamine on encoding could enhance the encoding of items that elicit phasic
of individual cues: enhanced tonic dopamine could facili- release, however, requires a retroactive mechanism
tate encoding of the context, the events that take place (Figure 4). One proposed mechanism is that of synaptic
within that context, and the relation between them. tagging: a process whereby excitation in the hippocampus
In support of the idea that tonic dopamine could play an can leave a tag on a specific synapse to facilitate enhance-
important role in the hippocampus, there is substantial ment of later, overlapping inputs exciting the same syn-
anatomical and ultrastructural evidence indicating a apse, even if they are weaker. Synaptic tagging could
prominent role for extrasynaptic (i.e. tonic) dopamine in potentially provide a powerful mechanism for dopamine-
the hippocampus, similar to patterns seen in prefrontal enhanced encoding of individual items or related experi-
cortex, and unlike those in striatum (Box 3). Additionally, ences and could be especially important for understanding
evidence of dopamine release in the hippocampus has been how the hippocampus facilitates encoding over temporal
obtained only via microdialysis, argued to preferentially delays to integrate information across multiple experi-
reflect tonic dopamine levels [65]. ences (e.g. in integrative encoding). Such mechanisms
In fact, evidence that phasic responses in midbrain could also mark synapses for slow structural changes on
dopamine neurons modulate the hippocampus is lacking: the order of minutes or hours [28,29,31,58] that underlie
disrupted phasic activity does not impact long-term spatial lasting memories [31] or potentiate changes during later
memory (as noted above [64]). Further, there is no evidence exposure to dopamine [30], for example in reactivation
from animals or from human fMRI that the hippocampus is [60,61]. These protracted effects predict that dopamine
transiently activated in situations known to elicit phasic might be particularly well-suited to enhancing memories
responses in midbrain dopamine neurons. This indicates after long delays. Pharmacological studies in animals sup-
that temporally-specific prediction error signals might be port this idea [31]. However, this remains an important
an inappropriate model for this system. Although this view open question for future research in humans.
is consistent with extant evidence, the predicted enhance- An important goal for future animal research will be to
ment not only of specific salient events, but also the con- determine how tonic and phasic responses in midbrain
texts in which they occur, remains to be tested. dopamine neurons affect synaptic and extrasynaptic dopa-
It is important to note that these two mechanisms are mine concentrations in the hippocampus to produce
not mutually exclusive. It is possible that midbrain dopa- changes in neural activity and plasticity, and in turn,
mine modulates hippocampal function via both tonic and influence memory formation. At present, all the evidence
phasic responses. Indeed, the range of behavioral contexts for cellular mechanisms comes from rodents. Although
and timescales across which dopamine has been implicated there are likely to be many commonalities across species,
in episodic memory formation seem unlikely to be there are also important differences (see Table I in Box 2)
that raise the need for future research that can bridge
these gaps.
Box 3. Questions for future research
Concluding remarks
 Do tonic and phasic dopamine have dissociable effects on the
hippocampus and long-term memory?
To summarize, extensive evidence indicates that dopamine
 Under what circumstances could these neurobiological mechan- release before, during, and after an event supports hippo-
isms produce maladaptive behavior? campal plasticity and episodic memory formation. Dopa-
 Does learning under appetitive motivation differ in quantity or mine thus seems to influence which episodic memories are
quality from learning under threat? formed and how they are represented, enabling memory for
 Norepinephrine and acetylcholine also signal salience and
uncertainty and also affect hippocampal memory. What are the
past experience to support future adaptive behavior. We
common and distinct mnemonic effects of these neurotransmit- use the term adaptive memory as a construct for this
ters? process to highlight the selectivity of memory and to
 How do other specialized memory systems based in the striatum consider whether this selectivity, which can seem quixotic
and amygdala, also innervated by dopamine, contribute to
or random, could be at least partially explained as a
adaptive memory?
 How is adaptive memory impacted by changes in dopamine with phenomenon that emerges from influences of motivational
healthy aging or in diseases such as schizophrenia? Does systems. We propose that consideration of this construct
dopaminergic medication impact adaptive memory? will enrich understanding of the intrinsic relations be-
 What are the implications of the adaptive memory construct for tween memory, motivation and decision making, at both
learning in educational settings and in everyday life?
the neural and cognitive levels.

470
Review Trends in Cognitive Sciences Vol.14 No.10

The framework we propose on the basis of the adaptive 20 Salamone, J.D. et al. (2007) Effort-related functions of nucleus
accumbens dopamine and associated forebrain circuits.
memory construct is also consistent with findings from the
Psychopharmacology 191, 461482
rich tradition of behavioral studies on memory, and could 21 Niv, Y. et al. (2007) Tonic dopamine: opportunity costs and the control
help resolve questions that have persisted in that litera- of response vigor. Psychopharmacology. 191, 507520
ture for decades without resolution. Wanting to remember 22 Robbins, T.W. and Everitt, B.J. (2007) A role for mesencephalic dopamine
has been an intuitively compelling but experimentally in activation: commentary on Berridge (2006). Psychopharmacology 191,
433437
elusive phenomenon. A framework incorporating dopa-
23 Redgrave, P. et al. (2008) What is reinforced by phasic dopamine
mine as a signal for learning, released in response not signals? Brain Res. Rev. 58, 322339
only to salient events but also to expectations, provides a 24 Samson, Y. et al. (1990) Catecholaminergic innervation of the
mechanistic neurobiological account of how motivation hippocampus in the cynomolgus monkey. J. Comp. Neurol. 298, 250263
influences memory beyond descriptors of internal states. 25 Gasbarri, A. et al. (1994) Anterograde and retrograde tracing of
projections from the ventral tegmental area to the hippocampal
Thus, the adaptive memory construct proposed here could formation in the rat. Brain Res. Bull. 33, 445452
provide insights into understanding not only what we 26 Otmakhova, N.A. and Lisman, J.E. (1998) D1/D5 dopamine receptors
remember, but why. inhibit depotentiation at CA1 synapses via cAMP-dependent
mechanism. J. Neurosci. 18, 12701279
Acknowledgements 27 Huang, Y.Y. and Kandel, E.R. (1995) D1/D5 receptor agonists induce a
The authors are grateful to Lauren Atlas, Nathan Clement, Lila Davachi, protein synthesis-dependent late potentiation in the CA1 region of the
Juliet Davidow, Karin Foerde, Elizabeth Johnson, Jeff Macinnes, Vishnu hippocampus. Proc. Natl. Acad. Sci. U. S. A. 92, 24462450
Murty, and G. Elliott Wimmer for insightful comments on an earlier 28 Frey, U. et al. (1990) Dopaminergic antagonists prevent long-term
draft. maintenance of posttetanic LTP in the CA1 region of rat
hippocampal slices. Brain Res. 522, 6975
References 29 OCarroll, C.M. et al. (2006) Dopaminergic modulation of the
1 Eichenbaum, H.E. and Cohen, N.J. (2001) From Conditioning to persistence of one-trial hippocampus-dependent memory. Learn.
Conscious Recollection: Memory Systems of the Brain, Oxford Mem. 13, 760769
University Press 30 Rossato, J.I. et al. (2009) Dopamine controls persistence of long-term
2 Davachi, L. (2006) Item, context and relational episodic encoding in memory storage. Science 325, 10171020
humans. Curr. Opin. Neurobiol. 16, 693700 31 Bethus, I. et al. (2010) Dopamine and memory: modulation of the
3 Squire, L.R. (1992) Memory and the hippocampus: a synthesis from persistence of memory for novel hippocampal NMDA receptor-
findings with rats, monkeys, and humans. Psychol. Rev. 99, 195231 dependent paired associates. J. Neurosci. 30, 16101618
4 Paller, K.A. and Wagner, A.D. (2002) Observing the transformation of 32 Neugebauer, F. et al. (2009) Modulation of extracellular monoamine
experience into memory. Trends. Cogn. Sci. 6, 93102 transmitter concentrations in the hippocampus after weak and strong
5 Scoville, W.B. and Milner, B. (1957) Loss of recent memory after tetanization of the perforant path in freely moving rats. Brain Res.
bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychiatry 20, 1273, 2938
1121 33 Lisman, J.E. and Grace, A.A. (2005) The hippocampal-VTA loop:
6 Gabrieli, J.D. (1998) Cognitive neuroscience of human memory. Annu. controlling the entry of information into long-term memory. Neuron
Rev. Psychol. 49, 87115 46, 703713
7 Wirth, S. et al. (2009) Trial outcome and associative learning signals in 34 Frank, M.J. et al. (2004) By carrot or by stick: cognitive reinforcement
the monkey hippocampus. Neuron 61, 930940 learning in parkinsonism. Science 306, 19401943
8 Johnson, A. and Redish, A.D. (2007) Neural ensembles in CA3 35 Knowlton, B.J. et al. (1996) A neostriatal habit learning system in
transiently encode paths forward of the animal at a decision point. humans. Science 273, 13991402
J. Neurosci. 27, 1217612189 36 Shohamy, D. et al. (2008) Basal ganglia and dopamine contributions to
9 Hassabis, D. et al. (2007) Patients with hippocampal amnesia cannot probabilistic category learning. Neurosci. Biobehav. Rev. 32, 219236
imagine new experiences. Proc. Natl. Acad. Sci. U. S. A. 104, 17261731 37 Schott, B.H. et al. (2008) Mesolimbic functional magnetic resonance
10 Schacter, D.L. and Addis, D.R. (2007) The cognitive neuroscience of imaging activations during reward anticipation correlate with reward-
constructive memory: remembering the past and imagining the future. related ventral striatal dopamine release. J. Neurosci. 28, 1431114319
Philos. Trans. R. Soc. Lond. B Biol. Sci. 362, 773786 38 Kumaran, D. and Maguire, E.A. (2006) An unexpected sequence of
11 Peters, J. and Buchel, C. (2010) Episodic future thinking reduces events: mismatch detection in the human hippocampus. PLoS Biol. 4,
reward delay discounting through an enhancement of prefrontal- e424
mediotemporal interactions. Neuron 66, 138148 39 Kirchhoff, B.A. et al. (2000) Prefrontal-temporal circuitry for episodic
12 Shohamy, D. and Wagner, A.D. (2008) Integrating memories in the encoding and subsequent memory. J. Neurosci. 20, 61736180
human brain: hippocampal-midbrain encoding of overlapping events. 40 Schott, B.H. et al. (2004) Activation of midbrain structures by
Neuron 60, 378389 associative novelty and the formation of explicit memory in humans.
13 Schultz, W. (1998) Predictive reward signal of dopamine neurons. Learn. Mem. 11, 383387
J. Neurophysiol. 80, 127 41 Axmacher, N. et al. (2010) Intracranial EEG correlates of expectancy
14 Daw, N.D. and Doya, K. (2006) The computational neurobiology of and memory formation in the human hippocampus and nucleus
learning and reward. Curr. Opin. Neurobiol. 16, 199204 accumbens. Neuron 65, 541549
15 Matsumoto, M. and Hikosaka, O. (2009) Two types of dopamine neuron 42 Schott, B.H. et al. (2006) The dopaminergic midbrain participates in
distinctly convey positive and negative motivational signals. Nature human episodic memory formation: evidence from genetic imaging. J.
459, 837841 Neurosci. 26, 14071417
16 Delgado, M.R. et al. (2009) Avoiding negative outcomes: tracking the 43 Li, S. et al. (2003) Dopamine-dependent facilitation of LTP induction in
mechanisms of avoidance learning in humans during fear conditioning. hippocampal CA1 by exposure to spatial novelty. Nat. Neurosci. 6, 526
Front. Behav. Neurosci. 3, 33 531
17 Carter, R.M. et al. (2009) Activation in the VTA and nucleus accumbens 44 Fenker, D.B. et al. (2008) Novel scenes improve recollection and recall
increases in anticipation of both gains and losses. Front. Behav. of words. J. Cogn. Neurosci. 20, 12501265
Neurosci. 3, 21 45 Zald, D.H. et al. (2008) Midbrain dopamine receptor availability is
18 Mazzoni, P. et al. (2007) Why dont we move faster? Parkinsons inversely associated with novelty-seeking traits in humans. J.
disease, movement vigor, and implicit motivation. J. Neurosci. 27, Neurosci. 28, 1437214378
71057116 46 Wittmann, B.C. et al. (2008) Mesolimbic interaction of emotional
19 Berridge, K.C. (2007) The debate over dopamines role in reward: the valence and reward improves memory formation. Neuropsychologia
case for incentive salience. Psychopharmacology 191, 391431 46, 10001008

471
Review Trends in Cognitive Sciences Vol.14 No.10

47 Guitart-Masip, M. et al. (2010) Contextual novelty changes reward 74 Popescu, A.T. et al. (2009) Coherent gamma oscillations couple the
representations in the striatum. J. Neurosci. 30, 17211726 amygdala and striatum during learning. Nat. Neurosci. 12, 801807
48 Bunzeck, N. et al. (2010) A common mechanism for adaptive scaling of 75 Dolcos, F. et al. (2004) Interaction between the amygdala and the
reward and novelty. Hum. Brain Mapp. 31, 13801394 medial temporal lobe memory system predicts better memory for
49 Bunzeck, N. et al. (2009) Reward motivation accelerates the onset of emotional events. Neuron 42, 855863
neural novelty signals in humans to 85 milliseconds. Curr. Biol. 19, 76 Ritchey, M. et al. (2008) Cereb. Cortex 18, 24942504
12941300 77 Poldrack, R.A. and Rodriguez, P. (2004) How do memory systems
50 Knutson, B. et al. (2005) Distributed neural representation of expected interact? Evidence from human classification learning. Neurobiol.
value. J. Neurosci. 25, 48064812 Learn. Mem. 82, 324332
51 DArdenne, K. et al. (2008) BOLD responses reflecting dopaminergic 78 Hartley, T. and Burgess, N. (2005) Complementary memory systems:
signals in the human ventral tegmental area. Science 319, 12641267 competition, cooperation and compensation. Trends Neurosci. 28, 169
52 Wittmann, B.C. et al. (2005) Reward-related FMRI activation of 170
dopaminergic midbrain is associated with enhanced hippocampus- 79 Kroener, S. et al. (2009) Dopamine modulates persistent synaptic
dependent long-term memory formation. Neuron 45, 459467 activity and enhances the signal-to-noise ratio in the prefrontal
53 Krebs, R.M. et al. (2009) The novelty exploration bonus and its cortex. PLoS One 4, e6507
attentional modulation. Neuropsychologia 47, 22722281 80 Peters, Y. et al. (2004) Prefrontal cortical up states are synchronized
54 Adcock, R.A. et al. (2006) Reward-motivated learning: mesolimbic with ventral tegmental area activity. Synapse 52, 143152
activation precedes memory formation. Neuron 50, 507517 81 Gao, W.J. and Goldman-Rakic, P.S. (2003) Selective modulation of
55 Kuhl, B.A. et al. (2010) Resistance to forgetting associated with excitatory and inhibitory microcircuits by dopamine. Proc. Natl. Acad.
hippocampus-mediated reactivation during new learning. Nat. Sci. U. S. A. 100, 28362841
Neurosci. 13, 501506 82 Beck, S.M. et al. (2010) Primary and secondary rewards differentially
56 Cohen, N.J. and Eichenbaum, H. (1993) Memory, amnesia, and the modulate neural activity dynamics during working memory. PLoS One
hippocampal system, MIT Press 5, e9251
57 Heckers, S. et al. (2004) Hippocampal activation during transitive 83 Delgado, M.R. et al. (2005) An fMRI study of reward-related probability
inference in humans. Hippocampus 14, 153162 learning. Neuroimage 24, 862873
58 OCarroll, C.M. and Morris, R.G. (2004) Heterosynaptic co-activation of 84 Sara, S.J. (2009) The locus coeruleus and noradrenergic modulation of
glutamatergic and dopaminergic afferents is required to induce cognition. Nat. Rev. Neurosci. 10, 211223
persistent long-term potentiation. Neuropharmacology 47, 324332 85 Kenney, J.W. and Gould, T.J. (2008) Modulation of hippocampus-
59 Otmakhova, N.A. and Lisman, J.E. (1996) D1/D5 dopamine receptor dependent learning and synaptic plasticity by nicotine. Mol.
activation increases the magnitude of early long-term potentiation at Neurobiol. 38, 101121
CA1 hippocampal synapses. J. Neurosci. 16, 74787486 86 Bergson, C. et al. (1995) Regional, cellular, and subcellular variations
60 Lansink, C.S. et al. (2009) Hippocampus leads ventral striatum in in the distribution of D1 and D5 dopamine receptors in primate brain.
replay of place-reward information. PLoS Biol. 7, e1000173 J. Neurosci. 15, 78217836
61 Singer, A.C. and Frank, L.M. (2009) Rewarded outcomes enhance 87 Floresco, S.B. et al. (2003) Afferent modulation of dopamine neuron
reactivation of experience in the hippocampus. Neuron 64, 910921 firing differentially regulates tonic and phasic dopamine transmission.
62 Lindsten, J. (1992) Nobel Lectures in Physiology or Medicine, World Nat. Neurosci. 6, 968973
Scientific Publishing Co. 690 88 Paspalas, C.D. and Goldman-Rakic, P.S. (2004) Microdomains for
63 Lodge, D.J. and Grace, A.A. (2006) The hippocampus modulates dopamine volume neurotransmission in primate prefrontal cortex. J.
dopamine neuron responsivity by regulating the intensity of phasic Neurosci. 24, 52925300
neuron activation. Neuropsychopharmacology 31, 13561361 89 Little, K.Y. et al. (1995) Characterization and localization of [125I]RTI-
64 Zweifel, L.S. et al. (2009) Disruption of NMDAR-dependent burst firing 121 binding sites in human striatum and medial temporal lobe. J.
by dopamine neurons provides selective assessment of phasic Pharmacol. Exp. Ther. 274, 14731483
dopamine-dependent behavior. Proc. Natl. Acad. Sci. U. S. A. 106, 90 Khan, Z.U. et al. (2000) Dopamine D5 receptors of rat and human brain.
72817288 Neuroscience 100, 689699
65 Floresco, S.B. (2007) Dopaminergic regulation of limbic-striatal 91 Camps, M. et al. (1990) Autoradiographic localization of dopamine D 1
interplay. J. Psychiatry Neurosci. 32, 400411 and D 2 receptors in the brain of several mammalian species. J. Neural
66 Astafiev, S.V. et al. (2010) Comment on Modafinil shifts human locus Transm. Gen. Sect. 80, 105127
coeruleus to low-tonic, high-phasic activity during functional MRI and 92 Lewis, D.A. et al. (2001) Dopamine transporter immunoreactivity in
Homeostatic sleep pressure and responses to sustained attention in monkey cerebral cortex: regional, laminar, and ultrastructural
the suprachiasmatic area. Science 328, 309 localization. J. Comp. Neurol. 432, 119136
67 Duzel, E. et al. (2009) Functional imaging of the human dopaminergic 93 Ciliax, B.J. et al. (2000) Dopamine D(5) receptor immunolocalization in
midbrain. Trends Neurosci. 32, 321328 rat and monkey brain. Synapse 37, 125145
68 Knutson, B. and Gibbs, S.E. (2007) Linking nucleus accumbens 94 Jiao, X. et al. (2003) Strain differences in the distribution of dopamine
dopamine and blood oxygenation. Psychopharmacology 191, 813822 transporter sites in rat brain. Prog. Neuropsychopharmacol. Biol.
69 Bertolino, A. et al. (2008) Epistasis between dopamine regulating genes Psychiatry 27, 913919
identifies a nonlinear response of the human hippocampus during 95 Fiorillo, C.D. et al. (2003) Discrete coding of reward probability and
memory tasks. Biol. Psychiatry 64, 226234 uncertainty by dopamine neurons. Science 299, 18981902
70 Shohamy, D. et al. (2006) L-dopa impairs learning, but spares 96 Williams, S.M. and Goldman-Rakic, P.S. (1993) Characterization of the
generalization, in Parkinsons disease. Neuropsychologia 44, 774784 dopaminergic innervation of the primate frontal cortex using a
71 Volkow, N.D. et al. (2008) Dopamine increases in striatum do not elicit dopamine-specific antibody. Cereb. Cortex 3, 199222
craving in cocaine abusers unless they are coupled with cocaine cues. 97 Oades, R.D. and Halliday, G.M. (1987) Ventral tegmental (A10) system:
Neuroimage 39, 12661273 neurobiology. 1. Anatomy and connectivity. Brain Res. 434, 117165
72 Cools, R. et al. (2009) Striatal dopamine predicts outcome-specific 98 Sesack, S.R. and Grace, A.A. (2010) Cortico-Basal Ganglia reward
reversal learning and its sensitivity to dopaminergic drug network: microcircuitry. Neuropsychopharmacology 35, 2747
administration. J. Neurosci. 29, 15381543 99 Morris, R.G. et al. (2003) Elements of a neurobiological theory of the
73 Bauer, E.P. et al. (2007) Gamma oscillations coordinate amygdalo- hippocampus: the role of activity-dependent synaptic plasticity in
rhinal interactions during learning. J. Neurosci. 27, 93699379 memory. Philos. Trans. R. Soc. Lond. B Biol. Sci. 358, 773786

472