Neurotherapeutics (2017) 14:687–697

DOI 10.1007/s13311-017-0525-z

REVIEW

Reward Circuitry in Addiction

Sarah Cooper 1 & A. J. Robison 1,2 & Michelle S. Mazei-Robison 1,2

Published online: 21 March 2017

# The American Society for Experimental NeuroTherapeutics, Inc. 2017

Abstract Understanding the brain circuitry that underlies re- these regions, with the postulation that increased knowledge
ward is critical to improve treatment for many common health of the plasticity within the drug reward circuit will lead to new
issues, including obesity, depression, and addiction. Here we and improved treatments for addiction.
focus on insights into the organization and function of reward
circuitry and its synaptic and structural adaptations in re- Key Words Cocaine . reward . dopamine . glutamate .
sponse to cocaine exposure. While the importance of certain ventral tegmental area . nucleus accumbens
circuits, such as the mesocorticolimbic dopamine pathway, are
well established in drug reward, recent studies using genetics-
based tools have revealed functional changes throughout the Introduction
reward circuitry that contribute to different facets of addiction,
such as relapse and craving. The ability to observe and ma- Understanding of the brain circuitry that mediates feelings of
nipulate neuronal activity within specific cell types and cir- reward or pleasure is of great interest, with myriad papers
cuits has led to new insight into not only the basic connections published in the last few years identifying regions and con-
between brain regions, but also the molecular changes within nections associated with a variety of natural rewards, includ-
these specific microcircuits, such as neurotrophic factor and ing food, sex, and social interaction. Many of these studies
GTPase signaling or α-amino-3-hydroxy-5-methyl-4- have taken advantage of increasingly sophisticated genetics-
isoxazolepropionic acid (AMPA) receptor function, that un- based tools to observe and manipulate neuronal activity within
derlie synaptic and structural plasticity evoked by drugs of specific cell types and circuits to ascertain their role in reward
abuse. Excitingly, these insights from preclinical rodent work processes. However, much of what we know of the structure
are now being translated into the clinic, where transcranial of the reward circuit, and the generation of pleasure, was orig-
magnetic simulation and deep brain stimulation therapies are inally identified in the context of drugs of abuse. Thus, our
being piloted in human cocaine dependence. Thus, this review understanding of the circuitry underlying the rewarding as-
seeks to summarize current understanding of the major brain pects of drug use, and maladaptive reward underlying addic-
regions implicated in drug-related behaviors and the molecu- tion, informs the greater understanding of general reward
lar mechanisms that contribute to altered connectivity between mechanisms. Moreover, while some specific nodes of the re-
ward circuitry, such as dopamine (DA) outputs from the ven-
tral tegmental area (VTA) to nucleus accumbens (NAc), and
* Michelle S. Mazei-Robison
their importance to drug reward, are well established, our un-
[email protected] derstanding of the complexity of the reward circuit underlying
various aspects of addiction, such as relapse and craving, has
1
Neuroscience Program, Michigan State University, East Lansing, MI, increased through the use of Bcircuit-busting^ opto- and
USA chemogenetic approaches. Thus, the goal of this review is to
2
Department of Physiology, Michigan State University, East summarize the current understanding of the major players
Lansing, MI, USA (brain regions) in drug reward and the role of the connections
688 Cooper et al.

between these regions (circuitry) in aspects relevant to addic-

tion and to connect greater understanding of these reward
circuit dynamics to potential improvements in the treatment
of drug addiction.

Wiring of the Reward Circuitry

VTA

The general understanding of the reward circuitry underlying

addiction begins with the VTA, a heterogeneous brain region
composed largely of DA (60–65%) and γ-amino butyric acid
(GABA; ~30–35%) neurons, with a smaller proportion of glu-
tamatergic neurons (2–3%) [1, 2]. Most studies have focused
on the VTA DA neurons, as stimulation of these neurons and
the release of DA in projection sites, most notably the NAc, is
Fig. 1 Schematic of brain reward circuitry implicated in addiction.
known to produce reward. Early work has shown that virtually
Dopaminergic (DA; green) and glutamatergic (Glut; red) inputs
all known drugs of abuse increase DA release in the NAc [3], converge on γ-aminobutyric acid (GABA)ergic (blue) neurons in the
and blocking the action of the DA (via receptor blockade) nucleus accumbens (NAc) to coordinate and regulate drug-related
blocks many behavioral effects of drugs [4]. Furthermore, behaviors. BLA = basolateral amygdala; D1 = dopamine type 1
receptor; D2 = dopamine type 2 receptor; LDT = laterodorsal tegmentum;
VTA DA activation is sufficient to support self-stimulation
LHb = lateral habenula; MDT = mediodorsal thalamus; PFC = prefrontal
and both noncontingent and contingent reward behavior, as cortex; RMTg = rostromedial tegmentum; STN = subthalamic nucleus;
has been recently demonstrated through the use of Thal = thalamus; vHPC = ventral hippocampus; VP = ventral pallidum;
optogenetics (detailed below; [5–7]). IL = infralimbic; PrL = prelimbic; OFC = orbitofrontal cortex;
VTA = ventral tegmental area;
However, through tracing and functional studies, it has
become clear that VTA DA neurons are not as homogenous
as previously thought, and can be subdivided by projection likely to address this key issue, as molecular profiling tech-
target. The 2 most extensively studied VTA DA neuron sub- niques such as translating ribosome affinity purification, and
types to date are those that project to the NAc (mesolimbic) more recently retro-translating ribosome affinity purification
and prefrontal cortex (PFC; mesocortical), though there are [12], have been pioneered and used to specifically isolate
additional projection sites, including the amygdala and hippo- mRNA from VTA DA neurons that project to NAc [13], for
campus (Fig. 1). Current evidence suggests that VTA DA example.
neuron subtypes have distinct electrophysiological properties While the focus of this review is on reward, and thus many
and functional outcomes [8, 9]. For example, a rewarding of the studies discussed here focus on these VTA DA projec-
stimulus (cocaine) selectively modulates excitatory input to tion neurons, it is also clear that VTA GABA neurons are
VTA DA neurons that project to NAc, in contrast to an aver- critical for reward processing. The role of these GABA neu-
sive stimulus (hindpaw formalin injection) that selectively rons in reward and aversion has been elegantly demonstrated
modulates synaptic input onto VTA DA neurons that project via optogenetics, where activation of VTA GABA neurons
to medial PFC (mPFC) [10]. Further, these differences are produces real-time aversion, and inhibition of these neurons
behaviorally relevant and constitute distinct microcircuits: ac- produces preference. These effects are thought to depend
tivation of glutamatergic neurons in laterodorsal tegmentum largely on the activity of VTA GABA interneurons, as VTA
synapsing onto VTA DA neurons projecting to NAc increases GABA activation inhibited VTA DA neuronal activity and
reward behavior, while activation of glutamatergic lateral produced a concomitant decrease in DA release in NAc, while
habenula neurons that innervate VTA DA neurons that project terminal activation of VTA GABA neurons in NAc was suf-
to the mPFC induces aversive behavior [11]. The lateral ficient to increase GABA concentration but not to alter reward
habenula neurons also innervate GABA neurons in the behavior [14, 15].
rostromedial tegmentum nucleus, or Btail^ of the VTA, which
exert inhibitory tone specifically on VTA DA neurons NAc
projecting to NAc (Fig. 1) [11]. Critically, while it is increas-
ingly clear that VTA DA neurons should be classified based As described above, a major reward-related output of VTA DA
on their connectivity, little is known about molecular differ- activity is NAc, or ventral striatum. Here, DA can exert its
ences between these subtypes of neurons. Future studies are effects via activation of DA receptors located on medium spiny
Reward Circuitry in Addiction 689

neurons (MSNs), the predominant cell type in NAc. MSNs are information from BLA and locus coeruleus [28–30]. Thus,
GABAergic projection neurons and consist of 2 (largely) sep- the vHIPP–NAc connection acts to provide contextually rele-
arate classes defined by their expression of either D1- or D2- vant emotional information to influence goal-directed behav-
like DA receptors [16, 17]. D1 and D2 MSNs in the dorsal ior. This circuit has been implicated in both reward and aver-
striatum are thought to have generally separate projections, sive behaviors, and its modulation has been shown to impact
with D1 MSNs constituting the Bdirect^ pathway (ultimately locomotor responses to drugs of abuse and cue-induced drug-
increasing thalamocortical drive) and D2 MSNs forming the seeking behavior [31, 32].
Bindirect^ pathway (ultimately decreasing thalamocortical
drive). However, evidence suggests that this direct versus indi-
rect pathway distinction is not as clear in NAc as it is in dorsal Amygdala
striatum, given that only NAc D1 MSNs project to the VTA,
while both D1 MSNs and D2 MSNs project to the ventral The amygdala also sends glutamatergic input to NAc, and
pallidum (VP) (Fig. 1) [18, 19]. Investigation of the role of its effects are thought to be mediated by D1 DA receptor
D1 versus D2 DA signaling in NAc and its role in addictive activation [33, 34]. In particular, activation of BLA–NAc
behavior has been of long standing interest in the field, initially projections facilitates reward seeking and supports posi-
through the use of agonists and antagonists [4], and more re- tive reinforcement [34, 35]. While there is a well-
cently using optogenetic approaches [20]. In addition to the established role of amygdala activation in emotional
cellular heterogeneity (D1 vs D2 cells), there is also regional learning, and for projections from the BLA in particular
heterogeneity, with distinct drug-associated behaviors and plas- to mediate fear- and anxiety-behaviors, these effects ap-
ticity differences noted between the NAc core and shell subre- pear to be mediated distinctly from those of BLA projec-
gions [21]. MSNs in NAc core appear to be critical for tions to NAc, as the BLA–NAc microcircuit drives reward
assigning motivational value to discrete stimuli associated with and reinforcement [36].
reward or aversion, and particularly updating these values as
circumstances change, while those in the NAc shell drive be-
havioral responses to repeated exposure to rewarding experi- Thalamus
ences, such as chronic drug administration [22]. While VTA
DA neurons provide a strong modulatory input onto NAc More recently, glutamatergic inputs into NAc from subre-
MSNs, they also receive a large amount of glutamatergic input gions of the thalamus, and especially midline thalamic nu-
from a variety of limbic and cortical regions, some of the most clei like the paraventricular nucleus (PVT), have been
notable being PFC, ventral hippocampus (vHIPP), and characterized. In contrast to glutamatergic inputs from the
basolateral amygdala (BLA) (Fig. 1) [23, 24]. regions described above (vHIPP, PFC, amygdala) where
stimulation is rewarding, direct activation of the PVT–
PFC NAc pathway is aversive, driving behavioral aversion in
a real-time place preference assay [37]. Additionally, alter-
PFC input to NAc is largely associated with executive control ation of PVT activity has been shown to alter drug-related
and thought to mediate goal-directed behaviors such as the behaviors like cocaine reward and seeking [38–40]. These
seeking and planning of action to obtain reward-related sub- behavioral changes are likely driven by changes in drug-
stances (like drugs of abuse) [25]. PFC subregions differ in induced plasticity within the medial thalamus–NAc circuit,
their projections to NAc, with infralimbic (IL) mPFC prefer- as cocaine experience has been shown to alter N-methyl-D-
entially projecting to NAc shell and prelimbic mPFC (PrL) to aspartate (NMDA) function and plasticity and increase si-
NAc core [23]. However, pharmacological and optogenetic lent synapses within this circuit, with plasticity changes
manipulation of specific PFC inputs to NAc has shown that dependent on both MSN cell type (D1 vs non-D1) and
the cocaine-related synaptic plasticity and behavior differs be- subregion (shell vs core) [41, 42].
tween PFC subregions [26, 27]. This fascinating complexity is Together, the regions described above form a highly
at the forefront of addiction research, and examples of these integrated circuit, the cortico-basal ganglia reward network
sometimes competing or complementary adaptations are de- [23], with drug-induced changes in plasticity within the
scribed below (in BPFC–NAc Stimulation^ subsection). circuit contributing to various facets of addiction [43].
The function of this circuitry is made all the more complex
vHIPP by the fact that these many regions, and the specific pro-
jection neurons connecting them, experience a wide range
vHIPP also sends glutamatergic projections to NAc and is of molecular and functional changes in response to drug
thought to act as a site of integration between spatial/ exposure, often with opposing or seemingly contradictory
contextual information from dorsal HIPP and emotional effects on behavior.
690 Cooper et al.

Molecular Mechanisms of Drug-Evoked Synaptic reinstatement. Interestingly, TC normalized cocaine-induced

and Structural Plasticity changes in phospho-TrkB in both NAc and PFC, normalizing
both the cocaine-induced increase in NAc and the cocaine-
It is well established that drug dependence and addiction in- induced decrease in PFC. The mechanisms driving the
volve changes in both synaptic and structural plasticity within brain-region specific changes in phospho-TrkB following co-
the reward circuitry [44, 45]. Some of the key molecular sub- caine self-administration remain unclear, as does the mecha-
strates underlying these changes in NAc have been identified, nism by which TC is able to reverse these changes in both
including alteration of growth factor and GTPase signaling brain regions, but this differential effect aligns with previous
and regulation of α-amino-3-hydroxy-5-methyl-4- data showing opposite effects of decreased TrkB signaling in
isoxazolepropionic acid (AMPA) receptor insertion and activ- NAc and VTA and suggests that TrkB antagonism may be a
ity. Excitingly, many studies are now expanding on this work viable option for treatment of cocaine addiction. Thus, future
to investigate circuit-specific changes in molecular mecha- studies should assess not only changes in signaling in specific
nisms that underlie synaptic and structural plasticity related reward brain regions, but also the overall effect throughout the
to drug addiction. Here, we present highlights from a few reward circuitry.
recent studies pushing the field to incorporate a circuit-level
understanding of molecular and functional changes underly- Small GTPase Signaling
ing drug responses.
Drugs of abuse also induce changes in neuronal structure, with
Brain-Derived Neurotrophic Factor one of the most-studied aspects being the change in number
and structure of dendritic spines in NAc after cocaine expo-
Neurotrophic factor signaling, and, in particular, that related to sure [44, 61, 63]. The maturation of dendritic spines from
brain-derived neurotrophic factor (BDNF) action at the immature, thin spines to mature mushroom-shaped spines is
tropomyosin kinase B (TrkB) receptor, and its downstream associated with protein synthesis and receptor insertion (such
effects, have been implicated in drug dependence and addic- as an increase in functional synaptic AMPA receptors) that
tion [46]. Multiple studies find that BDNF expression is in- corresponds to increased excitatory synaptic function [64].
creased in response to both self- and investigator-administered In NAc, changes in MSN spine structure and synaptic strength
cocaine throughout the reward circuitry, including in NAc, vary as a function of time. Initially following cocaine expo-
PFC, VTA, and amygdala [47–51]. However, the effect of sure there is an increase in immature spines and decreased
increased BDNF expression or signaling on cocaine-related synaptic strength (see discussion of silent synapses below),
behaviors is region-specific [52]. For example, increasing while at longer time points postexposure (weeks to months),
BDNF in the NAc versus PFC produces opposite effects on there is an increase in mature spines and synaptic strengthen-
cocaine intake. Increased BDNF in NAc leads to increased ing [61, 65–67]. Importantly, these changes also correlate with
cocaine reward, locomotor activity, self-administration, and behavior, as a similar early/late time course is observed in
reinstatement behaviors, while decreasing BDNF expression behavioral assays of cocaine seeking and reward [26,
or activity of the TrkB receptor decreases these same behav- 67–69]. Given that dendritic spines are the prominent site
iors [47, 53–56]. Alternatively, in PFC increasing BDNF re- for excitatory transmission, understanding circuit-specific
duces cocaine seeking, while decreasing BDNF results in in- changes in dendritic spine structure and function may yield
creased intake [57–60]. important insight into altered circuit activity induced by co-
Given that changes in BDNF could also mediate structural caine exposure that may underlie addiction-related behaviors,
and synaptic plasticity relevant to cocaine addiction through such as relapse to drug seeking.
known actions modulating dendritic spine morphology and To this end, a recent study describes the role of the small
synaptic plasticity [46, 61], the question of whether BDNF GTPase Rap1b in cocaine-mediated dendritic remodeling in
could be a therapeutic target for cocaine addiction is depen- NAc, where Rap1b is elevated 24 h following cocaine expo-
dent on whether competing changes within the reward circuit sure (administrator or self-administration) and is reduced
could abrogate benefits. Investigators have now begun to ex- 4 weeks later [70], analogous to the time course for spine
plore this possibility, as a recent paper described the use of a changes. Further, NAc-specific knockout of Rap1b is suffi-
brain-penetrant TrkB antagonist [Tat-cyclotraxin-B (TC)] in cient to prevent cocaine induction of immature spines, and
rats undergoing long access or short access cocaine self- increasing Rap1b is sufficient to increase cocaine-driven lo-
administration [62]. Systemic administration of TC decreased comotor activity and reward, suggesting it plays a key role in
cocaine intake in both long access and short access rats, both cocaine-induced changes in dendritic spine density and sub-
during fixed ratio and progressive ratio testing, without affect- sequent behavior. Critically, the authors then go on to explore
ing glucose/saccharin intake. Further, systemic administration which potential glutamatergic inputs are regulating these
of TC was also sufficient to reduce cocaine-induced changes in spine density, and they find that optogenetic
Reward Circuitry in Addiction 691

stimulation of glutamatergic terminals from IL, but not vHIPP [43, 73, 75]. Elucidation of common molecular mechanisms
or BLA, increases Rap1b in NAc. Further, IL–NAc terminal that facilitate drug-dependent remodeling of both specific re-
stimulation was sufficient to induce a significant cocaine con- ward microcircuits and general reward-related synapses re-
ditioned place preference (CPP) effect, and this was abolished mains a critical area of study, and a better understanding of
in mice with knockdown of Rap1b in NAc. These data nicely functional outputs of these mechanisms will be critical for
align with earlier results showing an increase in immature improved addiction treatment.
synapses in the IL-NAc circuit immediately following cocaine
self-administration (when Rap1b is increased), followed by
synaptic maturation after prolonged withdrawal, producing Optogenetic Dissection of Drug-Related Behaviors
an antirelapse effect [26], consistent with decreased Rap1b
expression at the same time point. Importantly, in this study, VTA DA Stimulation
there was also a circuit-specific effect, as remodeling of the
PrL NAc versus IL NAc projections following prolonged ab- Optogenetic tools allow direct control of a defined subset of
stinence yielded opposite effects, with PrL projections facili- neurons (via genetic signature) in freely moving animals
tating cocaine incubation/craving, and IL-NAc contributing to through the expression of light-activated channels that in-
an antirelapse phenotype. crease (channelrhodopsin, ChR2) or inhibit (archaerhodopsin;
halorhodpsin) neuronal activity [76]. Optogenetic approaches
AMPA Receptor Function have proven instrumental in identifying circuits necessary for
various aspects of reward and addiction [77]. For example,
This synaptic remodeling is not simply a change in the gross early work confirmed the role for VTA DA neuron activity
structure of the postsynaptic element (the spine), but also in- in reward-related behaviors, such that phasic optical stimula-
volves changes in the synaptic machinery that alter the func- tion of VTA DA neurons expressing channelrhodopsin was
tion of the glutamatergic synapse [71]. For instance, immedi- sufficient to produce CPP and increased DA release in the
ately after prolonged exposure to cocaine, NAc MSNs exhibit NAc [5]. Follow-up work demonstrated that optical activation
an increase in the number of Bsilent synapses^—a signature of of VTA DA neurons was also sufficient for development of
immature glutamatergic synapses that contain only NMDA positive reinforcement in a food-seeking operant task [78] and
receptors without stably expressed AMPA receptors [72]. to support intracranial self-stimulation [6]. More recently, the
However, after extended withdrawal from cocaine (ranging role of VTA DA activity in reward behaviors more relevant to
from 10 to 45 days), these synapses become Bunsilenced^ addiction, specifically compulsive use despite negative conse-
through synaptic insertion of AMPA receptors, including the quences, has been investigated using optogenetics [7]. In this
Ca2+-permeable GluA2-lacking variety, a process that directly study, mice were trained in an operant stimulation task in
correlates with the incubation of drug-seeking behavior, or which an active lever press resulted in a burst stimulation of
increased Bcraving^ [73]. VTA DA neurons. Mice readily acquired the optogenetic self-
Moreover, this effect appeared to be circuit-specific, as stimulation behavior, and this reinforcement appeared to rely
both the electrophysiological and behavioral outputs were di- on the same circuits as drugs of abuse, as an injection of
rectly linked to BLA inputs onto NAc MSNs [73, 74]. In cocaine was capable of decreasing self-stimulation in a dose-
parallel, another group investigated circuit-specific AMPA- dependent manner. In a further demonstration of the similarity
dependent plasticity of vHIPP and mPFC inputs onto NAc between optogenetic self-stimulation and cocaine self-admin-
MSNs in the context of 1 month of withdrawal from cocaine istration, mice exhibited similar cue-induced seeking behavior
self-administration [31]. They found that cue-induced cocaine and changes in NAc synaptic plasticity (e.g., an increased
seeking at this time point correlated with reduced AMPA/NMDA ratio in D1, but not D2, MSNs), following
AMPA/NMDA ratio at mPFC inputs onto NAc D1 MSNs, 30 days of abstinence from VTA DA self-stimulation, very
while AMPA/NMDA ratio was increased at vHIPP inputs. similar to that observed following cocaine [31, 79]. Finally,
Importantly, simultaneous reversal of these synaptic modifi- in a subset of mice, VTA DA optical self-stimulation was
cations at both inputs prevented cue-induced cocaine seeking, sufficient to induce continued responding in the face of an
while selective reversal at mPFC inputs impaired response electric foot shock punishment paired with the optical stimu-
discrimination and selective reversal at vHIPP inputs reduced lation, similar to results observed for cocaine self-
response vigor. administration in rat studies [80], suggesting that VTA DA
In combination, these studies demonstrate that plasticity of activity is sufficient to mediate the compulsive-like behavior
individual microcircuits may mediate specific aspects of drug- in the face of adverse consequences associated with addiction.
related behavior, but the entire reward circuitry is altered by The authors went on to explore which brain regions might
chronic exposure to, and withdrawal from, addictive drugs, contribute to the phenotype of continuing to self-stimulate in
and these alterations likely drive drug craving and relapse the presence of punishment, and identified the orbitofrontal
692 Cooper et al.

cortex (OFC) as a critical mediator of this response: the OFC PFC–NAc Stimulation
was activated specifically in mice resistant to punishment,
OFC pyramidal neurons were more excitable in resistant mice There is also evidence that glutamatergic afferents to the NAc
compared with naïve and yoked controls, and silencing OFC differentially support reward behavior. For example,
neurons was sufficient to convert punishment resistant mice to optogenetic stimulation of NAc axonal terminals originating
susceptible. These studies highlight the facility of optogenetic from BLA neurons was sufficient to promote self-stimulation,
approaches to confirm activity of specific neurons/brain re- indicative of reward, while simulation of terminals originating
gions in addictive behavior and use these approaches to model from mPFC neurons was not [34]. Moreover, work from nu-
addictive behavior itself, in the absence of drugs, to identify merous groups suggests that the PFC–NAc glutamatergic cir-
common underlying mechanisms. cuit is remodeled in rodent models of addiction, and that these
changes play an integral role in drug-seeking behavior follow-
ing forced abstinence or extinction from chronic cocaine.
NAc MSN Stimulation However, the role of the mPFC is complicated, as effects
differ between mPFC subregions. In general, evidence sug-
While the activity of VTA DA neurons is critical to drive gests that activation of PrL–NAc projections is important for
initial reward-related drug seeking, a critical nexus for the reinstatement behavior following cocaine extinction [88–90],
progression to addiction is thought to occur in NAc, where while inactivation of IL–NAc reinstates seeking behavior fol-
the reward-related information (DA) modulates glutamatergic lowing extinction [91, 92]. Similar results were recently ob-
inputs from cortical and limbic regions to produce behavior. served following forced withdrawal in the cocaine incubation
As discussed above, the predominant neurons in NAc are paradigm [54, 79]. While similar neuroplasticity (maturation
MSNs, and considerable effort has been put forth to under- of silent synapses) was induced in both the PrL–NAc and IL–
stand the changes in function induced by drugs of abuse in D1 NAc circuits, optogenetic reversal of the plasticity produced
versus D2 MSNs. While the use of D1- and D2-selective opposite behavioral effects within each circuit, as the IL–NAc
agonists and antagonists has been instrumental in defining reversal led to increased incubation of cocaine craving and
the often opposing effects of these 2 subclasses of neurons, PrL–NAc reversal inhibited incubation [26].
optogenetic tools offer a mechanism to investigate changes in Indeed, while it is clear that glutamatergic regulation of
the activity of these neurons, and their connections, in real NAc is critical for various aspects of addiction, including drug
time. For example, early optogenetic work found that in con- seeking and relapse, the roles of specific inputs are complex.
trast to VTA DA neuron optical activation, D1 or D2 MSN Results differ between the region of PFC examined (PrL, IL,
activation did not generate place preference [81]. However, or orbitofrontal), the cocaine-related paradigm and time point
activation of D1 MSNs during cocaine exposure was suffi- examined (reinstatement, resistance to punishment, seeking,
cient to induce CPP to a subthreshold dose of cocaine, while or escalation), the stimulation parameters, and the species ex-
D2 MSN activation produced the opposite result, decreasing amined [27, 93]. Thus, though current results present a com-
CPP when administered during training. While both D1 and plex and cloudy picture of the molecular mechanisms of
D2 MSNs receive dopaminergic and glutamatergic input from circuit-specific brain alterations underlying addiction, the
similar structures (e.g., VTA, BLA, mPFC, vHIPP, midline tools for clarifying and integrating these results are becoming
nuclei of the thalamus), the cocaine-induced changes in syn- available, and a clearer understanding of these mechanisms
aptic plasticity can be cell type-specific, as are their outputs, will be a critical step for uncovering circuit-specific therapies
which likely contribute to their differential effects on cocaine- for addiction.
related behavior [19]. This has recently been shown to be the
case in the VP, a robust target of both D1 and D2 NAc MSNs,
where cocaine exposure simultaneously potentiated D1 NAc– Circuit-Based Therapeutics for Addiction
VP output, but depressed D2 NAc–VP output [82]. Further,
these plasticity changes mediated distinct aspects of addictive- While considerable headway has been made in isolating the
like behaviors. Restoration of function at D1 NAc–VP synap- activity of specific neural circuits in aspects of addictive be-
ses eliminated cocaine locomotor sensitization, while rescue havior using preclinical models, the translation of this knowl-
at D2 NAc–VP synapses reversed cocaine withdrawal- edge to improve the diagnosis and treatment of addiction in
induced anhedonia, suggesting that NAc D1 versus D2 humans lags behind. The solution to this problem rests, at least
MSN efferents to VP regulate distinct behavioral states [82]. in part, in defining the effects of drug addiction on the human
This work highlights not only the differential impact of D1 reward circuitry using imaging approaches. Much work has
versus D2 MSN circuits, but also the importance of the NAc– been done in this area, with many reports of drug-induced
VP circuit in cocaine-related behaviors, a topic of great recent changes in both structure and function of regions within in
interest [83–87]. the reward circuitry and many excellent reviews summarizing
Reward Circuitry in Addiction 693

this work [94, 95]. In this section, we will specifically focus on depressive-, anxiety-, and sleep-related symptoms often expe-
PFC and striatum (dorsal and ventral) and the connections rienced when cocaine addicts stop using cocaine [100]. The
between them, as these have offered the primary targets for rTMS group had a significantly reduced relapse rate (all drug
approaches to alter human circuit function in the treatment of urine screens were negative during initial 4-week treatment
addiction. stage) compared with the control group. Moreover, in the sec-
In aggregate, current work suggests that PFC activity is ond stage of the study, the pharmacological control group was
decreased in drug abusers, with specific PFC subregions, in- given the option to receive rTMS treatment, which resulted in
cluding the dorsolateral PFC (dlPFC), mPFC, and OFC being a significantly reduced relapse rate in this group as well
of particular interest. Within these regions, there is evidence of (within-group comparison) that was indistinguishable from
both structural (e.g., decreased gray matter volume or cortical the relapse rate of patients receiving rTMS in stage 1. The
thickness) and functional (e.g., positron emission tomography, improvement in relapse rate was mirrored by decreased drug
functional magnetic resonance imaging, and blood-oxygen craving in the rTMS group. A second study examined the
level-dependent (BOLD) activity) abnormalities in cocaine effect on cocaine intake of bilateral PFC rTMS treatment of
abusers [94]. Alterations in the structure and function of these cocaine-addicted patients [101]. In contrast to the previous
subregions are thought to contribute to changes in limbic study, no differences in cocaine intake were observed between
arousal, executive function, and reward valuation that then rTMS and control (sham stimulation) groups immediately fol-
contribute to behaviors such as craving, risk taking, and out- lowing 4 weeks of treatment. However, there was a significant
come prediction errors that underlie addiction and relapse. time effect, as the rTMS group showed a trend for decreased
Moreover, altered activity within PFC subregions contributes cocaine intake 3 to 6 months following treatment. There were
to altered activity in the dorsal (caudate nucleus and putamen) multiple differences in the design of these 2 studies, which
and ventral (NAc) striatum. Changes in dorsal striatal activity may have contributed to the results, including the rTMS pro-
are most associated with changes in movement, cognitive pro- tocol (unilateral vs bilateral), control comparison (pharmaco-
cessing and habit formation, and stronger connectivity to the logical vs sham stimulus), and drug screen (urine vs hair anal-
dlPFC. In contrast, the ventral striatum is more highly con- ysis), that were likely amplified by the small sample sizes (32
nected to the mPFC and OFC, and changes in activity in this and 18 patients, respectively).
region are more relevant to limbic control, and arousal states While both of these studies involved a small number of
such as craving and Bhigh^. individuals, the results are promising and suggest that rTMS
can be safely used in cocaine-addicted patients. Further refine-
Repetitive Transcranial Magnetic Stimulation ment of TMS approaches may offer a novel treatment for
for Treatment of Addiction cocaine addiction, a critical advance given the lack of current-
ly approved treatments. This includes the use of rTMS in other
Research has also begun to translate findings from optogenetic regions, such as mPFC. In this case, low-frequency stimula-
studies in rodent models and apply them to human addicts. tion would be used to decrease cortical activity in order to
One strategy is to use repetitive transcranial magnetic stimu- normalize the increased activity induced by drugs of abuse
lation (rTMS), as this is a noninvasive approach to alter neu- [102]. Promising pilot data find that long-term depression
ronal activity that has been applied with success to other neu- (LTD) -like continuous theta burst stimulation decreased co-
ropsychiatric disorders such as depression [96]. Clinical trials caine craving in cocaine-dependent patients and decreased
using rTMS for addiction are ongoing, with studies investi- activity both in PFC and in projection regions such as the
gating whether rTMS treatment is capable of decreasing crav- ventral striatum (via BOLD) immediately following a single
ing for, or use of, multiple drugs of abuse, including alcohol, rTMS stimulation compared with nonstimulated controls.
cocaine, methamphetamine, and tobacco [97]. Most studies to While early results are encouraging, there are issues surround-
date have employed high-frequency rTMS (10–20 Hz) in the ing TMS use that could ultimately limit its effectiveness in the
dlPFC in order to increase neuronal excitability and cortical treatment of addiction including constraints on the specificity
activity [98, 99], as work from animal studies has shown that of stimulation target region, the frequency of treatments nec-
activity in PrL (the homologous PFC region in rodents) is essary, and the long-term persistence of any beneficial effects.
decreased by chronic cocaine use, and optogenetic stimulation
of PrL inhibits compulsive cocaine seeking [80]. Thus, rTMS Deep Brain Stimulation for Treatment of Addiction
of dlPFC in addicted individuals is expected to increase cor-
tical activity and improve cognitive/executive control. Another strategy proposed to translate the circuit-based find-
Two recent pilot studies have examined the effect of dlPFC ings of preclinical models to human addiction is the use of
rTMS on cocaine use in cocaine-addicted patients. The first deep brain stimulation (DBS) [71]. DBS is the electrical stim-
compared unilateral rTMS treatment of cocaine-addicted pa- ulation of discrete brain regions through surgical implantation
tients with a pharmacological control group, to offset potential of current passing electrodes. While pioneered for the
694 Cooper et al.

treatment of Parkinson’s disease, DBS has been employed appear to be drug-class selective [46, 111, 112], which could
more recently in the treatment of psychiatric disorders such present additional complexity in designing translational stud-
as depression [103]. However, DBS does not offer the selec- ies. However, the groundwork for a circuit model of addiction
tivity of preclinical optogenetic approaches, where activity of has been laid, and the insights gained from preclinical studies
specific cells or projections within a region can be modulated, have started to be translated to the clinic. Alteration of reward
and this specificity may be critical in regions such as NAc, circuit activity through rTMS and DBS offers a promising
where general stimulation could prove ineffective owing to new avenue for treatment of addiction, and is a welcome ad-
competing actions of multiple forms of plasticity (DA vs glu- vance given the lack of effective treatments for this devastat-
tamate or input projection-specific effects). While general ing disorder.
DBS approaches have been used with some success in pre-
clinical animal models [104–106], recent work has attempted Acknowledgments The work conducted in the Robison and Mazei-
to refine the specificity of DBS by designing a protocol to Robison laboratories is currently supported by grants from the National
Institutes on Mental Health (MH111604, AJR) and Drug Abuse
reverse specific forms of plasticity induced by drugs of abuse (DA039895, MMR).
in combination with the use of a pharmacological adjuvant
that helps to eliminate opposing effects from the general stim- Required Author Forms Disclosure forms provided by the authors are
ulation [107]. Termed optically inspired DBS, a stimulation available with the online version of this article.
protocol was developed to induce LTD of excitatory synapses
in the NAc in order to reverse cocaine-induced changes in
neuroplasticity. This stimulation was then combined with ad- References
ministration of D1 receptor antagonist to block competing
effects of the stimulation on DA release, thereby Bunmasking^ 1. Swanson LW. The projections of the ventral tegmental area and
adjacent regions: a combined fluorescent retrograde tracer and
the desired glutamatergic plasticity. Critically, this approach
immunofluorescence study in the rat. Brain Res Bull 1982; 9:
was sufficient to both reverse the cocaine-induced changes in 321-353.
synaptic plasticity and cocaine locomotor sensitization. This 2. Nair-Roberts RG, Chatelain-Badie SD, Benson E, White-Cooper
study serves as an intriguing example of an approach to inte- H, Bolam JP, Ungless MA. Stereological estimates of dopaminer-
grate the hard-earned knowledge of specific molecular and gic, GABAergic and glutamatergic neurons in the ventral tegmen-
tal area, substantia nigra and retrorubral field in the rat.
synaptic adaptations induced by drugs of abuse identified in Neuroscience 2008; 152:1024-1031.
animal models and apply them to improve treatment of addic- 3. Di Chiara G, Imperato A. Drugs abused by humans preferentially
tion. Given that pilot studies in the use of DBS to treat addic- increase synaptic dopamine concentrations in the mesolimbic sys-
tion have now begun for alcohol, cocaine, and heroin addic- tem of freely moving rats. Proc Natl Acad Sci U S A 1988; 85:
5274-5278.
tion with some promising initial results [108–110], the imple-
4. Anderson SM, Pierce RC. Cocaine-induced alterations in dopa-
mentation of drug-reward circuitry knowledge to affect thera- mine receptor signaling: implications for reinforcement and rein-
peutic potential may have the field poised for a breakthrough. statement. Pharmacol Ther 2005; 106:389-403.
5. Tsai HC, Zhang F, Adamantidis A, et al. Phasic firing in dopami-
nergic neurons is sufficient for behavioral conditioning. Science
2009; 324:1080-1084.
Conclusions and Future Directions
6. Witten IB, Steinberg EE, Lee SY, et al. Recombinase-driver rat
lines: tools, techniques, and optogenetic application to dopamine-
In summary, our understanding of the circuitry underlying mediated reinforcement. Neuron 2011; 72:721-733.
drug-related behaviors has increased greatly in recent years, 7. Pascoli V, Terrier J, Hiver A, Luscher C. Sufficiency of
driven by the development and implementation of opto- and mesolimbic dopamine neuron stimulation for the progression to
addiction. Neuron 2015; 88:1054-1066.
chemogenetic approaches. Excitingly, these circuit-based ap-
8. Lammel S, Lim BK, Malenka RC. Reward and aversion in a
proaches are now being used to explore the molecular mech- heterogeneous midbrain dopamine system. Neuropharmacology
anisms responsible for changes in drug-evoked structural and 2014; 76 Pt B:351-359.
synaptic plasticity that underlie the maladaptive behavior. 9. Lammel S, Hetzel A, Hackel O, Jones I, Liss B, Roeper J. Unique
While these genetics-based studies have allowed for elegant properties of mesoprefrontal neurons within a dual
mesocorticolimbic dopamine system. Neuron 2008; 57:760-773.
dissection and elucidation of specific microcircuits relevant to
10. Lammel S, Ion DI, Roeper J, Malenka RC. Projection-specific
various aspects of addiction, the challenge will be in integrat- modulation of dopamine neuron synapses by aversive and reward-
ing the findings from these distinct microcircuits back into a ing stimuli. Neuron 2011; 70:855-862.
general model of addiction, given the complexity and inter- 11. Lammel S, Lim BK, Ran C, et al. Input-specific control of reward
connectedness of the brain-reward circuitry they have already and aversion in the ventral tegmental area. Nature 2012; 491:212-
217.
revealed. Additionally, while many of the drug-evoked syn- 12. Nectow AR, Ekstrand MI, Friedman JM. Molecular characteriza-
aptic and structural plasticity changes described here are tion of neuronal cell types based on patterns of projection with
shared between different classes of abused drugs, others Retro-TRAP. Nat Protoc 2015; 10:1319-1327.
Reward Circuitry in Addiction 695

13. Ekstrand MI, Nectow AR, Knight ZA, Latcha KN, Pomeranz LE, 34. Stuber GD, Sparta DR, Stamatakis AM, et al. Excitatory transmis-
Friedman JM. Molecular profiling of neurons based on connectiv- sion from the amygdala to nucleus accumbens facilitates reward
ity. Cell 2014; 157:1230-1242. seeking. Nature 2011; 475:377-380.
14. van Zessen R, Phillips JL, Budygin EA, Stuber GD. Activation of 35. Ambroggi F, Ishikawa A, Fields HL, Nicola SM. Basolateral
VTA GABA neurons disrupts reward consumption. Neuron 2012; amygdala neurons facilitate reward-seeking behavior by exciting
73:1184-1194. nucleus accumbens neurons. Neuron 2008; 59:648-661.
15. Tan KR, Yvon C, Turiault M, et al. GABA neurons of the VTA 36. Janak PH, Tye KM. From circuits to behaviour in the amygdala.
drive conditioned place aversion. Neuron 2012; 73:1173-1183. Nature 2015; 517:284-292.
16. Gerfen CR, Engber TM, Mahan LC, et al. D1 and D2 dopamine 37. Zhu Y, Wienecke CF, Nachtrab G, Chen X. A thalamic input to the
receptor-regulated gene expression of striatonigral and nucleus accumbens mediates opiate dependence. Nature 2016;
striatopallidal neurons. Science 1990; 250:1429-1432. 530:219-222.
17. Gerfen CR, Surmeier DJ. Modulation of striatal projection sys- 38. Browning JR, Jansen HT, Sorg BA. Inactivation of the
tems by dopamine. Annu Rev Neurosci 2011; 34:441-466. paraventricular thalamus abolishes the expression of cocaine con-
18. Kupchik YM, Brown RM, Heinsbroek JA, Lobo MK, Schwartz ditioned place preference in rats. Drug Alcohol Depend 2014;
DJ, Kalivas PW. Coding the direct/indirect pathways by D1 and 134:387-390.
D2 receptors is not valid for accumbens projections. Nat Neurosci 39. James MH, Charnley JL, Flynn JR, Smith DW, Dayas CV.
2015; 18:1230-1232. Propensity to 'relapse' following exposure to cocaine cues is asso-
19. Smith RJ, Lobo MK, Spencer S, Kalivas PW. Cocaine-induced ciated with the recruitment of specific thalamic and epithalamic
adaptations in D1 and D2 accumbens projection neurons (a dichot- nuclei. Neuroscience 2011; 199:235-242.
omy not necessarily synonymous with direct and indirect path- 40. James MH, Charnley JL, Jones E, et al. Cocaine- and
ways). Curr Opin Neurobiol 2013; 23:546-552. amphetamine-regulated transcript (CART) signaling within the
20. Lenz JD, Lobo MK. Optogenetic insights into striatal function and paraventricular thalamus modulates cocaine-seeking behaviour.
behavior. Behav Brain Res 2013; 255:44-54. PLOS ONE 2010; 5:e12980.
21. Saddoris MP, Sugam JA, Cacciapaglia F, Carelli RM. Rapid do- 41. Joffe ME, Grueter BA. Cocaine experience enhances thalamo-
pamine dynamics in the accumbens core and shell: learning and accumbens N-methyl-D-aspartate receptor function. Biol
action. Front Biosci (Elite Ed) 2013; 5:273-288. Psychiatry 2016; 80:671-681.
42. Neumann PA, Wang Y, Yan Y, et al. Cocaine-induced synaptic
22. Meredith GE, Baldo BA, Andrezjewski ME, Kelley AE. The
alterations in thalamus to nucleus accumbens projection.
structural basis for mapping behavior onto the ventral striatum
Neuropsychopharmacology 2016; 41:2399-2410.
and its subdivisions. Brain Struct Funct 2008; 213:17-27.
43. Luscher C. The emergence of a circuit model for addiction. Annu
23. Sesack SR, Grace AA. Cortico-basal ganglia reward network:
Rev Neurosci 2016; 39:257-276.
microcircuitry. Neuropsychopharmacology 2010; 35:27-47.
44. Golden SA, Russo SJ. Mechanisms of psychostimulant-induced
24. Floresco SB. The nucleus accumbens: an interface between cog-
structural plasticity. Cold Spring Harb Perspect Med 2012; 2.
nition, emotion, and action. Annu Rev Psychol 2015; 66:25-52.
45. Luscher C, Malenka RC. Drug-evoked synaptic plasticity in ad-
25. Kalivas PW, Volkow N, Seamans J. Unmanageable motivation in diction: from molecular changes to circuit remodeling. Neuron
addiction: a pathology in prefrontal-accumbens glutamate trans- 2011; 69:650-663.
mission. Neuron 2005; 45:647-650.
46. Russo SJ, Mazei-Robison MS, Ables JL, Nestler EJ. Neurotrophic
26. Ma YY, Lee BR, Wang X, et al. Bidirectional modulation of in- factors and structural plasticity in addiction. Neuropharmacology
cubation of cocaine craving by silent synapse-based remodeling of 2009; 56(Suppl. 1):73-82.
prefrontal cortex to accumbens projections. Neuron 2014; 83: 47. Graham DL, Edwards S, Bachtell RK, DiLeone RJ, Rios M, Self
1453-1467. DW. Dynamic BDNF activity in nucleus accumbens with cocaine
27. Scofield MD, Heinsbroek JA, Gipson CD, et al. The nucleus ac- use increases self-administration and relapse. Nat Neurosci 2007;
cumbens: mechanisms of addiction across drug classes reflect the 10:1029-1037.
importance of glutamate homeostasis. Pharmacol Rev 2016; 68: 48. Graham DL, Krishnan V, Larson EB, et al. Tropomyosin-related
816-871. kinase B in the mesolimbic dopamine system: region-specific ef-
28. Oleskevich S, Descarries L, Lacaille JC. Quantified distribution of fects on cocaine reward. Biol Psychiatry 2009; 65:696-701.
the noradrenaline innervation in the hippocampus of adult rat. J 49. Grimm JW, Lu L, Hayashi T, Hope BT, Su TP, Shaham Y. Time-
Neurosci 1989; 9:3803-3815. dependent increases in brain-derived neurotrophic factor protein
29. French SJ, Totterdell S. Individual nucleus accumbens-projection levels within the mesolimbic dopamine system after withdrawal
neurons receive both basolateral amygdala and ventral subicular from cocaine: implications for incubation of cocaine craving. J
afferents in rats. Neuroscience 2003; 119:19-31. Neurosci 2003; 23:742-747.
30. French SJ, Hailstone JC, Totterdell S. Basolateral amygdala effer- 50. Le Foll B, Diaz J, Sokoloff P. A single cocaine exposure increases
ents to the ventral subiculum preferentially innervate pyramidal BDNF and D3 receptor expression: implications for drug-condi-
cell dendritic spines. Brain Res 2003; 981:160-167. tioning. Neuroreport 2005; 16:175-178.
31. Pascoli V, Terrier J, Espallergues J, Valjent E, O'Connor EC, 51. Fumagalli F, Moro F, Caffino L, et al. Region-specific effects on
Luscher C. Contrasting forms of cocaine-evoked plasticity control BDNF expression after contingent or non-contingent cocaine i.v.
components of relapse. Nature 2014; 509:459-464. self-administration in rats. Int J Neuropsychopharmacol 2013; 16:
32. Vezina P, Giovino AA, Wise RA, Stewart J. Environment-specific 913-918.
cross-sensitization between the locomotor activating effects of 52. McGinty JF, Whitfield TW, Jr., Berglind WJ. Brain-derived neu-
morphine and amphetamine. Pharmacol Biochem Behav 1989; rotrophic factor and cocaine addiction. Brain Res 2010; 1314:183-
32:581-584. 193.
33. Charara A, Grace AA. Dopamine receptor subtypes selectively 53. Horger BA, Iyasere CA, Berhow MT, Messer CJ, Nestler EJ,
modulate excitatory afferents from the hippocampus and amygda- Taylor JR. Enhancement of locomotor activity and conditioned
la to rat nucleus accumbens neurons. Neuropsychopharmacology reward to cocaine by brain-derived neurotrophic factor. J
2003; 28:1412-1421. Neurosci 1999; 19:4110-4122.
696 Cooper et al.

54. Lu L, Grimm JW, Hope BT, Shaham Y. Incubation of cocaine 73. Huang YH, Schluter OM, Dong Y. Silent synapses speak up:
craving after withdrawal: a review of preclinical data. updates of the neural rejuvenation hypothesis of drug addiction.
Neuropharmacology 2004; 47(Suppl. 1):214-226. Neuroscientist 2015; 21:451-459.
55. Li X, DeJoseph MR, Urban JH, et al. Different roles of BDNF in 74. Lee BR, Ma YY, Huang YH, et al. Maturation of silent synapses in
nucleus accumbens core versus shell during the incubation of cue- amygdala-accumbens projection contributes to incubation of co-
induced cocaine craving and its long-term maintenance. J caine craving. Nat Neurosci 2013; 16:1644-1651.
Neurosci 2013; 33:1130-1142. 75. Volkow ND, Morales M. The brain on drugs: from reward to
56. Bahi A, Boyer F, Chandrasekar V, Dreyer JL. Role of accumbens addiction. Cell 2015; 162:712-725.
BDNF and TrkB in cocaine-induced psychomotor sensitization, 76. Deisseroth K. Optogenetics. Nat Methods 2011; 8:26-29.
conditioned-place preference, and reinstatement in rats. 77. Stuber GD, Britt JP, Bonci A. Optogenetic modulation of neural
Psychopharmacology (Berl) 2008; 199:169-182. circuits that underlie reward seeking. Biol Psychiatry 2012; 71:
57. Whitfield TW, Jr., Shi X, Sun WL, McGinty JF. The suppressive 1061-1067.
effect of an intra-prefrontal cortical infusion of BDNF on cocaine- 78. Adamantidis AR, Tsai HC, Boutrel B, et al. Optogenetic interro-
seeking is Trk receptor and extracellular signal-regulated protein gation of dopaminergic modulation of the multiple phases of
kinase mitogen-activated protein kinase dependent. J Neurosci reward-seeking behavior. J Neurosci 2011; 31:10829-10835.
2011; 31:834-842. 79. Wolf ME. Synaptic mechanisms underlying persistent cocaine
58. Berglind WJ, See RE, Fuchs RA, et al. A BDNF infusion into the craving. Nat Rev Neurosci 2016; 17:351-365.
medial prefrontal cortex suppresses cocaine seeking in rats. Eur J 80. Chen BT, Yau HJ, Hatch C, et al. Rescuing cocaine-induced pre-
Neurosci 2007; 26:757-766. frontal cortex hypoactivity prevents compulsive cocaine seeking.
Nature 2013; 496:359-362.
59. Berglind WJ, Whitfield TW, Jr., LaLumiere RT, Kalivas PW,
81. Lobo MK, Covington HE, 3rd, Chaudhury D, et al. Cell type-
McGinty JF. A single intra-PFC infusion of BDNF prevents
specific loss of BDNF signaling mimics optogenetic control of
cocaine-induced alterations in extracellular glutamate within the
cocaine reward. Science 2010; 330:385-390.
nucleus accumbens. J Neurosci 2009; 29:3715-3719.
82. Creed M, Ntamati NR, Chandra R, Lobo MK, Luscher C.
60. Sadri-Vakili G, Kumaresan V, Schmidt HD, et al. Cocaine-induced Convergence of reinforcing and anhedonic cocaine effects in the
chromatin remodeling increases brain-derived neurotrophic factor ventral pallidum. Neuron 2016; 92:214-226.
transcription in the rat medial prefrontal cortex, which alters the 83. Heinsbroek JA, Neuhofer DN, Griffin WC, 3rd, et al. Loss of
reinforcing efficacy of cocaine. J Neurosci 2010; 30:11735-11744.
plasticity in the D2-accumbens pallidal pathway promotes cocaine
61. Russo SJ, Dietz DM, Dumitriu D, Morrison JH, Malenka RC, seeking. J Neurosci 2017; 37:757-767.
Nestler EJ. The addicted synapse: mechanisms of synaptic and 84. Wang L, Shen M, Yu Y, et al. Optogenetic activation of
structural plasticity in nucleus accumbens. Trends Neurosci GABAergic neurons in the nucleus accumbens decreases the ac-
2010; 33:267-276. tivity of the ventral pallidum and the expression of cocaine-
62. Verheij MM, Vendruscolo LF, Caffino L, et al. Systemic delivery context-associated memory. Int J Neuropsychopharmacol 2014;
of a brain-penetrant TrkB antagonist reduces cocaine self- 17:753-763.
administration and normalizes TrkB signaling in the nucleus ac- 85. James MH, Aston-Jones G. The ventral pallidum: proposed inte-
cumbens and prefrontal cortex. J Neurosci 2016; 36:8149-8159. grator of positive and negative factors in cocaine abuse. Neuron
63. Dietz DM, Dietz KC, Nestler EJ, Russo SJ. Molecular mecha- 2016; 92:5-8.
nisms of psychostimulant-induced structural plasticity. 86. Stefanik MT, Kupchik YM, Brown RM, Kalivas PW. Optogenetic
Pharmacopsychiatry 2009; 42(Suppl. 1):S69-78. evidence that pallidal projections, not nigral projections, from the
64. Tada T, Sheng M. Molecular mechanisms of dendritic spine mor- nucleus accumbens core are necessary for reinstating cocaine
phogenesis. Curr Opin Neurobiol 2006; 16:95-101. seeking. J Neurosci 2013; 33:13654-13662.
65. Dong Y, Nestler EJ. The neural rejuvenation hypothesis of cocaine 87. Root DH, Melendez RI, Zaborszky L, Napier TC. The ventral
addiction. Trends Pharmacol Sci 2014; 35:374-383. pallidum: subregion-specific functional anatomy and roles in mo-
66. Shen HW, Toda S, Moussawi K, Bouknight A, Zahm DS, Kalivas tivated behaviors. Prog Neurobiol 2015; 130:29-70.
PW. Altered dendritic spine plasticity in cocaine-withdrawn rats. J 88. Kalivas PW, McFarland K. Brain circuitry and the reinstatement
Neurosci 2009; 29:2876-2884. of cocaine-seeking behavior. Psychopharmacology (Berl) 2003;
67. Dietz DM, Sun H, Lobo MK, et al. Rac1 is essential in cocaine- 168:44-56.
induced structural plasticity of nucleus accumbens neurons. Nat 89. Kalivas PW. The glutamate homeostasis hypothesis of addiction.
Neurosci 2012; 15:891-896. Nat Rev Neurosci 2009; 10:561-572.
68. Pascoli V, Turiault M, Luscher C. Reversal of cocaine-evoked 90. McGlinchey EM, James MH, Mahler SV, Pantazis C, Aston-Jones
synaptic potentiation resets drug-induced adaptive behaviour. G. Prelimbic to accumbens core pathway is recruited in a
Nature 2011; 481:71-75. dopamine-dependent manner to drive cued reinstatement of co-
caine seeking. J Neurosci 2016; 36:8700-8711.
69. Gipson CD, Kupchik YM, Shen H, Reissner KJ, Thomas CA,
91. LaLumiere RT, Smith KC, Kalivas PW. Neural circuit competition
Kalivas PW. Relapse induced by cues predicting cocaine depends
in cocaine-seeking: roles of the infralimbic cortex and nucleus
on rapid, transient synaptic potentiation. Neuron. 2013; 77:867-
accumbens shell. Eur J Neurosci 2012; 35:614-622.
872.
92. Peters J, LaLumiere RT, Kalivas PW. Infralimbic prefrontal cortex
70. Cahill ME, Bagot RC, Gancarz AM, et al. Bidirectional synaptic
is responsible for inhibiting cocaine seeking in extinguished rats. J
structural plasticity after chronic cocaine administration occurs
Neurosci 2008; 28:6046-6053.
through Rap1 small GTPase signaling. Neuron 2016; 89:566-582.
93. Moorman DE, James MH, McGlinchey EM, Aston-Jones G.
71. Luscher C, Pascoli V, Creed M. Optogenetic dissection of neural Differential roles of medial prefrontal subregions in the regulation
circuitry: from synaptic causalities to blue prints for novel treat- of drug seeking. Brain Res 2015; 1628:130-146.
ments of behavioral diseases. Curr Opin Neurobiol 2015; 35:95- 94. Hanlon CA, Dowdle LT, Jones JL. Biomarkers for success: using
100. neuroimaging to predict relapse and develop brain stimulation
72. Huang YH, Lin Y, Mu P, et al. In vivo cocaine experience gener- treatments for cocaine-dependent individuals. Int Rev Neurobiol
ates silent synapses. Neuron 2009; 63:40-47. 2016; 129:125-156.
Reward Circuitry in Addiction 697

95. Volkow ND, Wang GJ, Fowler JS, Tomasi D. Addiction circuitry 104. Pierce RC, Vassoler FM. Deep brain stimulation for the treatment
in the human brain. Annu Rev Pharmacol Toxicol 2012; 52:321- of addiction: basic and clinical studies and potential mechanisms
336. of action. Psychopharmacology (Berl) 2013; 229:487-491.
96. Janicak PG, Dokucu ME. Transcranial magnetic stimulation for 105. Vassoler FM, Schmidt HD, Gerard ME, et al. Deep brain stimula-
the treatment of major depression. Neuropsychiatr Dis Treat 2015; tion of the nucleus accumbens shell attenuates cocaine priming-
11:1549-1560. induced reinstatement of drug seeking in rats. J Neurosci 2008; 28:
97. Gorelick DA, Zangen A, George MS. Transcranial magnetic stim- 8735-8739.
ulation in the treatment of substance addiction. Ann N Y Acad Sci 106. Vassoler FM, White SL, Hopkins TJ, et al. Deep brain stimulation
2014; 1327:79-93. of the nucleus accumbens shell attenuates cocaine reinstatement
98. Speer AM, Kimbrell TA, Wassermann EM, et al. Opposite effects through local and antidromic activation. J Neurosci 2013; 33:
of high and low frequency rTMS on regional brain activity in 14446-14454.
depressed patients. Biol Psychiatry 2000; 48:1133-1141. 107. Creed M, Pascoli VJ, Luscher C. Addiction therapy. Refining deep
99. Fitzgerald PB, Fountain S, Daskalakis ZJ. A comprehensive re- brain stimulation to emulate optogenetic treatment of synaptic
view of the effects of rTMS on motor cortical excitability and pathology. Science 2015; 347:659-664.
inhibition. Clin Neurophysiol 2006; 117:2584-2596. 108. Muller UJ, Sturm V, Voges J, et al. Nucleus accumbens deep brain
stimulation for alcohol addiction—safety and clinical long-term
100. Terraneo A, Leggio L, Saladini M, Ermani M, Bonci A,
results of a pilot trial. Pharmacopsychiatry 2016; 49:170-173.
Gallimberti L. Transcranial magnetic stimulation of dorsolateral
109. Kuhn J, Moller M, Treppmann JF, et al. Deep brain stimulation of
prefrontal cortex reduces cocaine use: A pilot study. Eur
the nucleus accumbens and its usefulness in severe opioid addic-
Neuropsychopharmacol 2016; 26:37-44.
tion. Mol Psychiatry 2014; 19:145-146.
101. Bolloni C, Panella R, Pedetti M, et al. Bilateral transcranial mag- 110. Goncalves-Ferreira A, do Couto FS, Rainha Campos A, Lucas
netic stimulation of the prefrontal cortex reduces cocaine intake: a Neto LP, Goncalves-Ferreira D, Teixeira J. Deep brain stimulation
pilot study. Front Psychiatry 2016; 7:133. for refractory cocaine dependence. Biol Psychiatry 2016; 79:e87-
102. Hanlon CA, Dowdle LT, Austelle CW, et al. What goes up, can e89.
come down: Novel brain stimulation paradigms may attenuate 111. Graziane NM, Sun S, Wright WJ, et al. Opposing mechanisms
craving and craving-related neural circuitry in substance depen- mediate morphine- and cocaine-induced generation of silent syn-
dent individuals. Brain Res 2015; 1628:199-209. apses. Nat Neurosci 2016; 19:915-925.
103. Holtzheimer PE, Mayberg HS. Deep brain stimulation for psychi- 112. Koo JW, Mazei-Robison MS, Chaudhury D, et al. BDNF is a
atric disorders. Annu Rev Neurosci 2011; 34:289-307. negative modulator of morphine action. Science 2012; 338:124-
128.