Neurobiology of craving, conditioned reward and relapse

Friedbert Weiss
Chronic vulnerability to relapse is a formidable challenge
for the treatment of drug addiction. The neurobiological
basis of relapse and its prevention has, therefore, attracted
major attention in addiction research. Current
conceptualizations of addiction recognize craving as a central
driving force for ongoing drug use, as well as for relapse
following abstinence. Progress has been made in
understanding experiential factors, neurocircuitry components
and signaling mechanisms that mediate conditioned
drug-seeking behaviour, craving and long-lasting susceptibility
to relapse. Importantly, stress contributes to drug craving,
and there is evidence for overlap between the neural and
neuroendocrine mechanisms implicated in drug desire evoked
by drug cues and stress. Recent research has substantially
advanced our understanding of the neurobiological factors
responsible for drug craving and relapse, with promising
therapeutic implications.
Addresses
The Scripps Research Institute, Department of Neuropharmacology
(CVN-15), 10550 North Torrey Pines Road, La Jolla, CA 92037, USA
Corresponding author: Weiss, F (bweiss@scripps.edu)

Current Opinion in Pharmacology 2005, 5:919

This review comes from a themed issue on
Neurosciences
Edited by Graeme Henderson, Hilary Little and Jenny Morton
Available online 21st December 2004
1471-4892/$ see front matter
# 2005 Elsevier Ltd. All rights reserved.
DOI 10.1016/j.coph.2004.11.001

Abbreviations
BDNF
brain-derived neurotrophic factor
BLA
basolateral amygdala
CRF
corticotropin-releasing factor
CS
conditioned stimuli
HPA
hypothalamic-pituitary-adrenal
LH
lateral hypothalamus
MAPK
mitogen-activated protein kinase
mGluR metabotropic glutamate receptor
NAc
nucleus accumbens
N/OFQ nociceptin/orphanin FQ
OFC
orbitofrontal cortex
PFC
prefrontal cortex
PLC
prelimbic cortex
VTA
ventral tegmental area

Introduction
Relapse is a phenomenon pivotal for the understanding
and treatment of drug addiction. The focus of addiction
www.sciencedirect.com

research has, therefore, shifted towards identifying the

experiential and neurobiological mechanisms responsible
for the chronically relapsing nature of addiction. Major
precipitating factors for relapse include drug craving and
stress [1,2], with an increasing recognition of cognitive
factors such as pre-attentive automatic reactions, and
attentional bias related to previous drug experiences
[3,4,5]. These risk factors are further aggravated by
protracted withdrawal symptoms resulting from druginduced neuroadaptation, but also by conditions such
as psychiatric comorbidity, socioeconomic conditions
and perceived drug availability [1,6,7].
Here, recent advances in both the neurocircuitry and the
neurochemical and molecular bases of craving and relapse
are reviewed, with emphasis on two fundamental questions: firstly, what factors are responsible for the distinctly
compulsive nature of drug-seeking and craving, as
opposed to behaviour motivated by natural rewards
essential for survival and well-being; and secondly, what
long-term changes develop in the brain during chronic
drug use that maintain craving and relapse risk through
years of abstinence and rapidly re-establish fullblown dependence after resumption of drug use? This
review focuses predominantly on cocaine and ethanol as
examples.

Associative learning as a factor in craving

and relapse
Critical for craving and relapse is the process of associative
learning, whereby environmental stimuli repeatedly
paired with drug consumption acquire incentive-motivational value, evoking expectation of drug availability and
memories of past drug euphoria [810]. Conditioned
responses to such stimuli can activate brain reward
mechanisms [11] and have been implicated both in
maintaining ongoing drug use and in eliciting drug desire
during abstinence, precipitating relapse. The definition
of craving and its measurement, as well as the predictive
link between craving and relapse, is still a matter of some
debate [12,13,14]. Moreover, as a hypothetical construct, craving is difficult to model with functional equivalence in animals [14]. However, reactivity to drug cues
has established significance for craving in clinical settings.
These conditioned responses represent a mechanism
leading to craving, and can therefore be exploited to
study this specific aspect of craving in animals [8].
Functional brain imaging in humans ([1517]; see review
by Lingford-Hughes, this issue) as well as lesion and
site-specific pharmacological manipulations in animals
[9,10,18] have implicated an interconnected set of cortical
Current Opinion in Pharmacology 2005, 5:919

10 Neurosciences

and limbic brain regions in associative learning underlying craving and relapse (see Table 1). Major components of this circuitry include the orbitofrontal cortex
(OFC), anterior cingulate, prelimbic cortex (PLC), basolateral amygdala (BLA), hippocampus, nucleus accumbens (NAc) and, more recently, the dorsal striatum, which
is thought to participate in consolidating stimulusresponse habits via the engagement of corticostriatal
loops [10].
Substantial advances have been made in elucidating key
components and signaling mechanisms within this circuitry that regulate specific aspects of drug-seeking using
animal models of conditioned reinforcement and conditioned reinstatement (see Boxes 13). The NAc, a functionally heterogeneous structure consisting of core and
shell subregions (for review, see [23]), has long been
considered critical for drug-seeking and reinforcement.
Existing evidence favours a differential role for these
subregions in reinstatement and second-order schedule
performance maintained by discrete conditioned stimuli
(CS) versus reinstatement elicited by contextual stimuli.
Selective inactivation of the NAc core, but not shell,
abolished cocaine CS-induced reinstatement [24], and
blockade of glutamate receptors in the NAc core eliminated conditioned reinforcement by cocaine CS on a
second order schedule [25]. Moreover, functional integrity of the core is required for the acquisition of conditioned cocaine reinforcement, whereas the shell appears
essential for the psychomotor stimulant effects of cocaine
[26]. Furthermore, the glutamatergic projection from
the BLA to the NAc core was found to be critical for the
expression of cocaine CS-maintained second order schedule performance [25,27]. Novel data also implicate the
dopaminergic projection from the ventral tegmental area
(VTA) to the NAc core in associative processes that
influence conditioned responding under second order
schedules [28]; however, in contradiction to the effects
of dopamine receptor blockade in the BLA [25], inactivation of the dopaminergic projection from the VTA to
the BLA failed to interfere with the conditioned reinforcing effects of cocaine CS [28]. Thus, dopaminergic
terminals in the NAc core and the BLA might differentially regulate CS-maintained cocaine reinforcement
[28], the behavioural significance of which awaits clarification.
In contrast to CS-maintained conditioned reinforcement
and reinstatement, recovery of extinguished drug-seeking by contextual stimuli is dependent upon the NAc
shell. During reinstatement evoked by a cocaine-predictive contextual stimulus, neurons in the shell exhibited
significantly greater activity than those in the core, the
latter responding indiscriminately to the drug cue or
neutral stimulus [29]. Shell neuron responses to the drug
cue were still evident after four weeks of abstinence,
illustrating that the neural representation of the motivaCurrent Opinion in Pharmacology 2005, 5:919

tional significance of these stimuli is highly resistant to

decay [29]. Importantly, inactivation of the dorsomedial
prefrontal cortex (PFC) and BLA impaired reinstatement
by cocaine CS [30,31], whereas inactivation of the dorsal
hippocampus was selective for contextual reinstatement
[30]. However, recent data also implicate the BLA in
contextual reinstatement [30,32,33] through its rostral
subdivision [34]. Thus, input from the rostral BLA to the
NAc shell is a neural link essential for encoding the
motivational value of drug-associated contextual stimuli
[29,35]. Overall, neural substrates for reinstatement associated with contextual and explicit CS overlap, but also
show distinct differences.
New understanding has accrued on the function of the
OFC and PLC in drug-related learning. The OFC has
been found to modulate cocaine-seeking in a subregionspecific manner, with CS-maintained reinstatement controlled by the lateral OFC and drug-induced (i.e. priming) reinstatement by the medial OFC [31]. In the
PLC, rats expressing conditioned preference for a
cocaine-paired environment (i.e. a contextual stimulus)
showed increased activation of inhibitory GABAergic
interneurons, paired with decreased excitatory output
from this site [36]. The decreased activation of excitatory
PLC efferents may add to the functional defects in
prefrontal cortical regions that develop with chronic
cocaine use [3740]. Considering the executive function of the PFC that includes risk-benefit analysis and
suppression of limbic impulses, the shift in activation
patterns of PLC cell populations could contribute to
impaired impulse control, and lack of judgement and
risk assessment [41], which are defining features of
addiction.

Determinants of craving and relapse as

persistent and compulsive phenomena
Clinically, a slip into limited drug use is unimportant for
ultimate therapeutic success, such that modeling craving
and reinstatement alone is not sufficient for investigating
the chronic and compulsive nature of addiction [8]. Evidence has accumulated that drug-seeking in animals
resembles along several dimensions the compulsive nature of addiction in humans. Reinstatement by drug cues is
resistant to extinction and persists over months of abstinence [42,43,44]. Even stimuli present during a single
cocaine experience elicited drug-seeking for up to one
year [45]. Importantly, the intensity of conditioned
reinstatement increases during the initial months of
abstinence [42,43,46,47]. The behavioural significance
of this incubation effect [48] was scrutinized by isolating
the respective contribution to reinstatement of simply
terminating access to cocaine versus subjecting rats to
extinction of cocaine-reinforced responding with, or without, presentation of drug-associated CS. Extinction
substantially decreased reinstatement (as measured by
CS-maintained responding on a second order schedule)
www.sciencedirect.com

Neurobiology of craving, conditioned reward and relapse Weiss 11

Table 1
Description of the major function of brain regions covered in this review, and their role in drug addiction, craving and relapse (italicized).
Bain region

Function

Prefrontal cortex

Executive and decision-making functions based on the recognition of survival-related challenges, and
execution of goal-directed actions.
Associative learning linked to rewarding and aversive stimuli. Integration of emotion and natural drive states
with behaviour. Assessment of reward value against previous experience influencing action and choice
- CS-based reinstatement
- Conditioned reinforcement
- Stress-induced reinstatement
- Cue-induced craving (human functional brain imaging)
Processing of pleasure and pain. Attention and cognitively demanding information processing.
Conditioned emotional learning, including attribution of emotional and motivational value to internal
and external stimuli guiding appropriate response selection.
Regulation of autonomic and endocrine function
- Conditioned reinforcement
- CS-based reinstatement
- Contextual reinstatement
- Cue-induced craving (human functional brain imaging)
Attention, working memory, detection/evaluation of actionoutcome relationships, response initiation and inhibition.
Regulation of autonomic and endocrine function
- Conditioned reinforcement
- CS-based reinstatement
- Contextual reinstatement
- Stress-induced reinstatement
- Putative final common pathway at the cortical level for stress-, cue- and drug-induced craving and reinstatement
- Cue-induced craving (human functional brain imaging)
Processing of emotionally significant stimuli guiding conditioned and unconditioned approach or avoidance behaviour
- Conditioned reinforcement
- CS-maintained reinstatement
- Contextual reinstatement
- Cue-induced craving (human functional brain imaging)
Contextual learning and memory consolidation. Retrieval of episodic memories.
Conditioning of contextual stimuli to fear or reward
- Contextual reinstatement
Limbic-motor interface, integrating converging input from limbic sites related to appetitive and activational
aspects of rewards for output to effector sites, initiating and sustaining behavioural responses.
Participation in the translation of motivation into action.
NAc core - Conditioned reinforcement
- CS-based reinstatement
NAc shell - Contextual reinstatement
- Molecular neuroadaptations following chronic cocaine with implications for craving and relapse
- Cue-induced craving (human functional brain imaging)
Part of the extrapyramidal motor system. Fine-tuning motoric functions.
Consolidation of stimulus-response habits via corticostriatal loops
- Presumed role in the switch from action to habit (i.e. the transition from drug abuse to drug addiction)
Origin of the mesocorticolimbic dopamine projection
- Site implicated in Group II mGluR-mediated contextual reinstatement
- Molecular neuroadaptations following chronic cocaine with implications for craving and relapse
Regulation of ingestive behaviour
Sensitive site for the rewarding effects of electrical brain stimulation reward and changes in brain reward
function as measured by intracranial self-stimulation reward thresholds
- Increased in brain reward function by exposure to cocaine cues
- Impaired brain reward function associated with escalated cocaine intake
- Molecular neuroadaptations linked to escalated cocaine intake
Processing of behavioural and physiological responses to stress. Key component of the extrahypothalamic
CRF stress system reciprocally connected to the BNST
Final common pathway processing stress and anxiety responses. Key component of the extrahypothalamic
stress system. The BNST also projects heavily to the PVN, modulating HPA axis activity
CRF-rich nucleus and apex of the HPA stress axis. CRF is the major stress-regulatory molecule in the brain
and integrates behavioural, endocrine and autonomic responses to stress
Release of CRF from PVN neurons projecting to the anterior pituitary results in secretion of ACTH (i.e. activation
of the HPA axis) leading to secretion of adrenal glucocorticoids, predominantly corticosterone. Stress-related
addictive behaviour is thought to be related to activation of the HPA axis by CRF
- Drug cue exposure leads to HPA axis activation correlated with craving in cocaine- or
ethanol-addicted individuals
- Inteference with HPA axis activation prevents cue-induced reinstatement

Orbitofrontal cortex

Anterior cingulate cortex

Prelimbic cortex

Basolateral amygdala

Hippocampus

Nucleus accumbens

Dorsal striatum

Ventral tegmental area

Lateral hypothalamus

Central nucleus
of the amygdala
Bed nucleus of the
stria terminalis (BNST)
Paraventricular nucleus
of the hypothalamus
Hypothalamicpituitary-adrenal axis

www.sciencedirect.com

Current Opinion in Pharmacology 2005, 5:919

12 Neurosciences

Box 1 Animal models of craving and relapse (1): CS-based

extinction-reinstatement model.

Box 3 Animal models of craving and relapse (3): contextual

extinction-reinstatement model.

The most widely employed method to model craving and relapse

(for review, see [19]). In this procedure, animals are trained to
respond at a lever or other operandum where completion of each
response requirement (i.e. one or more responses, depending on
the experimental objective) results in delivery of a small intravenous
(in the case of opioids or psychostimulants) or oral (in the case of
ethanol) dose of the drug. Each response producing a drug injection
is paired with brief (e.g. 520 s) presentation of one or more
environmental stimuli (e.g. a tone, cue light or both), referred to as
CS. Thus, both drug administration and presentation of the CS are
contingent upon a response by the animal (i.e. responsecontingent). Once reliable drug self-administration is acquired,
drug-reinforced responding (technically referred to as operant or
instrumental responding) is extinguished by ceasing to reinforce
responses by drug delivery and withholding presentation of the CS.
In contemporary applications of this procedure, extinction sessions
are conducted daily until responding ceases (i.e. is substantially
lower than that maintained by delivery of the drug and deemed
incidental rather than motivated by expectation of obtaining the
drug). Subsequently, reinstatement tests are conducted in which
the degree of recovery of responding at the previously drug-paired
operandum, now maintained by response-contingent presentation of
the conditioned CS only (i.e. without further drug delivery), is
operationally defined as a measure of craving or relapse or, more
generally, a measure of drug-seeking.

This model is identical in most aspects to the CS-based

reinstatement model, except that it utilizes contextual stimuli (i.e.
ambient olfactory, auditory, tactile or visual stimuli in the drug
self-administration environment) for conditioning that are not
discretely paired with drug infusions nor contingent upon a
response. Multiple contextual reinstatement procedures are
employed. The most prevalent contextual reinstatement model
utilizes cues referred to as discriminative stimuli where, during
learning, responses at the drug operandum are reinforced only in
the presence of these stimuli. Owing to their predictive nature for
drug availability, these stimuli set the occasion for engaging in
reward-seeking behavior (i.e. lead to the initiation of reponding) but
additionally, by virtue of their presence during drug consumption,
also become associated with the rewarding effects of the drug and
thus acquire incentive-motivational value (i.e. elicit memories of
previous drug euphoria and the magnitude or value of the
rewarding effect of drug consumption. Because of this dual
property, these stimuli are particularly powerful in eliciting drugseeking and reinstatement. Another contextual reinstatement
procedure is based on learning of a conditioned place preference
(CPP) for an environment paired with experience of the drug.
Following extinction of CPP, the re-establishment (technically
termed renewal) of CPP can be operationalized to reflect a
measure of relapse and craving. Non-contingent exposure to
contextual cues readily elicits reinstatement, whereas this is not the
case with non-contingent CS presentation in CS-based
reinstatement models [21,22]. Once initiated, however, both types of
stimuli maintain responding at a previously active drug operandum
[21]. These two variants of associative learning probably occur
simultaneously during drug use but may be coded, at least in
part, through different neural circuitries or activation patterns [9].

compared with one week of abstinence alone. However,

after four weeks of abstinence, CS-maintained responding in the extinguished group was increased (i.e. showed
incubation), and was identical to rats not subjected to
extinction [46]. Consequently, eliminating drug availability following extinction might, with time, augment
the efficacy of drug cues to facilitate reinstatement,
diminishing the effects of extinction. The time since
cessation of drug use might therefore be an important
factor in determining the efficacy of cue extinction

Box 2 Animal models of craving and relapse (2): second order

schedule of reinforcement.
This is another widely employed model of craving. Animals are first
trained to self-administer a drug, response-contingently paired with
presentation of a CS. Animals then are subjected to a procedure in
which only the CS is available response-contingently until
completion of a predetermined response-requirement (e.g. 2060
CS-reinforced responses), which then leads to a single
administration of the drug reinforcer itself, whereupon the
contingency reverts again to the non-drug-reinforced schedule
component with response-contingent presentation of the CS only,
until completion of the response requirement leading to the next
drug injection. The principal measure of interest in this procedure is
the degree of conditioned reinforcement maintained by the
drug-paired CS, before drug delivery that can be operationally
defined, again, as a measure of drug craving. Although not a formal
equivalence model of relapse because of typically lacking extinction
and abstinence manipulations [14], second order schedules have
proven highly effective for investigating the significance of
conditioning factors in ongoing drug-seeking and taking (for review,
see [20]).

Current Opinion in Pharmacology 2005, 5:919

therapies. Drug-seeking in animals is also consistent with

defining features of addiction other than persistence.
Rats with a history of prolonged, but not limited, cocaine
self-administration failed to show suppression of CSmaintained drug-seeking when concurrently presented
with a conditioned stressor [44,49], in contrast with
the well-established inhibition of appetitive behaviour
by aversive stimuli (conditioned suppression). Thus,
drug-related cues in rats having experienced prolonged
cocaine exposure sustain drug-directed behaviour
despite adverse consequences, one of the hallmarks of
substance dependence. Rats prone to addiction on one
index of vulnerability (cocaine priming) also proved
more susceptible to other addiction-like behaviours,
including escalation of cocaine intake, conditioned reinstatement and maintenance of cocaine-seeking despite
increased work requirements [44]. Thus, the inflexible
dimension that characterizes addiction can readily be
demonstrated in animals but requires prolonged drug
access, confirming original findings that such exposure
regimens are critical for the transition from drug use to
dependence [50].
An important experiential factor (i.e. one related to personal experience) for the development of compulsive
drug-seeking is that consumption of a drug during
withdrawal, an event inextricably linked to addiction,
www.sciencedirect.com

Neurobiology of craving, conditioned reward and relapse Weiss 13

introduces learning about a previously latent incentive

dimension of the drug its capacity is to alleviate
physically or emotionally adverse states. This experience
can enhance the incentive value of the drug and, therefore, its reinforcing efficacy particularly during subsequent withdrawal states. Sound support for this
hypothesis exists in animals [51]. This mechanism might
also be relevant for cue-induced craving and relapse,
considering that drug consumption during withdrawal
is likely to increase not only the incentive value of the
drug but also the conditioned incentive value of drugrelated stimuli. Indeed, ethanol-dependent rats allowed
to self-administer alcohol during withdrawal, and presented with ethanol CS three weeks after withdrawal
[52], showed greater conditioned reinstatement, whereas
this was not the case in dependent rats not having
experienced ethanol during withdrawal [53,54]. The
enhanced incentive value of drugs and associated environmental stimuli through withdrawal-related learning
could generalize to other states of aversion and negative
affect, because not only the ethanol CS but also footshock
and a conditioned stressor exacerbated reinstatement in
rats having consumed ethanol while in withdrawal, but
not in rats without this drug history [52,55]. Such generalizations would dramatically contribute to the compulsive
nature of drug addiction.

Craving and seeking-behaviour associated

with drug versus natural rewards
To understand drug compulsion, we must recognize
differences that exist between the stimulus control of
behaviour motivated by natural versus drug rewards (see
chapter by Salamone et al., this issue). Prolonged exposure to cocaine, but not sucrose, eliminated conditioned
suppression of reward-seeking by a footshock-predictive
cue, suggesting that the nature of the reward (drug versus
natural) determines whether compulsive behaviour
resistant to adverse environmental events develops
[44,49]. Cocaine-associated contextual stimuli elicited
reinstatement for one year, whereas the same stimuli
paired with a palatable sweet solution remained effective
for only three months [45]. Conversely, incubation of
CS-induced reinstatement developed with stimuli conditioned to either cocaine or sucrose [42,47]. Thus, timedependent variations in CS-induced reinstatement of
reward-seeking might be common to drug and natural
rewards but, by extrapolation from data above [46], could
be driven by incubation of spontaneous recovery as
an underlying process. Further understanding of the
mechanisms contributing to the differential control of
normal appetitive versus compulsive drug-seeking
behaviour might be gleaned from selective effects of
pharmacological manipulations. A group II metabotropic
glutamate receptor (mGluR) agonist and nociceptin/
orphanin FQ (N/OFQ), a member of the opioid peptide
family, reversed conditioned reinstatement by ethanol- or
cocaine-associated contextual cues, without altering
www.sciencedirect.com

appetitive behaviour motivated by palatable sweet solutions [56,57].

Neurobiological basis of long-lasting

susceptibility to relapse
Central to the understanding and treatment of craving
and relapse are cellular and molecular neuroadaptations
that mediate enduring associations between drugs and
associated stimuli. Increased expression of brain-derived
neurotrophic factor (BDNF) is a probable mechanism for
the incubation of conditioned reinstatement [42], a
hypothesis supported by long-lasting potentiation of
reinstatement following intra-VTA application of BDNF
[58]. The role of BDNF in reinstatement appears to
depend critically upon recruitment of mitogen-activated
protein kinase (MAPK) signal transduction [58], a pathway involved in long-term potentiation and synaptic
plasticity, processes that have been implicated in development of addiction [59,60]. Withdrawal from prolonged
cocaine self-administration is also associated with lasting
increases in glutamate N-methyl-D-aspartate receptor 1
and GluR2 receptor subunits in the NAc and VTA, as
well as transient elevation of cyclin-dependent protein
kinase 5 in the VTA [61]. Increased BNDF expression
initiates a plethora of signaling cascades with implications for addiction (for review, see [62]), including
changes in synaptic strength, neural signaling and dendrite formation. Repeated cocaine increases dendritic
sprouting in the NAC, PFC and caudate-putamen that
persists for months [6365] structural changes with a
presumptive role in sensitization and the persistence of
craving and relapse risk. The activation of cyclin-dependent protein kinase 5, also linked to proliferation of
dendritic sprouting [66], might be a contributing factor.
Chronic cocaine (or morphine) also alters postreceptor
signaling cascades for BDNF in the meso-accumbens
pathway, one of which sustains MAPK activation during
prolonged cocaine abstinence [62]. Pharmacological
interference with MAPK signaling reduces preference
for a cocaine-paired environment [67], thus supporting
the hypothesis that increases in both BDNF and MAPK
levels during cocaine withdrawal represent molecular
neuroadaptations contributing to craving and relapse
associated with drug cue exposure. Molecular neuroadaptations in two other regulatory systems are also of
interest. The first cocaine-induced alteration in
cystine-glutamate exchange in the NAc regulates
basal extracellular glutamate levels, resulting in reduced
levels of this amino acid during cocaine withdrawal, an
effect implicated in susceptibility to relapse [68]. The
second upregulation of an activator of G protein
signaling showed significant elevation in the PFC
and NAc after only one week of cocaine exposure, and
was still evident after two months of abstinence [69].
Although a behavioural function for these neuroadaptations has been demonstrated only for cocaine-induced
reinstatement, these molecular changes are promising
Current Opinion in Pharmacology 2005, 5:919

14 Neurosciences

targets to explore with respect to a role in cue-elicited

craving and relapse.
Finally, an emerging neuropharmacological mechanism
in craving and relapse is the endocannabinoid system that
functionally interacts with numerous other brain neurotransmitter systems with established roles in drug-seeking and addiction [70]. The endocannabinoid system has
been implicated in opiate, cocaine and alcohol addition
[71,72]. More importantly, pharmacological manipulation
of this system modifies both heroin reinforcement and
conditioned reinstatement [73].

Different neurocircuitries for drug versus

natural reward?
Conditioned responses, whether motivated by drugs of
abuse or natural rewards, are thought to be processed via
the same neurocircuitry [18,74,75]. It is unclear, however,
what differentiates neural signaling regulating normal
appetitive behaviour versus pathological behaviour
resulting from drug-related conditioning. NAc neurons,
activated during cocaine-reinforced responding, were also
responsive to cocaine CS, whereas cell populations activated only by water reinforcement were not (reviewed in
[76]). Positron emission tomography imaging data of
craving in heroin addicts indicate that both drug and
non-drug stimuli activate brain attentional and memory
circuits (primarily the anterior cingulate and OFC) but
with stronger effects evoked by drug cues [77]. Brain
regions activated by drug cues, therefore, might not be
specific to addiction-related events but components of
normal circuits activated to a greater degree, and may
reflect either the creation of new motivational states such
as those produced by withdrawal-related learning [74] or
the tilting of processes that normally govern responding
for natural rewards towards drug-directed behaviour.
Differential activation of BDNF-induced signaling might
be a contributory factor, as increased BDNF expression is
linked to incubation of conditioned cocaine- but not
sucrose-seeking behaviour [42].

Dysregulation of hedonic homeostasis

A widely held position in addiction theory views craving
and relapse as an opponent motivational process to avoid
negative affect (e.g. dysphoria, anhedonia and anxiety),
which is a consequence of withdrawal from most drugs of
abuse. The most recent conceptualization of this process
focuses on the development of allostasis, defined as a
state of progressive deviation of motivational regulatory
systems by chronic drug use from their normal function,
with the establishment of a new hedonic set point (for
review, see [50,78]). Novel evidence shows that increased
cocaine intake (escalation) that develops with prolonged
drug exposure [50] leads to long-lasting counteradaptive
deficits in brain reward function which reduce the hedonic impact of cocaine [79]. Allostasis has also been linked
to dysregulation in molecular signaling systems. TranCurrent Opinion in Pharmacology 2005, 5:919

scriptional profiling by DNA microarray revealed that

escalated cocaine intake is associated with alterations in
transcripts relevant for neurite extension and synaptogenesis, cell adhesion and nerve regeneration, as well as for
neuronal excitability and glutamate transmission (cited as
unpublished in [50]). Increases in ionotropic and mGluR
subunits during withdrawal support such changes in glutamate transmission during prolonged cocaine exposure
[80]. Interestingly, gene transcript changes in cocaine
escalated rats were confined largely to the lateral hypothalamus (LH); therefore, synaptic rearrangement and altered
neural excitability in the LH could play a role in compulsive cocaine-seeking [50] (for related discussion, see [81]).

Stress as a source of craving

Stress, although complex and influenced by multiple
factors, has an established role in relapse [2]. Footshock
(as a model of stress) reinstates extinguished responding
previously maintained by drugs of abuse. Several theoretical views exist on the motivational mechanisms by
which stress provokes relapse as discussed in several
excellent reviews [82,83,84]. Of interest here is the
growing evidence that drug cue and stress manipulations
induce a similar pattern of craving and cue reactivity,
including comparable increases in negative affect and
anxiety [85,86], findings that support the allostatic view
of addiction. Exposure to drug cues in cocaine-dependent
subjects produced increases in both craving and corticotropin-releasing factor/hypothalamic-pituitary-adrenal
(CRF/HPA)-axis-dependent behavioural and physiological measures of stress [86]. Consistent with these findings, corticosterone synthesis inhibition and blockade of
CRF1 receptors in rats attenuated cocaine CS-induced
reinstatement (i.e. a measure of craving), indicative of a
role for stress (i.e. HPA axis activation) in the responsereinstating actions of drug cues [22]. These observations
implicate overlap between neural substrates controlling
craving evoked by drug cues and stress. Stress, via HPA
axis activation, increases mesocorticolimbic dopamine
transmission. Both aversive and appetitive events are
processed by interactions among limbic subcortical and
prefrontal cortical regions, including the BLA and OFC,
with ultimate effects on extrahypothalamic and neuroendocrine stress responses through direct and indirect connections with the bed nucleus of the stria terminalis,
central nucleus of the amygdala, and hypothalamic paraventricular nucleus (reviewed in [2]). Thus, activation of
both brain motivational circuits and stress-regulatory systems by drug cues and stress implicates common neural
substrates through which these challenges enhance drugseeking. Transient lesion and site-specific pharmacological manipulation of PFC regions have implicated the
PLC and OFC as components of a circuitry mediating
footshock-induced drug-seeking, with the PLC representing a possible common pathway for stress-, cueand drug-induced reinstatement [87]. Lastly, evidence
for overlap between the neural substrates mediating stress
www.sciencedirect.com

Neurobiology of craving, conditioned reward and relapse Weiss 15

and drug cue effects exists also at the pharmacological

level. Several agents that attenuate conditioned reinstatement [56,57] also have marked anti-stress and anxiolytic
profiles. These include Group II mGluR agonists [8890]
and N/OFQ, which exerts strong CRF antagonist actions in
the bed nucleus of the stria terminalis [91].

Interactions among risk factors for craving

and relapse
Although craving and stress are undoubtedly important
for relapse, a predictive relationship between subjective
reports of stress or craving and subsequent drug use is a
matter of some dispute [2,12,13,14]. Risk factors for
relapse are typically studied in isolation, whereas abstinent drug users are frequently exposed to multiple external risk factors, while experiencing varying degrees of
protracted withdrawal symptoms. Of relevance for this
issue, reinstatement of alcohol-seeking was found markedly enhanced when rats were subjected to footshock
stress before response-contingent exposure to ethanol CS
[52,55]. Both the individual and interactive effects of
stress and the ethanol cue were greatly augmented in
previously ethanol-dependent rats tested three weeks
after withdrawal, a time when behaviourally relevant
neuroadaptive changes in CRF and endogenous opioid
function are present [54,92]. It is therefore likely that the
probability of relapse varies as a function of the number
and intensity of risk factors operative at any given time,
with relapse occurring when the sum of these motivating
forces reaches a critical threshold. Systematic investigation of such interactions could substantially advance
current understanding of the relapse process.

craving and relapse show that repeated withdrawal from

escalated cocaine intake leads to long-lasting deficits in
brain reward function, possibly involving synaptic rearrangement and altered neural excitability in the LH.
Clearly, major gaps remain in our knowledge of the
control of craving and relapse. For example, only the
time profile of BDNF corresponded well with incubation
of drug-seeking. The role of other molecular signaling
mechanisms, both in incubation and vulnerability to
relapse in general, remains to be established. Similarly,
the presumptive role of the LH in the consequences of
escalated drug intake will require demonstration of a link
between gene expression changes and allostatic load.
The selective reversal of conditioned reinstatement by a
Group II mGluR agonist (see also Update) and by N/
OFQ warrants acceleration of research devoted to these
systems, both with regard to a mechanistic role in addiction and as drug targets for relapse prevention. Furthermore, the important function postulated for the dorsal
striatum in consolidating addiction-related stimulusresponse habits would benefit from confirmation as to
whether this role applies specifically to ongoing drugseeking maintained by conditioned reinforcers or also to
relapse. This question is important because inactivation
of the dorsolateral striatum failed to interfere with conditioned reinstatement of cocaine-seeking following
extinction and abstinence [93]. Lastly, investigation of
how interactions among risk factors exacerbate the likelihood of resumption of drug use could substantially
advance our current understanding of the relapse process.

Update
Conclusions
The past three years have seen major advances in understanding of the experiential and neurobiological basis of
craving and vulnerability to relapse. Behaviourally, prolonged exposure to drugs of abuse in animals has consequences consistent with defining features of substance
dependence in humans, providing crucial support for the
utility of animal models as tools for investigating the
neural basis of specific aspects of addiction. In particular,
drug-related environmental stimuli, once conditioned to
drug-taking experiences, acquire the ability to elicit craving and precipitate relapse in a manner that is resistant to
extinction, impervious to punishment, and increases in
strength over prolonged abstinence. Novel data indicate
strong overlap between neural substrates controlling drug
desire evoked by drug cues and stress, the latter emerging
as an important factor in drug craving. Neuroadaptations
previously implicated in addiction using forced drug
administration regimens have been verified and extended
in animals voluntarily self-administering drugs of abuse
with findings of persistent perturbations in BDNF and its
signaling cascades, as well as in glutamate receptor subunits and gene transcripts linked to regulation of excitatory
neurotransmission. Data supporting the allostatic view of
www.sciencedirect.com

Recent work has confirmed that Group II mGLuR activation by the mGlu2/3 agonist LY379268 attenuates
context-induced reinstatement of heroin-seeking [94].
Moreover, these authors have identified Group II
mGluR-mediated glutamate transmission in the VTA
as important in contextual heroin reinstatement. Importantly, these results also verified the selectivity of Group
II mGluR activation for drug-seeking behavior, as
LY379268 had no effect on responding for oral sucrose.

Acknowledgements
The author acknowledges National Institutes of Health funding
through National Institute on Drug Abuse grants DA07348, DA08467
and DA017097; National Institute on Alcohol Abuse and Alcoholism
(NIAAA) grants AA10531 and AA014351; and a component of NIAAA
Alcohol Research Center grant AA06420. The author would like to thank
Mike Arends for his editorial assistance.

References and recommended reading

Papers of particular interest, published within the annual period of
review, have been highlighted as:
 of special interest
 of outstanding interest
1.

OBrien CP, Childress AR, Ehrman R, Robbins SJ: Conditioning

factors in drug abuse: can they explain compulsion?
J Psychopharmacol 1998, 12:15-22.
Current Opinion in Pharmacology 2005, 5:919

16 Neurosciences

2.

Sinha R: How does stress increase risk of drug abuse and

relapse? Psychopharmacology (Berl) 2001, 158:343-359.

3.


Franken IH: Drug craving and addiction: integrating

psychological and neuropsychopharmacological approaches.
Prog Neuropsychopharmacol Biol Psychiatry 2003, 27:563-579.
Integrating neurobiological and psychological findings, the author proposes that the psychological mechanism of attentional bias explains
craving and relapse in humans. According to this hypothesis, drug-paired
CS increase dopamine levels in the corticostriatal circuit, particularly in
the anterior cingulate gyrus, amygdala, and NAc. This doaminergic
activation draws attention to the drug-related stimulus, resulting in motor
preparation and a hyperattentive state towards drug-related stimuli that,
ultimately, promotes further craving and relapse.
4.

Ingjaldsson JT, Thayer JF, Laberg JC: Preattentive processing of

alcohol stimuli. Scand J Psychol 2003, 44:161-165.

5.

Tiffany ST: Cognitive concepts of craving. Alcohol Res Health

1999, 23:215-224.

6.

Altman J, Everitt BJ, Glautier S, Markou A, Nutt D, Oretti R,

Phillips GD, Robbins TW: The biological, social and clinical
bases of drug addiction: commentary and debate.
Psychopharmacology (Berl) 1996, 125:285-345.

7.


Wilson SJ, Sayette MA, Fiez JA: Prefrontal responses to

drug cues: a neurocognitive analysis. Nat Neurosci 2004,
7:211-214.
This paper reviews the functional neuroimaging literature of cue-elicited
craving in the context of a framework derived from behavioural data
indicating that perceived drug use opportunity significantly affects
responses to the presentation of drug cues. Consideration of this variable
provides a means by which discrepant findings among brain-imaging
studies of cue-induced craving can be reconciled.
8.

Littleton J: Can craving be modeled in animals? The relapse

prevention perspective. Addiction 2000, 95(Suppl 2):S83-S90.

9.

See RE: Neural substrates of conditioned-cued relapse to

drug-seeking behavior. Pharmacol Biochem Behav 2002,
71:517-529.

10. Everitt BJ, Dickinson A, Robbins TW: The neuropsychological

basis of addictive behaviour. Brain Res Brain Res Rev 2001,
36:129-138.
11. Kenny PJ, Koob GF, Markou A: Conditioned facilitation of brain

reward function after repeated cocaine administration.
Behav Neurosci 2003, 117:1103-1107.
Cocaine self-administration lowers brain stimulation reward thresholds,
reflecting increased brain reward function. This paper examined whether
cocaine-associated CS also reduce brain stimulation reward thresholds.
Substitution of intravenous cocaine with saline (the vehicle solution for
cocaine acting as drug CS) significantly lowered brain stimulation reward
thresholds, suggesting that cocaine-predictive stimuli produce conditioned facilitation of brain reward function.
12. Drummond DC: Theories of drug craving, ancient and modern.
Addiction 2001, 96:33-46.
13. Tiffany ST, Carter BL, Singleton EG: Challenges in the
manipulation, assessment and interpretation of craving
relevant variables. Addiction 2000, 95(Suppl 2):S177-S187.
14. Katz JL, Higgins ST: The validity of the reinstatement model of
 craving and relapse to drug use. Psychopharmacology (Berl)
2003, 168:21-30.
This insightful commentary discusses validity issues surrounding the
reinstatement procedure as a model of craving and relapse. The authors
conclude that the reinstatement model has reasonable face validity (i.e.
functional equivalence in terms of resembling relapse as it occurs
outside the laboratory). Concern is expressed, however, about the sparse
support favouring the predictive validity (i.e. the degree of correlation
between the outcomes of experimental manipulations in the laboratory
compared with clinical settings) and the functional equivalence (i.e.
mechanistic similarity of the laboratory phenomenon to relapse in drug
addicts) of the model. The most serious limitation identified by the authors
is that, as a model of craving, the reinstatement procedure might not
reach the level of even formal equivalence (i.e. face validity). Overall, a
need for caution is emphasized in generalizing results from the reinstatement model to clinical conditions.
15. Daglish MR, Nutt DJ: Brain imaging studies in human addicts.
Eur Neuropsychopharmacol 2003, 13:453-458.
Current Opinion in Pharmacology 2005, 5:919

16. Miller NS, Goldsmith RJ: Craving for alcohol and drugs in
animals and humans: biology and behavior. J Addict Dis 2001,
20:87-104.
17. Goldstein RZ, Volkow ND: Drug addiction and its underlying
neurobiological basis: neuroimaging evidence for the
involvement of the frontal cortex. Am J Psychiatry 2002,
159:1642-1652.
18. Cardinal RN, Parkinson JA, Hall J, Everitt BJ: Emotion and
motivation: the role of the amygdala, ventral striatum, and
prefrontal cortex. Neurosci Biobehav Rev 2002, 26:321-352.
19. Shaham Y, Shalev U, Lu L, De Wit H, Stewart J: The
reinstatement model of drug relapse: history, methodology
and major findings. Psychopharmacology (Berl) 2003, 168:3-20.
20. Everitt BJ, Robbins TW: Second-order schedules of drug
reinforcement in rats and monkeys: measurement of
reinforcing efficacy and drug-seeking behaviour.
Psychopharmacology (Berl) 2000, 153:17-30.
21. Di Ciano P, Everitt BJ: Differential control over drug-seeking
behavior by drug-associated conditioned reinforcers and
discriminative stimuli predictive of drug availability.
Behav Neurosci 2003, 117:952-960.
22. Goeders NE, Clampitt DM: Potential role for the hypothalamopituitary-adrenal axis in the conditioned reinforcer-induced
reinstatement of extinguished cocaine seeking in rats.
Psychopharmacology (Berl) 2002, 161:222-232.
23. Di Chiara G: Nucleus accumbens shell and core dopamine:
differential role in behavior and addiction. Behav Brain Res
2002, 137:75-114.
24. Fuchs RA, Evans KA, Parker MC, See RE: Differential

involvement of the core and shell subregions of the nucleus
accumbens in conditioned cue-induced reinstatement of
cocaine seeking in rats. Psychopharmacology (Berl) 2004,
176:459-465.
Transient inactivation of the NAc core, but not the shell, by a muscimolbaclofen (GABA agonist) mixture abolished conditioned reinstatement of
cocaine-seeking, implicating the NAc core as a critical accumbal subregion for CS-induced reinstatement.
25. Di Ciano P, Everitt BJ: Direct interactions between the
 basolateral amygdala and nucleus accumbens core underlie
cocaine-seeking behavior by rats. J Neurosci 2004,
24:7167-7173.
An elegant study utilizing a neuropharmacological disconnection procedure to investigate the neural control of conditioned cocaine-seeking on a
second order schedule of reinforcement. Unilateral infusion of a dopamine
receptor antagonist into the BLA (which on its own had no effect) and
concurrent application into the contralateral NAc core of an AMPA-kainate
receptor antagonist (which on its own also had no effect) greatly reduced
cocaine seeking. Complementing these findings, bilateral infusion of the
dopamine receptor antagonist into the BLA, but not the NAc core, or of the
AMPA-kainate receptor antagonist in the NAc core, but not the BLA, also
diminished cocaine seeking. The results identify a BLA-NAc core neural
link that regulates stimulus-controlled ongoing cocaine-seeking.
26. Ito R, Robbins TW, Everitt BJ: Differential control over cocaine seeking behavior by nucleus accumbens core and shell. Nat
Neurosci 2004, 7:389-397.
Selective excitotoxic lesion of the NAc core produced only minor impairments in acquisition of cocaine-reinforced responding under a schedule
of continuous reinforcement, but profoundly interfered with acquisition of
drug-seeking maintained by cocaine-paired conditioned reinforcers
under a second-order schedule contingency. By contrast, selective lesion
of the NAc shell did not impair cocaine self-administration or acquisition
of conditioned cocaine-seeking, but attenuated cocaines psychomotor
stimulant effects.
27. See RE, Kruzich PJ, Grimm JW: Dopamine, but not glutamate,
receptor blockade in the basolateral amygdala attenuates
conditioned reward in a rat model of relapse to cocaineseeking behavior. Psychopharmacology (Berl) 2001,
154:301-310.
28. Di Ciano P, Everitt BJ: Contribution of the ventral tegmental

area to cocaine-seeking maintained by a drug-paired
conditioned stimulus in rats. Eur J Neurosci 2004, 19:1661-1667.
Transient inactivation of the VTA or NAc core by a baclofen-muscimol
mixture decreased conditioned cocaine-seeking. Inactivation of the BLA,
www.sciencedirect.com

Neurobiology of craving, conditioned reward and relapse Weiss 17

however, had no effect. Although these data further confirm a selective

role for the NAc core in conditioned reinforcement, they are somewhat
difficult to reconcile with an earlier finding by the same authors [26]
implicating dopaminergic transmission in the BLA as critical in CSmaintained cocaine-seeking.
29. Ghitza UE, Fabbricatore AT, Prokopenko V, Pawlak AP, West MO:

Persistent cue-evoked activity of accumbens neurons after
prolonged abstinence from self-administered cocaine.
J Neurosci 2003, 23:7239-7245.
Using single-unit recording in rats, the authors demonstrate that NAc shell
neurons participate in the process of discriminating the motivational value
of cocaine-predictive versus neutral contextual stimuli, whereas NAc core
neurons do not. Importantly, NAc shell neurons process information on
reward-related contextual stimuli even after weeks of abstinence.
30. McLaughlin J, See RE: Selective inactivation of the dorsomedial
prefrontal cortex and the basolateral amygdala attenuates
conditioned-cued reinstatement of extinguished cocaineseeking behavior in rats. Psychopharmacology (Berl) 2003,
168:57-65.
31. Fuchs RA, Evans KA, Parker MP, See RE: Differential

involvement of orbitofrontal cortex subregions in conditioned
cue-induced and cocaine-primed reinstatement of cocaine
seeking in rats. J Neurosci 2004, 24:6600-6610.
Using fibre-sparing lesion or functional inactivation methodologies, a
differential role of OFC subregions was identified in conditioned versus
cocaine-induced reinstatement in rats. The central finding of this thorough and elaborate study implicates the medial OFC in cocaine-induced
reinstatement and the lateral OFC in cue-induced reinstatement, the latter
by a role in evaluating the current motivational significance of cocaine CS.
32. Fuchs RA, Evans KA, Ledford CC, Parker MP, Case JM, Mehta RH,

See RE: The role of the dorsomedial prefrontal cortex,
basolateral amygdala, and dorsal hippocampus in contextual
reinstatement of cocaine seeking in rats.
Neuropsychopharmacology 2004, in press.
Transient inactivation of the dorsal hippocampus by tetrodotoxin abolished the effects of contextual stimuli on reinstatement, but not those of
cocaine or an explicit CS. Contextual reinstatement was also abolished
by inactivation of the BLA or dorsomedial PFC. These data implicate
critical roles for the dorsal hippocampus, BLA and dorsomedial PFC in
contextual reinstatement. However, the data show also that, although
distinct substrates mediating reinstatement by drug-associated contextual stimuli versus explicit CS versus cocaine can be identified, there is
also overlap among substrates mediating the effects of these diverse
reinstatement-evoking manipulations.
33. Yun IA, Fields HL: Basolateral amygdala lesions impair both

cue- and cocaine-induced reinstatement in animals trained
on a discriminative stimulus task. Neuroscience 2003,
121:747-757.
The main finding of this study is that, in rats, a discriminative stimulus (SD;
a type of contextual stimulus) that signals cocaine availability is a more
robust trigger of reward-seeking than a Pavlovian CS signaling reward
delivery. The authors show also that excitotoxic lesion of the BLA
abolishes SD-induced (i.e. contextual) reinstatement, consistent with
the results of [32].
34. Kantak KM, Black Y, Valencia E, Green-Jordan K,
Eichenbaum HB: Dissociable effects of lidocaine inactivation of
the rostral and caudal basolateral amygdala on the
maintenance and reinstatement of cocaine-seeking behavior
in rats. J Neurosci 2002, 22:1126-1136.
35. Ghitza UE, Fabbricatore AT, Prokopenko VF, West MO:
Differences between accumbens core and shell neurons
exhibiting phasic firing patterns related to drug-seeking
behavior during a discriminative-stimulus task.
J Neurophysiol 2004, 92:1608-1614.
36. Miller CA, Marshall JF: Altered prelimbic cortex output during
cue-elicited drug seeking. J Neurosci 2004, 24:6889-6897.
37. Franklin TR, Acton PD, Maldjian JA, Gray JD, Croft JR, Dackis CA,
OBrien CP, Childress AR: Decreased gray matter
concentration in the insular, orbitofrontal, cingulate, and
temporal cortices of cocaine patients. Biol Psychiatry 2002,
51:134-142.
38. Schoenbaum G, Saddoris MP, Ramus SJ, Shaham Y, Setlow B:
Cocaine-experienced rats exhibit learning deficits in a task
sensitive to orbitofrontal cortex lesions. Eur J Neurosci 2004,
19:1997-2002.
www.sciencedirect.com

39. Volkow ND, Fowler JS: Addiction, a disease of compulsion

and drive: involvement of the orbitofrontal cortex.
Cereb Cortex 2000, 10:318-325.
40. Fein G, Di Sclafani V, Meyerhoff DJ: Prefrontal cortical volume
reduction associated with frontal cortex function deficit in
6-week abstinent crack-cocaine dependent men.
Drug Alcohol Depend 2002, 68:87-93.
41. Bechara A, Damasio AR, Damasio H, Anderson SW: Insensitivity
to future consequences following damage to human
prefrontal cortex. Cognition 1994, 50:7-15.
42. Grimm JW, Lu L, Hayashi T, Hope BT, Su TP, Shaham Y:

Time-dependent increases in brain-derived neurotrophic
factor protein levels within the mesolimbic dopamine system
after withdrawal from cocaine: implications for incubation of
cocaine craving. J Neurosci 2003, 23:742-747.
This research examines the relationship between behavioural incubation
of cocaine-seeking and changes in BDNF protein levels in the mesolimbic
dopaminergic system. BDNF, but not nerve growth factor, levels in the
NAc, amygdala, and VTA increased during cocaine, but not sucrose,
withdrawal in parallel with behavioural incubation of cocaine-seeking. The
correspondence between the time profiles of incubation of cocaineseeking and BDNF expression was strongest in the NAc and weakest
in the VTA.
43. Lu L, Grimm JW, Dempsey J, Shaham Y: Cocaine seeking over
extended withdrawal periods in rats: different time courses of
responding induced by cocaine cues versus cocaine priming
over the first 6 months. Psychopharmacology (Berl) 2004,
176:101-108.
44. Deroche-Gamonet V, Belin D, Piazza PV: Evidence for addiction like behavior in the rat. Science 2004, 305:1014-1017.
A seminal paper reporting that behaviours reflective of three essential
diagnostic criteria for addiction in humans can be demonstrated to
develop with prolonged cocaine self-administration in rats: drug use
despite adverse outcomes, difficulty in limiting drug intake, and preoccupation with procurement and consumption of the drug. This set of
behaviours was evident only in a small proportion of rats, with the majority
of animals being non-responsive to any of the addiction-like behaviours
measured. These findings strikingly resemble human epidemiological
data showing that only a small proportion of drug users progress to an
addicted state.
45. Ciccocioppo R, Martin-Fardon R, Weiss F: Stimuli associated

with a single cocaine experience elicit long-lasting
cocaine-seeking. Nat Neurosci 2004, 7:495-496.
This paper presents evidence that contextual stimuli associated with a
singe life-time cocaine self-administration experience in rats elicit strong
drug-seeking that persists for up to one year. By contrast, contextual
stimuli associated with availability of a highly palatable conventional
reinforcer produced only modest seeking behaviour that extinguished
rapidly. Thus, conditioned responses to drug cues not only facilitate
relapse during abstinence in drug-experienced subjects, but perhaps
also provoke craving related to initial sporadic drug experiences, leading
to more frequent drug use.
46. Di Ciano P, Everitt BJ: Reinstatement and spontaneous
recovery of cocaine-seeking following extinction and different
durations of withdrawal. Behav Pharmacol 2002, 13:397-405.
47. Grimm JW, Shaham Y, Hope BT: Effect of cocaine and sucrose
withdrawal period on extinction behavior, cue-induced
reinstatement, and protein levels of the dopamine transporter
and tyrosine hydroxylase in limbic and cortical areas in rats.
Behav Pharmacol 2002, 13:379-388.
48. Grimm JW, Hope BT, Wise RA, Shaham Y: Neuroadaptation.
incubation of cocaine craving after withdrawal. Nature 2001,
412:141-142.
49. Vanderschuren LJ, Everitt BJ: Drug seeking becomes
 compulsive after prolonged cocaine self-administration.
Science 2004, 305:1017-1019.
One of a series of recent seminal papers providing preclinical evidence for
the progression from casual to compulsive drug use. Here, rats with an
extended cocaine self-administration history no longer showed conditioned suppression of drug-seeking in the presence of aversive CS. The
authors show that this effect was not accounted for by either impaired
fear conditioning or an increase in cocaines incentive value. As in [44],
the maintenance of drug-seeking by rats in the face of adverse consequences a hallmark feature of addiction is consistent with the
establishment of compulsive behaviour.
Current Opinion in Pharmacology 2005, 5:919

18 Neurosciences

50. Koob GF, Ahmed SH, Boutrel B, Chen SA, Kenny PJ, Markou A,
ODell LE, Parsons LH, Sanna PP: Neurobiological mechanisms
in the transition from drug use to drug dependence. Neurosci
Biobehav Rev 2004, 27:739-749.
51. Hutcheson DM, Everitt BJ, Robbins TW, Dickinson A: The role of
withdrawal in heroin addiction: enhances reward or promotes
avoidance? Nat Neurosci 2001, 4:943-947.
52. Liu X, Weiss F: Additive effect of stress and drug cues on
reinstatement of ethanol seeking: exacerbation by history
of dependence and role of concurrent activation of
corticotropin-releasing factor and opioid mechanisms.
J Neurosci 2002, 22:7856-7861.
53. Liu X, Weiss F: Reversal of ethanol-seeking behavior by D1 and
D2 antagonists in an animal model of relapse: differences
in antagonist potency in previously ethanol-dependent
versus nondependent rats. J Pharmacol Exp Ther 2002,
300:882-889.
54. Ciccocioppo R, Lin D, Martin-Fardon R, Weiss F: Reinstatement
of ethanol-seeking behavior by drug cues following single
versus multiple ethanol intoxication in the rat: effects of
naltrexone. Psychopharmacology (Berl) 2003, 168:208-215.
55. Liu X, Weiss F: Stimulus conditioned to foot-shock stress
reinstates alcohol-seeking behavior in an animal model of
relapse. Psychopharmacology (Berl) 2003, 168:184-191.
56. Baptista MA, Martin-Fardon R, Weiss F: Preferential effects of

the metabotropic glutamate 2/3 receptor agonist LY379268 on
conditioned reinstatement versus primary reinforcement:
comparison between cocaine and a potent conventional
reinforcer. J Neurosci 2004, 24:4723-4727.
Group II mGluRs have received attention with respect to a role in
addiction and as potential treatment drug targets. The data in this report
support this hypothesis in that pharmacological activation of Group II
mGluRs preferentially reduced the reinforcing effects of cocaine compared with a potent natural reinforcer, which remained unaltered, and
preferentially reduced conditioned reinstatement as opposed to primary
reinforcement by cocaine and the natural reinforcer.
57. Ciccocioppo R, Economidou D, Fedeli A, Angeletti S, Weiss F,

Heilig M, Massi M: Attenuation of ethanol self-administration
and of conditioned reinstatement of alcohol-seeking
behaviour by the antiopioid peptide nociceptin/orphanin FQ in
alcohol-preferring rats. Psychopharmacology (Berl) 2004,
172:170-178.
Evidence is presented that N/OFQ the endogenous ligand of the
opioid-like orphan receptor reduces alcohol self-administration under
continuous and progressive ratio reinforcement contingencies, and
attenuates contextual conditioned reinstatement. Importantly, the effects
of N/OFQ were selective for drug-related behaviour, because the peptide
did not alter behaviour maintained by sucrose.
58. Lu L, Dempsey J, Liu SY, Bossert JM, Shaham Y: A single

infusion of brain-derived neurotrophic factor into the ventral
tegmental area induces long-lasting potentiation of cocaine
seeking after withdrawal. J Neurosci 2004, 24:1604-1611.
This report extends the findings in [42] with data further supporting
the view that neuroadaptation in BDNF and/or signaling cascades
initiated by BDNF contribute to the incubation phenomenon during
cocaine withdrawal.
59. Thomas GM, Huganir RL: MAPK cascade signaling and synaptic
plasticity. Nat Rev Neurosci 2004, 5:173-183.
60. Wolf ME: LTP may trigger addiction. Mol Interv 2003, 3:248-252.
61. Lu L, Grimm JW, Shaham Y, Hope BT: Molecular
neuroadaptations in the accumbens and ventral tegmental
area during the first 90 days of forced abstinence from cocaine
self-administration in rats. J Neurochem 2003, 85:1604-1613.
62. Bolanos CA, Nestler EJ: Neurotrophic mechanisms in drug

addiction. Neuromolecular Med 2004, 5:69-83.
An excellent brief review of the role of neurotrophins and other neurotrophic factors, as well as the signaling pathways they activate, in longterm molecular, cellular, and behavioural adaptations associated with
drug addiction.
63. Li Y, Kolb B, Robinson TE: The location of persistent
amphetamine-induced changes in the density of dendritic
spines on medium spiny neurons in the nucleus accumbens
Current Opinion in Pharmacology 2005, 5:919

and caudate-putamen. Neuropsychopharmacology 2003,

28:1082-1085.
64. Robinson TE, Gorny G, Mitton E, Kolb B: Cocaine selfadministration alters the morphology of dendrites and
dendritic spines in the nucleus accumbens and neocortex.
Synapse 2001, 39:257-266.
65. Robinson TE, Gorny G, Savage VR, Kolb B: Widespread but
regionally specific effects of experimenter- versus selfadministered morphine on dendritic spines in the nucleus
accumbens, hippocampus, and neocortex of adult rats.
Synapse 2002, 46:271-279.
66. Norrholm SD, Bibb JA, Nestler EJ, Ouimet CC, Taylor JR,
Greengard P: Cocaine-induced proliferation of dendritic spines
in nucleus accumbens is dependent on the activity of
cyclin-dependent kinase-5. Neuroscience 2003, 116:19-22.
67. Valjent E, Corvol JC, Pages C, Besson MJ, Maldonado R, Caboche
J: Involvement of the extracellular signal-regulated kinase
cascade for cocaine-rewarding properties. J Neurosci 2000,
20:8701-8709.
68. Baker DA, McFarland K, Lake RW, Shen H, Tang XC, Toda S,
Kalivas PW: Neuroadaptations in cystine-glutamate exchange
underlie cocaine relapse. Nat Neurosci 2003, 6:743-749.
69. Bowers MS, McFarland K, Lake RW, Peterson YK, Lapish CC,
Gregory ML, Lanier SM, Kalivas PW: Activator of G protein
signaling 3: a gatekeeper of cocaine sensitization and drug
seeking. Neuron 2004, 42:269-281.
70. Tanda G, Goldberg SR: Cannabinoids: reward, dependence,
and underlying neurochemical mechanismsa review of
recent preclinical data. Psychopharmacology (Berl) 2003,
169:115-134.
71. Naassila M, Pierrefiche O, Ledent C, Daoust M: Decreased
alcohol self-administration and increased alcohol sensitivity
and withdrawal in CB1 receptor knockout mice.
Neuropharmacology 2004, 46:243-253.
72. Gonzalez S, Schmid PC, Fernandez-Ruiz J, Krebsbach R,
Schmid HH, Ramos JA: Region-dependent changes in
endocannabinoid transmission in the brain of morphinedependent rats. Addict Biol 2003, 8:159-166.
73. De Vries TJ, Homberg JR, Binnekade R, Raaso H,
Schoffelmeer AN: Cannabinoid modulation of the reinforcing
and motivational properties of heroin and heroin-associated
cues in rats. Psychopharmacology (Berl) 2003, 168:164-169.
74. Kelley AE, Berridge KC: The neuroscience of natural
rewards: relevance to addictive drugs. J Neurosci 2002,
22:3306-3311.
75. Cardinal RN, Everitt BJ: Neural and psychological mechanisms
underlying appetitive learning: links to drug addiction. Curr
Opin Neurobiol 2004, 14:156-162.
76. Carelli RM: Nucleus accumbens cell firing during goal-directed
behaviors for cocaine vs natural reinforcement. Physiol Behav
2002, 76:379-387.
77. Daglish MR, Weinstein A, Malizia AL, Wilson S, Melichar JK,
 Lingford-Hughes A, Myles JS, Grasby P, Nutt DJ: Functional
connectivity analysis of the neural circuits of opiate
craving: more rather than different? Neuroimage 2003,
20:1964-1970.
The authors employed positron emission tomography scanning to investigate connectivity patterns associated with the primary brain regions
activated by drug-craving memories (the anterior cingulate) and correlated with opiate craving (the OFC). The data revealed an ability of drugrelated stimuli to activate attentional and memory circuits to a greater
degree than non-drug-related stimuli, leading the authors to conclude
that neural circuits of dependence and craving are not specific craving
or addiction circuits, but normal circuits activated to a greater degree.
78. Koob GF, Le Moal M: Drug addiction, dysregulation of
reward, and allostasis. Neuropsychopharmacology 2001,
24:97-129.
79. Ahmed SH, Kenny PJ, Koob GF, Markou A: Neurobiological
evidence for hedonic allostasis associated with escalating
cocaine use. Nat Neurosci 2002, 5:625-626.
www.sciencedirect.com

Neurobiology of craving, conditioned reward and relapse Weiss 19

80. Tang W, Wesley M, Freeman WM, Liang B, Hemby SE: Alterations

in ionotropic glutamate receptor subunits during binge
cocaine self-administration and withdrawal in rats.
J Neurochem 2004, 89:1021-1033.
81. DiLeone RJ, Georgescu D, Nestler EJ: Lateral hypothalamic
neuropeptides in reward and drug addiction. Life Sci 2003,
73:759-768.
82. Lu L, Shepard JD, Hall FS, Shaham Y: Effect of environmental
stressors on opiate and psychostimulant reinforcement,
reinstatement and discrimination in rats: a review.
Neurosci Biobehav Rev 2003, 27:457-491.
83. Le A, Shaham Y: Neurobiology of relapse to alcohol in rats.
Pharmacol Ther 2002, 94:137-156.
84. Goeders NE: The impact of stress on addiction.

Eur Neuropsychopharmacol 2003, 13:435-441.
A concise review summarizing both from a preclinical and clinical perspective the significance of stress in various aspects of addiction, with
emphasis on the role of the HPA axis.
85. Foltin RW, Haney M: Conditioned effects of environmental
stimuli paired with smoked cocaine in humans.
Psychopharmacology (Berl) 2000, 149:24-33.
86. Sinha R, Talih M, Malison R, Cooney N, Anderson GM, Kreek MJ:
 Hypothalamic-pituitary-adrenal axis and sympatho-adrenomedullary responses during stress-induced and drug
cue-induced cocaine craving states. Psychopharmacology
(Berl) 2003, 170:62-72.
An elaborate study that scrutinizes the response of brain stress circuits in
cocaine-dependent subjects exposed to guided imagery procedures
known to increase cocaine craving. Stress and drug imagery manipulations each produced significant increases in craving and subjective
anxiety, pulse rate, systolic blood pressure and levels of adrenocorticotropic hormone, cortisol, prolactin and norepinephrine, compared with
neutral imagery. Of interest, these data confirm not only that similar
craving states are elicited by stress and drug cue exposure, but also
that the neurobiological basis of these reactions is shared with a significant activation of the CRF/HPA axis and noradrenergic/sympathoadreno-medullary system response.
87. Capriles N, Rodaros D, Sorge RE, Stewart J: A role for the

prefrontal cortex in stress- and cocaine-induced
reinstatement of cocaine seeking in rats. Psychopharmacology
(Berl) 2003, 168:66-74.

www.sciencedirect.com

This paper addresses the issue of common neural substrates for cocaine
craving associated with different risk factors, with a focus on the role of
the PFC in cocaine-seeking induced by footshock stress and priming
injections. Based on the effects of transient tetrodotoxin lesions, sitespecific pharmacological manipulations and findings in related literature,
a picture emerges that implicates the PLC and OFC as part of the circuitry
specifically mediating the effects of footshock stress on reinstatement.
The PLC is proposed as a possible common pathway for cue, foot-shock
and cocaine-induced reinstatement of drug seeking.
88. Scaccianoce S, Matrisciano F, Del Bianco P, Caricasole A,
Di Giorgi Gerevini V, Cappuccio I, Melchiorri D, Battaglia G,
Nicoletti F: Endogenous activation of group-II metabotropic
glutamate receptors inhibits the hypothalamic-pituitaryadrenocortical axis. Neuropharmacology 2003, 44:555-561.
89. Linden AM, Greene SJ, Bergeron M, Schoepp DD: Anxiolytic
activity of the mGlu2/3 receptor agonist LY354740 on the
elevated plus maze is associated with the suppression of
stress-induced c-Fos in the hippocampus and increases in
c-Fos induction in several other stress-sensitive brain
regions. Neuropsychopharmacology 2004, 29:502-513.
90. Swanson CJ, Perry KW, Schoepp DD: The mGlu2/3 receptor
agonist, LY354740, blocks immobilization-induced increases
in noradrenaline and dopamine release in the rat medial
prefrontal cortex. J Neurochem 2004, 88:194-202.
91. Ciccocioppo R, Fedeli A, Economidou D, Policani F, Weiss F,
Massi M: The bed nucleus is a neuroanatomical substrate for
the anorectic effect of corticotropin-releasing factor and for
its reversal by nociceptin/orphanin FQ. J Neurosci 2003,
23:9445-9451.
92. Valdez GR, Zorrilla EP, Roberts AJ, Koob GF: Antagonism of
corticotropin-releasing factor attenuates the enhanced
responsiveness to stress observed during protracted ethanol
abstinence. Alcohol 2003, 29:55-60.
93. Kantak KM, Black Y, Valencia E, Green-Jordan K,
Eichenbaum HB: Stimulus-response functions of the lateral
dorsal striatum and regulation of behavior studied in a
cocaine maintenance/cue reinstatement model in rats.
Psychopharmacology (Berl) 2002, 161:278-287.
94. Bossert JM, Liu SY, Lu L, Shaham Y: A role of ventral tegmental
area glutamate in contextual cue-induced relapse to heroin
seeking. J Neurosci 2004, 24:in press.

Current Opinion in Pharmacology 2005, 5:919