Embedded Files
Drug addiction, also known as substance use disorder (SUD), is a disease with a problematic pattern of drug use leading to clinically significant impairment or distress defined by diagnostic criteria in DSM-5. The criteria pertain to loss of volitional control over substance use and continued use despite significant physical/psychological harm and significant social, occupational, and interpersonal problems. The criteria also include tolerance and withdrawal.
Substance use disorder mainly involves changes in the prefrontal cortex and dorsal and ventral striatum including nucleus accumbens. Typically, drug use is initiated through the explicit (declarative) learning system involving the hippocamal complex and prefrontal cortex. The dorsomedial and dorsolateal striatum are involved in the goal-directed and habitual behavior respectively. Drug use causes biochemical changes in these neural circuits and lead to habitual and compulsive drug use in susceptible individuals. Clinical manifestations in diseases affecting the dopamine system include deficits in emotional, cognitive, and motor function. The parallel organization of specific corticostriatal pathways is well documented. The ventromedial striatum (limbic) projects to a wide range of the dopamine cells and receives a relatively small dopamine input. In contrast, the dorsolateral striatum (DLS, motor) receives input from a broad expanse of dopamine cells and has a confined input to the substantia nigra (SN). The central striatum (CS, associative) receives input from and projects to a relatively wide range of the SN. There is an interface between different striatal regions via the midbrain dopamine cells that forms an ascending spiral between regions. The shell influences the core, the core influences the central striatum, and the central striatum influences the dorsolateral striatum (Haber et al, 2000). Information, coded as neural signals, flows independently through the three systems, the hippocampus, the matrix compartment of the dorsal striatum (caudate-putamen), and the amygdala. The systems interact in at least two ways: by simultaneous parallel influence on behavioral output and by directly influencing each other (White and McDonald, 2002).
Drugs act as "instrumental reinforcers". They increase the likelihood of responses that procure them, resulting in drug self-administration or "drug taking". Environmental stimuli that are closely associated in time and space with the effects of self-administered drugs gain incentive salience through the process of pavlovian conditioning. CSs that predict natural reinforcers, such as a light that predicts food, can have several effects on behavior, in addition to eliciting pavlovian (that is, automatic or reflexive) elements of approach and consummatory behavior, such as the locomotor stimulation produced by amphetamine and cocaine. Selective lesions of the nucleus accumbens core or infusions of NMDA or dopamine receptor antagonists into the nucleus accumbens core during training (reference) greatly retard the acquisition of pavlovian approach responses, whereas infusions of NMDA or dopamine D1 receptor antagonists into this region after a training trial disrupt the consolidation of this response into memory (reference). Lesions of the nucleus accumbens core also abolish PIT (reference), and increasing dopamine in the nucleus accumbens shell potentiates PIT (reference). In contrast to the Pavlovian process, motivational states do not influence the instrumental process directly; rather, the agent has to learn about the value of an outcome in a given motivational state by exposure to it while in that state.
Drug self-administration behavior, including drug seeking under second-order schedules of reinforcement, initially involves action-outcome learning, before extended training additionally leads to the formation of stimulus-response (‘habit’) associations that help maintain responding (references). For example, a second-order fixed-ratio schedule of drug delivery can be organized into 10 blocks with 5 responses per block. Every fifth response produces a conditioned reinforcer (e.g., a visual stimulus) and the completion of the tenth block (50 responses) produces the conditioned reinforcer accompanied by drug delivery. The nucleus accumbens shell is connected to the full network of descending neural influences over reflexive autonomic and motor responses (references). This region is necessary for the direct psychomotor stimulant effects of the cocaine, including response rate–enhancing and locomotor activity effects (reference). The mixed dopamine receptor antagonist α-flupenthixol infused into the dorsal striatum greatly reduces well-established cocaine seeking under a second-order schedule, yet it has no effect when infused into the nucleus accumbens core (references). This is consistent with the presence of ‘error prediction’ dopamine neurons innervating the entire striatum, including its dorsal as well as ventral regions (references). The ventral striatum is implicated in pavlovian conditioning and the dorsal striatum with instrumental learning (reference). Models have consistently suggested that it is the conditioned reinforcing, rather than other, effects of Pavlovian drug stimuli that most profoundly influence drug seeking. However, neither approach to a CS predictive of a drug, nor enhancement of drug seeking by the unexpected presentation of a drugassociated CS has been clearly demonstrated in laboratory studies of drug seeking or relapse, although both are readily seen in animals responding for natural rewards.
Selective lesions of the basolateral amygdala or the nucleus accumbens core impair the acquisition of cocaine or heroin seeking under a second-order schedule. Basolateral amygdala lesions, like nucleus accumbens core lesions, increase the choice of small, immediate rewards over larger, delayed rewards—indicating greater impulsivity (reference). Dopamine (but not AMPA) receptor blockade bilaterally in the basolateral amygdala impairs cocaine seeking under a second-order schedule, whereas AMPA (but not dopamine) receptor blockade bilaterally in the nucleus accumbens core has a similar effect, suggesting that associative information in the basolateral amygdala is translated into goal-directed, drug seeking behavior via its interactions with the nucleus accumbens core. Selective lesions of the orbital prefrontal cortex (OFC) also impair the acquisition of cocaine seeking suggesting that the basolateral amygdala and OFC cooperate to regulate goal-directed behavior.
The amygdala mediates conditioning to discrete CSs, the hippocampal formation underlies conditioning to contextual or spatial stimuli and may therefore underlie the motivational impact of contextual stimuli on drug seeking. We hypothesize that, in psychological terms, hippocampal mechanisms provide the contextual background that defines the motivational arousal upon which goal-directed responding occurs. Inactivating this mechanism at the hippocampus or nucleus accumbens shell level reduces exploration, activity and contextual conditioning and also the potentiation of these responses by psychomotor stimulants—providing, therefore, an additional basis for understanding the reinforcing effects of drugs acting on the dopamine and other systems in the nucleus accumbens. The involvement of the mPFC in the reinstatement of drug seeking depends on glutamate release in the nucleus accumbens core and also on the integrity of the ventral pallidum, providing clear evidence of the function of specific limbic cortical-ventral striatopallidal circuits. Lesions of the mPFC (including the prelimbic and infralimbic cortex) result in increased responding for cocaine under a second-order schedule of reinforcement and also enhance the acquisition of cocaine self-administration. Chronic drug abusers show deficits in tests of inhibitory control and decision making. Overall, we hypothesize that the transition from voluntary actions (governed mainly by their consequences) to more habitual modes of responding in drug seeking behavior represents a transition from prefrontal cortical to striatal control over responding, and from ventral to more dorsal striatal subregions.
A decade ago, we suggested that aversive motivational states, as in withdrawal phenomena (Koob & LeMoal 1997), drug-induced sensitization (Robinson & Berridge 1993), and a loss of top-down inhibitory response control ( Jentsch & Taylor 1999, Robbins & Everitt 1999) might all be contributory factors. One of the aims of this review is to re-examine the relative importance of these factors in the light of recent empirical evidence in experimental studies of both animals and humans.
It rapidly became apparent that responses to drugs can acquire motivational significance by being associated with environmental stimuli through Pavlovian conditioning (Gawin & Kleber 1986, O’Brien et al. 1998) and that these drug-associated CSs exert a marked influence on the instrumental behaviors of drug seeking and use (Everitt et al. 2001), induce subjective craving states, and precipitate relapse long into abstinence (Childress et al. 1999, Garavan et al. 2000, Grant et al. 1996). As a result, concepts of addiction incorporating learning theory accounts of Pavlovian conditioning mechanisms (Everitt et al. 2001, Robinson & Berridge 1993, Saunders & Robinson 2013, Stewart et al. 1984) and, more recently, instrumental learning mechanisms (Everitt et al. 2001, Everitt & Robbins 2005, Robbins & Everitt 1999), have become more prominent.
Using a novel method in which cocaine-seeking responses on one lever gave access to a second, “taking” lever delivering i.v. cocaine, we showed that devaluation achieved by extinguishing specifically the taking link of the chain reduced seeking early in training, which confirmed that it was goal directed (Olmstead et al. 2001). This demonstration was later confirmed and extended by Zapata et al. (2010), who showed additionally that cocaine-seeking behavior became insensitive to devaluation, i.e., habitual, after extended training (Zapata et al. 2010). This effect of extended training to promote habitual drug seeking has further been shown in rats self-administering nicotine (Clemens et al. 2014). Alcohol-dependent human subjects also exhibit overreliance on S-R representations, as shown in a computer-based task distinguishing between goal-directed and habitual control (Sjoerds et al. 2013). In fact, observations in humans and animals indicate that even noncontingent (i.e., not self-administered) addictive drug exposure can tip the balance between A-O and S-R associative mechanisms to favor the latter. Thus, for humans undergoing instrumental training for chocolate reward, noncontingent alcohol administration attenuated goal-directed control over chocolate choice and accelerated habit learning (Hogarth et al. 2013). In rats, noncontingent alcohol exposure accelerated the development of habitual control over natural reward seeking (Corbit et al. 2012), whereas repeated noncontingent amphetamine treatment resulted in extremely rapid development of habitual control over responding for sucrose (Nelson & Killcross 2006). The goal-directed system in both rats and humans depends on interactions between the medial prefrontal cortex (mPFC) and the posterior dorsomedial striatum (pDMS) (Shiflett et al. 2010, Yin et al. 2005). In contrast, the habit system implicates the anterior dorsolateral striatum (aDLS), or putamen in humans, and perhaps motor cortical areas (Balleine&O’Doherty 2010, Yin et al. 2004). The aDLS is not required for initial cocaine seeking, when it is goal directed, but gradually becomes dominant in the control over this behavior when it is well established and habitual (Murray et al. 2012). By contrast, DA receptor blockade in the pDMS impaired the acquisition of cocaine seeking when goal directed but had no effect after extended training (Murray et al. 2012). An overreliance on S-R habit learning in alcohol-dependent individuals was associated with the decreased activation of areas of the brain implicated in A-O, goal-directed learning, such as the ventromedial PFC and anterior putamen, but with the increased activation of the posterior putamen that mediates habit learning (Sjoerds et al. 2013).
In human subjects engaged in learning a virtual maze task that could dissociate spatial and S-R response navigational strategies, response learners had increased dorsal striatal gray matter volume and activity measured using functional magnetic resonance imaging (fMRI), whereas spatial learners had increased hippocampal gray matter and activity. Furthermore, response learners had greater use of abused substances in comparison with spatial learners, including double the lifetime alcohol consumption, a greater number of cigarettes smoked, and a greater lifetime use of cannabis (Bohbot et al. 2013). Cocaine-dependent individuals and their non-cocaine-abusing siblings had a significantly enlarged left putamen (Ersche et al. 2011a, 2013a), suggesting that greater dorsal striatal (putamen) volume may be associated with a predisposition to acquire drug-seeking and -taking habits.
The neural circuitry by which the amygdala influences habitual instrumental behavior in the DLS is unclear, since it does not project there directly. The two main routes by which amygdala-DLS interactions could occur are (a) via glutamatergic basolateral amygdala (BLA) projections to the NAcbC, which can thereby influence the spiraling dopaminergic circuitry linking the core with the DLS, and (b) via the central amygdala (CeN) projections directly to the substantia nigra pars compacta that dopaminergically innervates the DLS. The latter amygdala CeN-DLS system has been shown functionally to be important for Pavlovian conditioned orienting (Han et al. 1997) and for the acquisition of food-reinforced habits (Lingawi & Balleine 2012). By disconnecting either the BLA or the CeN from the DLS (by combining a unilateral lesion of either the CeN or BLA with infusion of a DA receptor antagonist into the contralateral aDLS), Murray and colleagues (2013) have demonstrated the functional importance of this polysynaptic and indirect circuitry in the acquisition and maintenance of cue-controlled cocaine-seeking habits. Moreover, it was further shown electrophysiologically that activation of the BLA can both up- and downregulate cortically driven aDLS medium spiny neuron activity via its projections to the NAcbC (Belin-Rauscent et al. 2013).
Negative emotional states, such as those instantiated by stress, can also influence habit learning.Thus, rats subjected to chronic stress rapidly developed insensitivity to outcome value and werer elatively impervious to changes in A-O contingency. These changes biased rats toward habitual behavioral strategies and resulted in atrophy of the mPFC and the associative striatum and hypertrophy of the DLS (Dias-Ferreira et al. 2009). An intriguing possibility, then, is that drug withdrawal stress that is known to result in raised reward thresholds and long-term changes in hedonic state (Koob 2008) may, through activation of stress systems in the amygdala (Koob 2008), also influence the development of S-R drug-seeking habits by facilitating its couplingwith theDLS (Lingawi & Balleine 2012, Murray et al. 2013). Indeed, instrumental avoidance learning, which presumably contributes to withdrawal-motivated negative reinforcement, becomes impervious to extinction and may resemble compulsive habits in OCD (Gillan et al. 2015).
The phenomenon of sensitization has now clearly been demonstrated in humans exposed relatively few times to amphetamine, leading to very long-lasting enhancements in striatal DA responses to both drugs and drug CSs (Leyton 2007, Leyton & Vezina 2014, Vezina & Leyton 2009). One consequence of this process is that drug CSs, through their enhanced ability to increase DA release in the ventral striatum, may lead to subjective craving states and whatmight be assumed to be a voluntary urge to seek and take drugs (Leyton & Vezina 2014). The enhanced DA transmission underlying sensitization is not restricted to the ventral striatum but is also seen in the dorsal striatum and is associated with the potentiated expression of motor stereotypies. The latter depend upon the dorsal striatum (Kelly et al. 1975) and are putatively a form of compulsive responding. Stimulant sensitization may therefore lead both to potentiated motivational and Pavlovian associative processes and a parallel enhancement of S-R learning mediated by upregulation of DA in the ventral and dorsal striatum, respectively.
In rats (as in humans), about 20% exhibit compulsive drug seeking despite adverse consequences, but only after chronic drug use (Belin et al. 2008, Deroche-Gamonet et al. 2004, Pelloux et al. 2007). We modified our previously established cocaine seeking-taking chained schedule (Olmstead et al. 2001) to introduce unpredictable footshock punishment and therefore required rats to risk these adverse consequences when seeking the opportunity to take cocaine (Pelloux et al. 2007). After a brief cocaine history, all rats stopped seeking cocaine when the punishment contingency was introduced (i.e., they abstained from drug use), but after a long history of cocaine self-administration, some 20% of rats continued to seek cocaine (i.e., were compulsive) (Pelloux et al. 2007). The extent of exposure to cocaine, rather than the degree of conditioning through Pavlovian pairings of CS and drug, was further shown to be a critical factor in determining the development of cocaine seeking under punishment ( Jonkman et al. 2012b). Compulsive cocaine seeking in a vulnerable subgroup of rats has now been demonstrated in different strains of rats and in different laboratories (Belin et al. 2008, Cannella et al. 2013, Chen et al. 2013, Deroche-Gamonet et al. 2004, Pelloux et al. 2007) using three-criteria or seeking-under-threat-of-punishment procedures. Chen and colleagues (2013) showed that in vivo optogenetic stimulation of the prelimbic cortex decreased compulsive cocaine seeking in compulsive animals, whereas the 80% subpopulation of rats that had suppressed their cocaine seeking during punishment subsequently increased their cocaine seeking under punishment (i.e., became compulsive) after optogenetic inhibition of the prelimbic cortex. These data (Chen et al. 2013, Jonkman et al. 2012a, Pelloux et al. 2012), together with the demonstration of anaplasticity in NAcb neurons in three-criteria “addicted” rats (Kasanetz et al. 2010), suggest altered corticostriatal mechanisms and disrupted top-down or inhibitory control in compulsive cocaine seeking.
In general, the changes in frontal brain function are consistent with impairments in decisionmaking cognition in chronic drug abusers (Rogers et al. 1999), and these impairments resemble in some ways the effects of frontal lesions in clinical patients (see also Bechara et al. 2001, Clark et al. 2008). Deficits in decision making can be due to disruption of several distinct contributory processes. The ventromedial PFC (and medial OFC) is implicated in reward-related processing in fMRI studies (both in the anticipation of reward and in its reinforcing outcome). Damage to the ventromedial PFC in humans causes impairments in the assessment of value and choice outcomes, as well as gross impairments in decision-making tasks such as the Iowa Gambling Task and the Cambridge Gambling Task (Bechara et al. 2001, Clark et al. 2008). Functional imaging of stimulant and opiate abusers has shown changes in the way the OFC is activated during risky decision making (Ersche et al. 2005). In addition to impairments in decision making, drug abusers are commonly impaired in several facets of “cold” executive function, including working memory, cognitive flexibility, and response inhibition (Friedman et al. 2006, Ornstein et al. 2000, Rogers & Robbins 2001). Of these, impairments of inhibitory response control are of obvious interest as potentially leading to relapse, impulsivity, and compulsion (Morein-Zamir & Robbins 2014).
Somatic markers are essentially interoceptive Pavlovian cues that can both elicit conditioned responses and contribute to PIT. Such cues, as well as more readily identified exteroceptive CSs, are well known to elicit drug-seeking behavior and concomitant subjective craving, as well as aversive withdrawal phenomena, both mediated in part via limbic-striatal circuitry including the amygdala. This may explain the remarkable observation of a blockade of craving in nicotine-dependent individuals following damage to the insula caused, for example, by strokes (Naqvi et al. 2007). A subsequent analysis of neuroimaging studies by Naqvi & Bechara (2009) has provided broad support for this original observation over a number of drug classes, including nicotine, cocaine, alcohol, and heroin. Of the 16 studies examined, the insula was the only brain region to be consistently activated by urges to seek drugs, although the anterior cingulate and OFC were often activated, too. Animal studies involving inactivation of the insula have suggested a causal role in an animal model of craving (Contreras et al. 2007).
In humans, the anterior cingulate and inferior frontal cortex (especially in the right hemisphere) are generally considered to be cortical components of a neural circuit mediating inhibitory response control, which also includes the striatum, subthalamic nucleus, and supplementary motor cortex. The impairments are most often quantified in terms of go/no-go or stop-signal reaction time performance (Aron et al. 2014) but may have an obvious influence on decision making, especially decision making involving conflict or the need to reflect on information processing. Functional connectivity studies of chronic stimulant abusers, in parallelwith studies of patientswith obsessivecompulsive disorder (OCD), have shown that frontal zones of connectivity including the OFC are correlated in both cases with measures of compulsivity [the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) and the Obsessive Compulsive Drug Use Scale (OCDUS)] in the stimulant abusers (Meunier et al. 2012). Such studies encourage the notion that compulsivity associated with a neural circuit regulated by the OFC may be a general construct of neuropsychiatric disorders, including addiction. These cortical regions are also implicated in the production of compulsive stimulant drug seeking in rats that results in adverse consequences such as electric footshock (Chen et al. 2013; Pelloux et al. 2012, 2013).
The high-impulsive rat phenotype has been subsequently refined. These rats do not exhibit obvious alterations in appetitive Pavlovian conditioning (autoshaping or sign tracking) and are not especially susceptible to novelty stress or anxiety (Molander et al. 2011) or to impairments on the rodent stop-signal reaction time task, another index of impulsive action (Robinson et al. 2009). They do, however, exhibit steeper reward discounting compared to controls, enhanced intake of nicotine or sucrose, and a mild preference for novelty (see review by Dalley et al. 2011). Other studies (Perry&Carroll 2008) have also shown that changes in the temporal discounting of reward, sometimes defined as impulsive choice, can be predictive of future drug self-administration.Most recently, a similar high-impulsive phenotype based on premature responding in the 5-CSRTT has been shown in the ethanol-preferring B6 strain of mice (Sanchez-Roige et al. 2014), although the high-impulsive rat does not exhibit binge self-administration of heroin (McNamara et al. 2010), indicating that this relationship does not exist for all major drugs of abuse.
Page updated
Report abuse