Stahl's Essential Psychopharmacology: Neuroscientific Basis and Practical Applications, 4th Ed.

Chapter 14. Impulsivity, compulsivity, and addiction

   Overview of impulsive–compulsive disorders

   Neurocircuitry and the impulsive–compulsive disorders

    The mesolimbic dopamine circuit as the final common pathway of reward

   Substance addictions




    Sedative hypnotics




    Club drugs and others

   Obesity as an impulsive–compulsive disorder

   Impulsive–compulsive disorders of behavior

    Obsessive–compulsive disorder


Recent advances in understanding the neurocircuitry of impulsivity and compulsivity has led to the notion that many different psychiatric disorders share these two dimensions of psychopathology. Here we take a look not only at drug addiction, the best-known set of disorders in this category, but also briefly at other “impulsive–compulsive disorders” including obsessive–compulsive disorder (OCD), trichotillomania, gambling, aggression, obesity, and other disorders thought to be related in part to inefficient information processing in prefrontal cortex/striatal circuitry (Figure 14-1 and Table 14-1). Earlier chapters have already touched upon impulsivity in attention deficit hyperactivity disorder (ADHD, discussed in Chapter 12) and in bipolar mania (discussed in Chapters 6 and 7), and the discussion in this chapter applies to impulsivity in these disorders as well.

Figure 14-1. Impulsive–compulsive disorder construct. Impulsivity and compulsivity are seen in a wide variety of psychiatric disorders. Impulsivity can be thought of as the inability to stop the initiation of actions and involves a brain circuit centered on the ventral striatum, linked to the thalamus (T), to the ventromedial prefrontal cortex (VMPFC), and to the anterior cingulate cortex (ACC). Compulsivity can be thought of as the inability to terminate ongoing actions and hypothetically is centered on a different brain circuit, namely the dorsal striatum, thalamus (T), and orbitofrontal cortex (OFC). Impulsive acts such as drug use, gambling, and obesity can eventually become compulsive due to neuroplastic changes that engage the dorsal habit system and theoretically cause impulses in the ventral loop to migrate to the dorsal loop.

Table 14-1 Possible categorization of impulsivity and compulsivity endophenotypes as impulsive–compulsive disorders

Obsessive–compulsive related spectrum disorders

Substance/behavioral addictions

Disruptive/impulse control


Obsessive–compulsive disorder (OCD)

Hair pulling (trichotillomania; TTM)

Skin picking

Body dysmorphic disorder (BDD)


Tourette’s syndrome/tic disorders

Stereotyped movement disorders

Autism spectrum disorders



Drug addiction


Internet addiction

Food addiction (binge eating, obesity)

Compulsive shopping



Intermittent explosive disorder

Impulsive violence

Borderline personality disorder

Self harm/parasuicidal behavior

Antisocial behavior

Conduct disorder

Oppositional defiant disorder





Here we review the hypothetical shared neurobiology of impulsive–compulsive disorders and discuss the treatments that are available for some of these conditions. Although serotonergic treatments of OCD are well known, psychopharmacologists have generally been reluctant to embrace therapeutics for substance abuse, whereas family physicians and subspecialty substance abuse experts use the available psychopharmacological treatments more frequently. Perhaps the lack of highly effective psychopharmacologic treatments for many impulsive–compulsive disorders has led to a certain amount of therapeutic nihilism towards psychopharmacologic approaches to these conditions. Nevertheless, an explosion of neurobiological understanding of the symptom dimensions of impulsivity and compulsivity now sets the stage for novel therapeutic interventions to be discovered in the future, making it worthwhile understanding contemporary neurobiological formulations of addiction, compulsions, and impulsivity. Full clinical descriptions and formal criteria for how to diagnose the numerous known diagnostic entities should be obtained by consulting standard reference sources.

Overview of impulsive–compulsive disorders

Impulsivity and compulsivity are proposed as endophenotypes, namely symptoms linked to specific brain circuits and that are present trans-diagnostically as a dimension of psychopathology that cuts across many psychiatric disorders (Table 14-1Figure 14-1). Simply put, impulsivity and compulsivity are both symptoms that result from the brain having a hard time saying “no.” In fact, these two symptom constructs can perhaps be best differentiated by how they both fail to control responses: impulsivity as the inability to stop initiating actions, and compulsivity as the inability to terminate ongoing actions. Both impulsivity and compulsivity are therefore forms of cognitive inflexibility.

More precisely, impulsivity is defined as acting without forethought; the lack of reflection on the consequences of one’s behavior; the inability to postpone reward with preference for immediate reward over more beneficial but delayed reward; a failure of motor inhibition, often choosing risky behavior; or (less scientifically) lacking the willpower not to give in to temptations (see definitions in Table 14-2).

Table 14-2 Definitions of key terms


Self-administration of any drug in a culturally disapproved manner that causes adverse consequences.


A behavioral pattern of drug abuse characterized by overwhelming involvement with the use of a drug (compulsive use), the securing of its supply, and a high tendency to relapse after discontinuation.


Repetitive actions inappropriate to the situation that persist, that have no obvious relationship to the overall goal, and that often result in undesirable consequences; behavior that results in perseveration in responding in the face of adverse consequences; perseveration in responding in the face of incorrect responses in choice situations or persistent reinitiation of habitual acts.

Cross-tolerance and cross-dependence

The ability of one drug to suppress the manifestations of physical dependence produced by another drug and to maintain the physically dependent state.


The physiological state of adaptation produced by repeated administration of certain drugs such as alcohol, heroin, and benzodiazepines when they are abruptly discontinued, and are associated with physical drug withdrawal distinct from the motivational changes of acute withdrawal and protracted abstinence, which is part of addiction.


Responses triggered by environmental stimuli regardless of the current desirability of the consequences. This conditioned response to a stimulus has been reinforced and strengthened either by past experience with reward (positive reinforcement) or by the omission of an aversive event (negative reinforcement).


The tendency to act prematurely without foresight; actions which are poorly conceived, prematurely expressed, unduly risky, or inappropriate to the situation and that often result in undesirable consequences; predisposition toward rapid, unplanned responses to internal and external stimuli without regard for the negative consequences of those reactions to themselves or others. Impulsivity is often measured in two domains: the choice of a small, immediate reward over a larger delayed reward, or the inability to inhibit behavior to change the course of action or to stop a response once it is initiated.


The exaggerated expression of the original condition sometimes experienced by patients immediately after cessation of an effective treatment.


The tendency of a pleasure-producing drug to lead to repeated self-administration.


The reoccurrence, upon discontinuation of an effective medical treatment, of the original condition from which the patient suffered.


Tolerance has developed when, after repeated administration, a given dose of a drug produces a decreased effect, or, conversely, when increasingly larger doses must be administered to obtain the effects observed with the original use.


The psychologic and physiologic reactions to abrupt cessation of a dependence-producing drug.

On the other hand, compulsivity is defined as actions inappropriate to the situation but which nevertheless persist, and which often result in undesirable consequences. In fact, compulsions are characterized by the inability to adapt behavior after negative feedback. Habits are a type of compulsion, and can be seen as responses triggered by environmental stimuli regardless of the current desirability of the consequences of that response (Table 14-2). Habits can be seen as conditioned responses (such as drug seeking, food seeking, gambling) to a conditioning stimulus (such as being around people or places or items associated with drugs, food, or gambling in the past) that have been reinforced and strengthened either by past experience with reward (positive reinforcement) or with the omission of an aversive event (loss of the negative reinforcement that comes from withdrawal or craving). Whereas goal-directed behavior is mediated by knowledge of and desire for the consequences, in contrast, habits are controlled by external stimuli through stimulus–response associations that are stamped into brain circuits through behavioral repetition and formed after considerable training, can be automatically triggered by stimuli, and are defined by their insensitivity to their outcomes. Given that goal-directed actions are relatively cognitively demanding, for daily routines it can be adaptive to rely on habits that can be performed with minimal conscious awareness. However, habits can also represent severely maladaptive perseveration of behaviors (Figure 14-1Table 14-1).

Neurocircuitry and the impulsive–compulsive disorders

Why can’t impulses and compulsions be stopped in various psychiatric disorders (Table 14-1Figure 14-1)? The answer may lie in a problem in cortical circuits that normally suppress these behaviors. An oversimplification of this notion is that impulsivity and compulsivity are hypothetically neurobiological drives that are “bottom-up,” with impulsivity coming from the ventral striatum, compulsivity coming from the dorsal striatum, and different areas of prefrontal cortex acting “top-down” to suppress these drives (Figures 14-1 through 14-3). Inhibitory volitional control is thus exerted top-down by cortical mechanisms, implying that impulsivity and compulsivity could result from a relaxation of this control. According to this formulation of impulsivity and compulsivity, behavioral output is thus controlled by a balance between dual and sometimes competing neurobehavioral systems. What actually happens depends upon the balance between “top-down” and “bottom-up,” with both impulsivity and compulsivity being either caused by a failure of response inhibition systems (i.e., inadequate top-down cognitive control: Figures 14-1through 14-3) or the result of too much pressure coming bottom-up from the ventral striatum for impulsivity (Figure 14-2) or from the dorsal striatum for compulsivity (Figure 14-3).

Figure 14-2. Circuitry of impulsivity and reward. The “bottom-up” circuit that drives impulsivity (shown in pink) is a loop with projections from the ventral striatum to the thalamus, from the thalamus to the ventromedial prefrontal cortex (VMPFC), and from the VMPFC back to the ventral striatum. This circuit is usually modulated “top-down” from the prefrontal cortex (PFC). If this top-down response inhibition system is inadequate or is overcome by activity from the bottom-up ventral striatum, impulsive behaviors may result.

Figure 14-3. Circuitry of compulsivity and motor response inhibition. The “bottom-up” circuit that drives compulsivity (shown in pink) is a loop with projections from the dorsal striatum to the thalamus, from the thalamus to the orbitofrontal cortex (OFC), and from the OFC back to the dorsal striatum. This habit circuit can be modulated “top-down” from the OFC, but if this top-down response inhibition system is inadequate or is overcome by activity from the bottom-up dorsal striatum, compulsive behaviors may result.

Neuroanatomically, impulsivity and compulsivity are seen as engaging different neuronal loops: impulsivity as an action–outcome ventrally dependent learning system (Figure 14-2) and compulsivity a habit system that is dorsal (Figure 14-3). Many behaviors start out as impulses in the ventral loop of reward and motivation (Figure 14-2). Over time, however, some of these behaviors migrate dorsally (Figure 14-3) due to a cascade of neuroadaptations and neuroplasticity that engage the habit system by means of which an impulsive act eventually becomes compulsive (Figure 14-1). These spirals of information from one neuronal loop to another also appear to involve regulatory input from hippocampus and amygdala and other areas of prefrontal cortex (Figure 14-4).

Figure 14-4. Spiraling circuits of impulsivity and compulsivity. The progression from occasional, impulsive drug use to compulsive use and addiction involves both the dysregulation of bottom-up reward circuits and insufficient top-down inhibition of these circuits. The amygdala and hippocampus provide regulatory input to this system as well. ACC, anterior cingulate cortex; DLPFC, dorsolateral prefrontal cortex; OFC, orbitofrontal cortex; T, thalamus; VMPFC, ventromedial prefrontal cortex.

A well-known example of ventral to dorsal migration is drug addiction. Although initial drug use is thought to be voluntary and linked to trait impulsivity, drug abusers gradually lose control over drug-seeking and drug-taking behavior, which becomes compulsive (Figure 14-5). Impulses to take drugs or to perform certain behaviors initially give a “high” or at least great pleasure and satisfaction (Figure 14-5). If this happens infrequently enough so as notto trigger neuroplastic cascades from ventral to dorsal, this remains under relative control, and can be seen as being occasionally “naughty” (Figure 14-5). However, impulsive drug use or impulsive behaviors repeated too often may progress to compulsive use driven by a desire to reduce the distressing symptoms of withdrawal that develop over time as the drug/behavior is repeated many times (Figure 14-5). Individuals with drug or behavioral addictions experience tension and arousal in anticipation of performing the behavior but dysphoric mood (but no physiological withdrawal) when prevented from performing the behavior or taking the drug. The pleasure and gratification that the drug/behavior initially causes, however, diminish over time, perhaps requiring increasing doses (e.g., gambling higher dollar amounts, more amounts or frequency of drug ingestion) in order to achieve the same effects (akin to tolerance) (Figure 14-5).

Figure 14-5. Shifting from impulsivity to compulsivity. Drug addiction provides a good example of the shift from impulsivity to compulsivity that comes with migration from ventral to dorsal circuits. The impulse to take a drug initially leads to great pleasure and satisfaction (a “high”). If this happens infrequently, the behavior may be a bit “naughty” but will not necessarily progress to compulsivity. With chronic substance use, compulsivity may develop as an individual’s drive turns from seeking pleasure to seeking relief from distressing symptoms of withdrawal and anticipation of obtaining the drug.

Maybe the first dose of any drug will always be the best one, most reinforcing and without penalty. However, individuals do not usually just take one dose of a drug, or perform satisfying behaviors just once. High impulsivity predisposes to the development of compulsions and is predictive of over-reliance on habit learning. Accelerated habit formation may underlie the transition in individuals who have high impulsivity to compulsions and habits. Compulsivity is clearly a maladaptive perseveration of behavior. This is not so much being naughty and giving into temptations (Figure 14-5) as being more like one of Pavlov’s dogs with a mindless involuntary conditioned compulsive response, with willpower being totally inadequate to interrupt the potentially destructive perseverations of behavior once they have become compulsive habits (Table 14-2 and Figure 14-5).

The mesolimbic dopamine circuit as the final common pathway of reward

All drugs that can lead to addiction increase dopamine (DA) in ventral striatum, also called the nucleus accumbens. This area of the brain is familiar to readers, as it is the same area discussed in Chapter 4 on psychosis and also known as the mesolimbic dopamine pathway hypothesized to be overly active in psychosis and to mediate the positive symptoms of schizophrenia (see Figures 4-124-134-304-31). The final common pathway of reinforcement and reward in the brain is also hypothesized to be this same mesolimbic dopamine pathway (Figure 14-6). Some even consider this to be the “center of hedonic pleasure” of the brain and dopamine to be the “neurotransmitter of hedonic pleasure.” There are many natural ways to trigger your mesolimbic dopamine neurons to release dopamine, ranging from intellectual accomplishments to athletic accomplishments, to enjoying a good symphony, to experiencing an orgasm. These are sometimes called “natural highs” (Figure 14-6). The inputs to the mesolimbic pathway that mediate these natural highs include a most incredible “pharmacy” of naturally occurring substances ranging from the brain’s own morphine/heroin (endorphins), to the brain’s own marijuana (anandamide), to the brain’s own nicotine (acetylcholine), to the brain’s own cocaine and amphetamine (dopamine itself) (Figure 14-7).

Figure 14-6. Dopamine is central to reward. Dopamine (DA) has long been recognized as a major player in the regulation of reinforcement and reward. Specifically, the mesolimbic pathway from the ventral tegmental area (VTA) to the nucleus accumbens seems to be crucial for reward. Naturally rewarding activities, such as achieving major accomplishments or enjoying a good meal, can cause fast and robust increases in DA in the mesolimbic pathway. Drugs of abuse also cause DA release in the mesolimbic pathway. In fact, drugs of abuse can often increase dopamine in a manner that is more explosive and pleasurable than that which occurs naturally. Unfortunately, unlike a natural high, the activation caused by drugs of abuse can eventually cause changes in reward circuitry that are associated with a vicious cycle of drug preoccupation, craving, addiction, dependence, and withdrawal. This conceptualization has similarities to many impulsive–compulsive disorders such as pathological gambling. That is, individuals with these disorders experience tension and arousal in anticipation of performing the behavior and dysphoric mood (but no physiological withdrawal) when prevented from performing the behavior. In addition, the pleasure and gratification that is initially experienced when performing the behavior seems to diminish over time, perhaps requiring increasing “doses” (e.g., gambling higher dollar amounts) in order to achieve the same effects (akin to tolerance).

Figure 14-7. Neurotransmitter regulation of mesolimbic reward. The final common pathway of reward in the brain is hypothesized to be the mesolimbic dopamine pathway. This pathway is modulated by many naturally occurring substances in the brain in order to deliver normal reinforcement to adaptive behaviors (such as eating, drinking, sex) and thus to produce “natural highs,” such as feelings of joy or accomplishment. These neurotransmitter inputs to the reward system include the brain’s own morphine/heroin (i.e., endorphins such as enkephalin), the brain’s own cannabis/marijuana (i.e., anandamide), the brain’s own nicotine (i.e., acetylcholine), and the brain’s own cocaine/amphetamine (i.e., dopamine itself), among others. The numerous psychotropic drugs of abuse that occur in nature bypass the brain’s own neurotransmitters and directly stimulate the brain’s receptors in the reward system, causing dopamine release and a consequent “artificial high.” Thus alcohol, opioids, stimulants, marijuana, benzodiazepines, sedative hypnotics, hallucinogens, and nicotine all affect this mesolimbic dopaminergic system. PFC, prefrontal cortex; PPT/LDT, pedunculopontine tegmental and laterodorsal tegmental nuclei.

The numerous psychotropic drugs of abuse also have a final common pathway of causing the mesolimbic pathway to release dopamine, often in a manner more explosive and pleasurable than that which occurs naturally. Also, it now appears that potentially maladaptive behaviors as well as drugs can result in the release of dopamine that in turn stimulates the reward system (Figure 14-7). These are included in the impulsive–compulsive disorder construct (Figure 14-1) and include behaviors such as gambling, using the internet, shopping, and even eating. Drugs bypass the brain’s own neurotransmitters and directly stimulate the brain’s own receptors for these drugs, causing dopamine to be released. Since the brain already uses neurotransmitters that resemble drugs of abuse, it is not necessary to earn your reward naturally since you can get a much more intense reward in the short run and upon demand from a drug of abuse than you can from a natural high with the brain’s natural system. However, unlike a natural high, a drug-induced reward can start a cascade of neuroadaptation in the ventral striatum loop that migrates to the dorsal striatal loop (Figure 14-1), such that the initial high caused by impulsive early use of a drug leads to withdrawal, craving, and preoccupation with finding drug, thus beginning a vicious cycle of abuse, addiction, dependence, and withdrawal (Figure 14-5).

Substance addictions

Not everyone who takes a drug once gets addicted to it. Why? For one thing, some drugs seem to be intrinsically more addicting than others (Table 14-3). For another, some individuals may be more impulsive by nature or have a genetically dysfunctional reward system. It seems that impulsive traits and a dysfunctional reward system may confer a propensity towards drug use and abuse, and when drugs are ingested frequently, impulsive drug use can recruit the involvement of the habit system perhaps in some individuals more readily than in others, triggering neuroplasticity in the compulsivity circuit, which hypothetically is the means by which drug ingestion eventually becomes compulsive in some individuals (Figures 14-1 and 14-5).

Table 14-3 How addicting are different substances?

Probability of becoming dependent when you have tried a substance at least once




















The speed with which a drug enters the brain dictates the degree of the subjective “high” (Figure 14-8). This may be the reason why drugs that are inhaled, snorted, or injected, thus entering the brain in a sudden explosive manner, are usually much more reinforcing than when those same drugs are taken orally, where speed of entry to the brain is considerably slowed by the process of gastrointestinal absorption. Cocaine is not even active orally, so users have learned over the years to take it intranasally – where the drug rapidly enters the brain directly, bypassing the liver, and thus can have a more rapid onset this way than even with intravenous administration. The most rapid and robust way to deliver drugs to the brain is to smoke those that are compatible with this route of administration, as this avoids first-pass metabolism through the liver and is somewhat akin to giving the drug by intra-arterial/intra-carotid bolus via immediate absorption across the massive surface area of the lung. The faster the drug’s entry into the brain, the stronger are its reinforcing effects, probably because this form of drug delivery triggers phasic DA firing, the type associated with reward and saliency (see Chapter 12 for discussion, and Figures 12-1012-29, and 12-31).

Figure 14-8. Dopamine, pharmacokinetics, and reinforcing effects. Acute drug use causes dopamine (DA) release in the striatum. However, the reinforcing effects of the drug are largely determined not only by the presence of DA but also by the rate at which DA increases in the brain, which in turn is dictated by the speed at which the drug enters and leaves the brain, in this case targeting the dopamine transporter (DAT). An abrupt and large increase in DA (such as that caused by drugs of abuse that block DAT) mimic the phasic DA firing associated with conveying information about reward and saliency. The rate of drug uptake is subject to the route of administration, with intravenous administration and inhalation producing the fastest drug uptake, followed by snorting. In addition, different drugs of abuse have different “reward values” (i.e., different rates at which they increase DA) based on their individual mechanisms of action.

On the other hand, some of these very same stimulants taken at low doses orally, especially within controlled-release formulations that minimize peak absorption, slow the rate of absorption, and prolong the duration of drug exposure, are not particularly reinforcing, but instead are therapeutic agents for treating ADHD, as discussed in Chapter 12 and illustrated in Figures 12-29 and 12-30. As discussed in Chapter 12, hypothetically, stimulants administered in this slow-release manner act to “tune” inefficient brain circuits by targeting the prefrontal cortex, enhancing tonic dopamine firing for motivation and attention, and reducing impulses and hyperactivity, while allowing sufficient phasic dopamine firing for learning and for facilitating appropriate goal-directed behaviors/reward (Figures 12-10 through 12-31). Although therapeutic actions of stimulants are thought to be directed at the prefrontal cortex to enhance both norepinephrine and dopamine neurotransmission there (Figures 12-13 through 12-18; Figure 12-31), the reinforcing effects and abuse of stimulants are thought to be directed at reward circuits, especially at dopamine release from mesolimbic dopamine neurons in the nucleus accumbens (Figure 12-31).

It turns out that in the long run it is not the reward of drug, but the anticipation of the reward, that is associated with drug seeking, or food seeking, or seeking numerous other types of situations involved in a wide range of impulsive–compulsive disorders (Table 14-1). Dopamine neurons actually stop responding to the primary reinforcer (i.e., the drug, the food, gambling) and instead begin to respond to the conditioned stimulus (i.e., the sight of the drug, the refrigerator door, the gambling casino). Conditioned responses underlie craving and compulsive use, and the increased dopamine migrates to the dorsal striatum (Figures 14-1 and 14-4). Drugs and behaviors may initially lead to dopamine increase in the ventral striatum and reward (Figures 14-114-214-414-614-7), but with repeated administration, as habits develop, dopamine increases shift from the drug/behavior to the conditioned response/environmental trigger, as the dopamine increases shift from the ventral striatum/nucleus accumbens (Figure 14-2) to the dorsal striatum (Figure 14-3).

Dopamine is associated with motivation, and the motivation to procure drugs is the hallmark of addiction. Drug seeking and drug taking become the main motivational drive when one is addicted, and thus the addicted subject is aroused and motivated when seeking to procure the drug, but is withdrawn and apathetic when exposed to non-drug-related activities (Figures 14-5 and 14-8). What starts out as increased DA release leading to increased ventral striatum and anterior cingluate cortex (ACC) activity with reward may end up as a compulsive drive with escalating dosing in an attempt to get increased reward stimulation to restore a resultant DA deficiency. The discrepancy between expectation for drug effects and the blunted DA effects maintains drug taking in an attempt to achieve the expected reward. High doses of stimulants can cause tremor, emotional lability, restlessness, irritability, panic, and repetitive stereotyped behavior. At even higher repetitive doses, stimulants can induce paranoia and hallucinations, with hypertension, tachycardia, ventricular irritability, hyperthermia, and respiratory depression. In overdose, stimulants can cause acute heart failure, stroke, and seizures.

Not only methylphenidate and amphetamines, but also cocaine are all inhibitors of the dopamine transporter (DAT) and the norepinephrine transporter (NET) (see discussion in Chapter 12, and Figures 12-25and 12-28). Cocaine also inhibits the serotonin transporter (SERT) and is also a local anesthetic, which Freud himself exploited to help dull the pain of his cancer of the jaw and mouth. He may have also exploited the second property of the drug, which is to produce euphoria, reduce fatigue, and create a sense of mental acuity due to inhibition of dopamine reuptake at the dopamine transporter, at least for a while, until drug-induced reward is replaced by drug-induced compulsivity.

Although there are no approved treatments for stimulant addicts, in the future there may be a cocaine vaccine that removes the drug before it reaches the brain so there are no more reinforcing effects that accompany drug ingestion. Theoretically, it may also be possible to administer intravenously a long-acting form of the enzyme cocaine esterase that destroys cocaine before it can exert its reinforcing effects, as has been shown in animal models. Naltrexone, a µ-opioid antagonist approved for the treatment of both opioid and alcohol addiction, is also being investigated for patients with stimulant addiction, particularly for those patients with polydrug dependence on both the opioid heroin and the stimulant amphetamine. Buprenorphine, a synthetic opioid used for treatment of pain and for opioid addiction, stimulates as a partial agonist both µ- and κ-opioid receptors, and can decrease cocaine use in opioid addicts. It is also being studied in combination with naltrexone for cocaine addicts who do not have opioid addiction. The combination results in stimulation only of κ-opioid receptors and not of µ-opioid receptors and may decrease compulsive cocaine self-administration in animals without producing opioid addiction – suggesting that, at least in this case, three drugs might be better than one!


How common is smoking in clinical psychopharmacology practices? Some estimates are that more than half of all cigarettes are consumed by patients with a concurrent psychiatric disorder, and that smoking is the most common comorbidity among seriously mentally ill patients. It is estimated that about 20% of the general population (in the US) smoke, about 30% of people who regularly see general physicians smoke, but that 40–50% of patients in a psychopharmacology practice smoke, including 60–85% of patients with ADHD, schizophrenia, and bipolar disorder. Unfortunately, histories of current smoking are often not carefully taken nor recorded as one of the diagnoses for smokers in mental health practices, and only about 10% of smokers report being offered treatment proactively by psychopharmacologists and other clinicians.

Nicotine acts directly upon nicotinic cholinergic receptors in reward circuits (Figure 14-7). Cholinergic neurons and the neurotransmitter acetylcholine are discussed in Chapter 13 and illustrated in Figures 13-17 through 13-24. Nicotinic receptors are illustrated in Figures 13-20 through 13-22. There are two major subtypes of nicotinic receptors that are known to be present in the brain, the α4β2 subtype and the α7subtype (discussed in Chapter 13 and illustrated in Figure 13-20). Nicotine’s actions in the ventral tegmental area are those that are theoretically linked to addiction, namely at α4β2-nicotinic postsynaptic receptors on dopamine neurons, leading to dopamine release in the nucleus accumbens, and at α7-nicotinic presynaptic receptors on glutamate neurons, which causes glutamate release, and in turn dopamine release in the nucleus accumbens (Figure 14-9). The release-promoting actions of presynaptic α7-nicotinic receptors on glutamate neurons are discussed in Chapter 13 and illustrated in Figure 13-20. Nicotine also appears to desensitize α4β2 postsynaptic receptors on inhibitory GABAergic interneurons in the VTA (Figure 14-9); this also leads to DA release in nucleus accumbens by disinhibiting dopaminergic mesolimbic neurons. Actions of nicotine on postsynaptic α7-nicotinic receptors in the prefrontal cortex may be linked to the pro-cognitive and mentally alerting actions of nicotine, but not to addictive actions.

Figure 14-9. Actions of nicotine. Nicotine directly causes dopamine release in the nucleus accumbens by binding to α4β2-nicotinic postsynaptic receptors on dopamine neurons in the ventral tegmental area (VTA). In addition, nicotine binds to α7-nicotinic presynaptic receptors on glutamate neurons in the VTA, which in turn leads to dopamine release in the nucleus accumbens. Nicotine also seems to desensitize α4β2 postsynaptic receptors on GABA interneurons in the VTA; the reduction of GABA neurotransmission disinhibits mesolimbic dopamine neurons and thus is a third mechanism for enhancing dopamine release in the nucleus accumbens. PFC, prefrontal cortex; PPT/LDT, pedunculopontine tegmental and laterodorsal tegmental nuclei.

The α4β2-nicotinic receptors adapt to the chronic intermittent pulsatile delivery of nicotine in a way that leads to addiction (Figure 14-10). That is, initially these receptors in the resting state are opened by delivery of nicotine, which in turns leads to dopamine release and reinforcement, pleasure, and reward (Figure 14-10A). By the time the cigarette is finished, these receptors become desensitized, so that they cannot function temporarily and thus cannot react either to acetylcholine or to nicotine (Figure 14-10A). In terms of obtaining any further reward, you might as well stop smoking at this point. An interesting question to ask is: how long does it take for the nicotinic receptors to desensitize? The answer seems to be: about as long as it takes to inhale all the puffs of a standard cigarette and burn it down to a butt. Thus, it is probably not an accident that cigarettes are the length that they are. Shorter does not maximize the pleasure. Longer is a waste since by then the receptors are all desensitized anyway (Figure 14-10A).

Figure 14-10. Reinforcement and α4β2 nicotinic receptors. (A) In the resting state α4β2 nicotinic receptors are closed (left). Nicotine administration, as by smoking a cigarette, causes the receptor to open, which in turn leads to dopamine release (middle). Long-term stimulation of these receptors leads to their desensitization, such that they temporarily can no longer react to nicotine (or to acetylcholine); this occurs in approximately the same length of time it takes to finish a single cigarette (right). As the receptors resensitize, they initiate craving and withdrawal due to the lack of release of further dopamine. (B) With chronic desensitization, α4β2 receptors upregulate to compensate. (C) If one continues smoking, however, the repeated administration of nicotine continues to lead to desensitization of all of these α4β2receptors, and thus the upregulation does no good. In fact, the upregulation can lead to amplified craving as the extra receptors resensitize to their resting state.

The problem for the smoker is that when the receptors resensitize to their resting state, this initiates craving and withdrawal due to the lack of release of further dopamine (Figure 14-10A). Another interesting question is: how long does it take to resensitize nicotinic receptors? The answer seems to be: about the length of time that smokers take between cigarettes. For the average one pack per day smoker awake for 16 hours, that would be about 45 minutes, possibly explaining why there are 20 cigarettes in a pack (i.e., enough for an average smoker to keep his or her nicotinic receptors completely desensitized all day long).

Putting nicotinic receptors out of business by desensitizing them causes neurons to attempt to overcome this lack of functioning receptors by upregulating the number of receptors (Figure 14-10B). That, however, is futile, since nicotine just desensitizes all of them the next time a cigarette is smoked (Figure 14-10C). Furthermore, this upregulation is self-defeating because it serves to amplify the craving that occurs when the extra receptors are resensitizing to their resting state (Figure 14-10C).

From a receptor point of view, the goal of smoking is to desensitize all nicotinic α4β2 receptors, get the maximum dopamine release at first, but eventually mostly to prevent craving. Positron emission tomography (PET) scans of α4β2-nicotinic receptors in human smokers confirm that nicotinic receptors are exposed to just about enough nicotine for just about long enough from each cigarette to accomplish this. Craving seems to be initiated at the first sign of nicotinic receptor resensitization. Thus, the bad thing about receptor resensitization is craving. The good thing from a smoker’s point of view is that as the receptors resensitize, they are available to release more dopamine and cause pleasure or suppress craving and withdrawal again.

Treating nicotine dependence is not easy. There is evidence that nicotine addiction begins with the first cigarette, with the first dose showing signs of lasting a month in experimental animals (e.g., activation of the anterior cingulate cortex for this long after a single dose). Craving begins within a month of repeated administration. Perhaps even more troublesome is the finding that the “diabolical learning” that occurs from substance abuse of all sorts including nicotine may be very, very long-lasting once exposure to nicotine is stopped. Some evidence suggests that these changes even last a lifetime, with a form of “molecular memory” to nicotine, even in long-term abstinent former smokers. One of the first successful agents proven to be effective is nicotine itself, but in a route of administration other than smoking: gums, lozenges, nasal sprays, inhalers, and transdermal patches. Delivering nicotine by these other routes does not attain the high levels or the pulsatile blasts that are delivered to the brain by smoking, so they are not very reinforcing. However, these alternative forms of nicotine delivery can help to reduce craving due to a steady amount of nicotine that is delivered, presumably desensitizing an important number of resensitizing and craving nicotinic receptors.

Another treatment for nicotine dependence is varenicline, a selective α4β2-nicotinic acetylcholine receptor partial agonist (Figures 14-11 and 14-12). Figure 14-11 contrasts the effects of nicotinic partial agonists (NPAs) with nicotinic full agonists and nicotinic antagonists on the cation channel associated with nicotinic cholinergic receptors. Nicotinic full agonists include acetylcholine, a short-acting full agonist, and nicotine, a long-acting full agonist. They open the channel fully and frequently (Figure 14-11, left). By contrast, nicotinic antagonists stabilize the channel in the closed state, but do not desensitize these receptors (Figure 14-11, right). NPAs stabilize nicotinic receptors in an intermediate state which is not desensitized and where the channel is open less frequently than with a full agonist, but more frequently than with an antagonist (Figure 14-11, middle).

Figure 14-11. Molecular actions of a nicotinic partial agonist (NPA). Full agonists at α4β2 receptors, such as acetylcholine and nicotine, cause the channels to open frequently (left). In contrast, antagonists at these receptors stabilize them in a closed state, such that they do not become desensitized (right). Nicotinic partial agonists (NPAs) stabilize the channels in an intermediate state, causing them to open less frequently than a full agonist but more frequently than an antagonist (middle).

Figure 14-12. Varenicline actions on reward circuits. Varenicline is a nicotinic partial agonist (NPA) selective for the α4β2 receptor subtype. Its actions at α4β2-nicotinic receptors – located on dopamine neurons and GABA interneurons in the VTA – are all shown. PFC, prefrontal cortex; PPT/LDT, pedunculopontine tegmental and laterodorsal tegmental nuclei.

How addicting is tobacco, and how well do NPAs work to achieve cessation of smoking? About two-thirds of smokers want to quit, one-third try, but only 2–3% succeed long-term. Of all the substances of abuse, some surveys show that tobacco has the highest probability of making you become dependent when you have tried a substance at least once (Table 14-3). It could be argued, therefore, that nicotine might be the most addicting substance known. The good news is that the NPA varenicline triples or quadruples the 1-month, 6-month, and 1-year quit rates compared to placebo; the bad news is that this means only about 10% of smokers who have taken varenicline are still abstinent a year later. Many of these patients are prescribed varenicline for only 12 weeks, which might be far too short a period of time for maximal effectiveness.

Another approach to the treatment of smoking cessation is to try to reduce the craving that occurs during abstinence by boosting dopamine with the norepinephrine–dopamine reuptake inhibitor (NDRI) bupropion (see Chapter 7, and Figures 7-35 through 7-37). The idea is to give back some of the dopamine downstream to the craving postsynaptic D2 receptors in the nucleus accumbens while they are readjusting to the lack of getting their dopamine “fix” from the recent withdrawal of nicotine (Figure 14-13). Thus, while smoking, dopamine is happily released in the nucleus accumbens because of the actions of nicotine on α4β2 receptors on the VTA dopamine neuron (Figure 14-13A). During smoking cessation, resensitized nicotinic receptors no longer receiving nicotine are craving due to an absence of dopamine release in the nucleus accumbens (“where’s my dopamine?” – Figure 14-13B). When the NDRI bupropion is administered, theoretically a bit of dopamine is now released in the nucleus accumbens, making the craving less but usually not eliminating it (Figure 14-13C). How effective is bupropion in smoking cessation? Quit rates for bupropion are about half that of the NPA varenicline. Quit rates for nicotine in alternative routes of administration such as transdermal patches are similar to those of bupropion. Novel approaches to treating nicotine addiction include the investigation of nicotine vaccines and other direct-acting nicotinic cholinergic agents.

Figure 14-13. Mechanism of action of bupropion in smoking cessation. (A) A regular smoker delivers reliable nicotine (circle), releasing dopamine in the limbic area at frequent intervals, which is rewarding to the limbic dopamine D2 receptors on the right. (B) However, during attempts at smoking cessation, dopamine will be cut off when nicotine no longer releases it from the mesolimbic neurons. This upsets the postsynaptic D2 limbic receptors and leads to craving and what some call a “nicotine fit.” (C) A therapeutic approach to diminishing craving during the early stages of smoking cessation is to deliver a bit of dopamine itself by blocking dopamine reuptake directly at the nerve terminal with bupropion. Although not as powerful as nicotine, it does take the edge off and can make abstinence more tolerable.


The famous artist Vincent van Gogh reportedly drank ruinously, some speculating that he self-medicated his bipolar disorder this way, a notion reinforced by his explanation, “If the storm within gets too loud, I take a glass too much to stun myself.” Alcohol may stun but it does not treat psychiatric disorders adaptively long term. Unfortunately, many alcoholics who have comorbid psychiatric disorders continue to self-medicate with alcohol rather than seeking treatment to receive a more appropriate psychopharmacologic agent. In addition to frequent comorbidity with psychiatric disorders, it is estimated that 85% of alcoholics also smoke.

An oversimplified view of alcohol’s mechanism of action is that it enhances inhibition at GABA synapses and reduces excitation at glutamate synapses. Alcohol actions at GABA synapses enhance GABA release via blocking presynaptic GABAB receptors, and also directly stimulate postsynaptic GABAA receptors, especially those of the δ subtype that are responsive to neurosteroid modulation but not to benzodiazepine modulation, either via direct actions or by releasing neurosteroids (Figure 14-14). Delta subtypes of GABAA receptors are discussed in Chapter 9 and illustrated in Figure 9-21. Alcohol also acts at presynaptic metabotropic glutamate receptors (mGluRs) and presynaptic voltage-sensitive calcium channels (VSCCs) to inhibit glutamate release (Figure 14-15). mGluRs are introduced in Chapter 4and illustrated in Figures 4-22 and 4-23. VSCCs and their role in glutamate release are introduced in Chapter 3 and illustrated in Figures 3-22 through 3-24. Alcohol may also have some direct or indirect effects on reducing the actions of glutamate at postsynaptic NMDA receptors and at postsynaptic mGluR receptors (Figure 14-15). Alcohol’s reinforcing effects are theoretically mediated not only by its effects at GABA and glutamate synapses but also by actions at opioid synapses within mesolimbic reward circuitry (Figure 14-15). Opioid neurons arise in the arcuate nucleus and project to the VTA, synapsing on both glutamate and GABA neurons. The net result of alcohol actions on opioid synapses is thought to be the release of dopamine in the nucleus acccumbens (Figure 14-15). Alcohol may do this either by directly acting upon µ-opioid receptors or by releasing endogenous opioids such as enkephalin. These actions of alcohol create the rationale for blocking µ-opioid receptors with antagonists such as naltrexone (Figure 14-16). Figure 14-7 also shows the presence of presynaptic cannabinoid receptors at both glutamate and GABA synapses, where alcohol may have actions. Cannabinoid antagonists such as rimonabant, which blocks CB1 receptors, can reduce alcohol consumption and reduce craving in animals dependent upon alcohol.

Figure 14-14. Binding sites for sedative hypnotic drugs. (A) Benzodiazepines (BZs) and barbiturates both act at GABAA receptors, but at different binding sites. Benzodiazepines do not act at all GABAAreceptors; rather, they are selective for the α1, α2, α3, and α5 subtypes of receptors that also contain γ but not δ subunits. (B) General anesthetics, alcohol, and neurosteroids may bind to other types of GABAAreceptors, particularly those containing δ subunits.

Figure 14-15. Actions of alcohol in the ventral tegmental area (VTA). Opioid neurons synapse in the VTA with GABAergic interneurons and with presynaptic nerve terminals of glutamate neurons. Inhibitory actions of opioids at µ-opioid receptors there cause disinhibition of dopamine release in the nucleus accumbens. Alcohol either directly acts upon µ receptors or causes release of endogenous opioids such as enkephalin. Alcohol also acts at presynaptic metabotropic glutamate receptors (mGluRs) and presynaptic voltage-sensitive calcium channels (VSCCs) to inhibit glutamate release. Finally, alcohol enhances GABA release by blocking presynaptic GABAB receptors and through direct or indirect actions at GABAA receptors.

Figure 14-16. Actions of naltrexone in the ventral tegmental area (VTA). Opioid neurons form synapses in the VTA with GABAergic interneurons and with presynaptic nerve terminals of glutamate neurons. Alcohol either acts directly upon µ receptors or causes release of endogenous opioids such as enkephalin; in either case, the result is increased dopamine release to the nucleus accumbens. Naltrexone is a µ-opioid receptor antagonist; thus it blocks the pleasurable effects of alcohol mediated by µ-opioid receptors.

Several therapeutic agents exploit the known pharmacology of alcohol and are approved for treating alcohol dependence. One of these, naltrexone, blocks µ-opioid receptors (Figure 14-16). As for opioid abuse, µ-opioid receptors theoretically also contribute to the euphoria and “high” of heavy drinking. It is therefore not surprising that a µ-opioid antagonist would block the enjoyment of heavy drinking and increase abstinence by its actions upon reward circuitry (Figure 14-16). This theory is supported by clinical trials, which show that naltrexone not only increases the chances of attaining complete abstinence from alcohol, but also reduces “heavy drinking” (defined as five or more drinks per day for a man and four or more for a woman).

Outcomes for patients with alcohol dependence who take naltrexone may be more favorable when the form of naltrexone that is administrated is given once monthly by intramuscular injection, called XR-naltrexone. This may be due to the fact that this method of drug administration forces compliance for at least a month. Monthly rather than daily drug administration may be just what the reward circuitry needs for someone with a substance-abuse problem. As discussed earlier in this chapter, patients addicted to various substances lose their ability to make rational decisions, and instead respond immediately and impulsively to the desire to seek drugs, and have vast capacity for denial of the maladaptive nature of their compulsive decisions. It is hard enough to get a patient with a substance-abuse disorder to enter treatment or take medications at all, let alone make that person decide every day not only to stay abstinent but also to take a medication. Addiction and human nature being what they are, it is not surprising that patients frequently drop out of treatment and resume substance abuse. If you drink when you take naltrexone, the opioids released do not lead to pleasure, so why bother drinking? Some patients may also of course say, why bother taking naltrexone? – and relapse back into drinking alcohol. However, if you have been given an injection that lasts for a month, and have an irresistible impulse to drink, and you “slip” and start to drink, you are not able to discontinue your naltrexone. Thus, if you “drink over” your naltrexone, you may discover that you do not get the buzz or enjoyment out of intoxication, and therefore might stop after a few drinks. You might even become abstinent for several days again.

Acamprosate is a derivative of the amino acid taurine and interacts with both the glutamate system, to inhibit it, and with the GABA system, to enhance it, a bit like a form of “artificial alcohol” (compare Figure 14-15 with Figure 14-17). Thus, when alcohol is taken chronically and then withdrawn, the adaptive changes that it causes in both the glutamate system and the GABA system create a state of glutamate overexcitement and even excitotoxicity as well as GABA deficiency. Too much glutamate can cause neuronal damage, as discussed in Chapter 13 and illustrated in Figures 13-28 and 13-29. To the extent that acamprosate can substitute for alcohol in patients during withdrawal, the actions of acamprosate mitigate the glutamate hyperactivity and the GABA deficiency (Figure 14-17). This occurs because acamprosate appears to have direct blocking actions on certain glutamate receptors, particularly mGlu receptors (specifically mGluR5 and perhaps mGluR2). One way or another, acamprosate apparently reduces the glutamate release associated with alcohol withdrawal (Figure 14-17). Actions, if any, at NMDA receptors may be indirect, as are actions at GABA systems, both of which may be secondary downstream effects from acamprosate’s actions on mGlu receptors (Figure 14-17).

Figure 14-17. Actions of acamprosate in the ventral tegmental area (VTA). Acamprosate seems to block glutamate receptors, particularly metabotropic glutamate receptors (mGluRs) and perhaps also N-methyl-D-aspartate (NMDA) receptors. When alcohol is taken chronically and then withdrawn, the adaptive changes that it causes in both the glutamate system and the GABA system create a state of glutamate overexcitation as well as GABA deficiency. By blocking glutamate receptors, acamprosate may thus mitigate glutamate hyperexcitability during alcohol withdrawal.

Disulfiram is the classic drug for treating alcoholism. It is an irreversible inhibitor of aldehyde dehydrogenase and, when alcohol is ingested, results in the build-up of toxic levels of acetaldehyde. This creates an aversive experience with flushing, nausea, vomiting, and hypotension, hopefully conditioning the patient to a negative rather than positive response to drinking. Obviously, compliance is a problem with this agent, and its aversive reactions are occasionally dangerous.

Experimental agents that show some promise in treating alcohol dependence include the anticonvulsant topiramate (discussed in more detail below in the section on obesity), the 5HT3 antagonists (mechanism discussed in Chapter 7 and illustrated in Figure 7-46), and cannabinoid CB1 receptor antagonists. New opioid antagonists such as nalmefene (Selinco) are also in late-stage clinical testing. The subject of how to treat alcohol abuse and dependence is obviously complex, and the psychopharmacological treatments are most effective when integrated with structured therapies such as 12-step programs, a topic which is beyond the scope of this text. Hopefully, clinicians will learn how to better leverage the various treatments for alcoholism that are available today, and determine whether they can be used to treat this devastating illness to attain far better outcomes than are available when no treatment is provided, accepted, or sustained.

Sedative hypnotics

Sedative hypnotics include barbiturates and related agents such as ethclorvynol and ethinamate, chloral hydrate and derivatives, and piperidinedione derivatives such as glutethimide and methyprylon. Experts often include alcohol, benzodiazepines (discussed in Chapter 9), and Z-drug hypnotics (discussed in Chapter 11) in this class as well. The mechanism of action of sedative hypnotics is basically the same as those described in Chapter 9 and illustrated in Figure 9-23 for the action of benzodiazepines: namely, they are positive allosteric modulators (PAMs) for GABAA receptors. Actions of sedative hypnotics are at GABAA receptor sites in reward circuits (Figure 14-7). Molecular actions of all sedative hypnotics are similar, but benzodiazepines and barbiturates seem to work at different sites from each other, and also only on some GABAA receptor subtypes, namely those with α1, α2, α3, or α5 subunits (Figure 14-14). Barbiturates are much less safe in overdose than benzodiazepines, cause dependence more frequently, are abused more frequently, and produce much more dangerous withdrawal reactions. Apparently, the receptor site at GABAA receptors mediating the pharmacologic actions of barbiturates is even more readily desensitized with even more dangerous consequences than the benzodiazepine receptor (Figure 14-14). The barbiturate site must also mediate a more intense euphoria and a more desirable sense of tranquility than the benzodiazepine receptor site. Since benzodiazepines are generally an adequate alternative to barbiturates, psychopharmacologists can help to minimize abuse of barbiturates by prescribing them rarely if ever. In the case of withdrawal reactions, reinstituting and then tapering the offending barbiturate under close clinical supervision can assist the detoxification process.


Opioids act as neurotransmitters released from neurons that arise in the arcuate nucleus and project both to the VTA and to the nucleus accumbens, and release enkephalin (Figure 14-18). Naturally occurring endogenous opioids act upon a variety of receptor subtypes. The three most important receptor subtypes are the µ-, δ-, and κ-opioid receptors (Figure 14-18). The brain makes a variety of its own endogenous opioid-like substances, sometimes referred to as the “brain’s own morphine.” These are all peptides derived from precursor proteins called pro-opiomelanocortin (POMC), proenkephalin, and prodynorphin (Figure 14-18). Parts of these precursor proteins are cleaved off to form endorphins, enkephalins, or dynorphins, stored in opioid neurons, and presumably released during neurotransmission to mediate endogenous opioid-like actions, including a role in mediating reinforcement and pleasure in reward circuitry (Figure 14-7).

Figure 14-18. Endogenous opioid neurotransmitters. Opioid drugs act on a variety of receptors called opioid receptors, the most important of which are µ, δ, and κ. Endogenous opioid-like substances are peptides derived from precursor proteins called POMC (pro-opiomelanocortin), proenkephalin, and prodynorphin. Parts of these precursor proteins are cleaved off to form endorphins, enkephalins, or dynorphins, which are then stored in opioid neurons and presumably released during neurotransmission to mediate reinforcement and pleasure.

Exogenous opioids in the form of pain relievers (such as oxycodone, hydrocodone, and many others) or drugs of abuse (such as heroin) are also thought to act as agonists at µ-, δ-, and κ-opioid receptors, particularly at µ sites. At and above pain-relieving doses, the opioids induce euphoria, which is the main reinforcing property of the opioids. Opioids can also induce a very intense but brief euphoria, sometimes called a “rush,” followed by a profound sense of tranquility which may last several hours, followed in turn by drowsiness (“nodding”), mood swings, mental clouding, apathy, and slowed motor movements. In overdose, these same agents act as depressants of respiration, and can also induce coma. The acute actions of opioids can be reversed by synthetic opioid antagonists, such as naloxone and naltrexone, which compete as antagonists at opioid receptors.

When given chronically, opioids readily cause both tolerance and dependence. Adaptation of opioid receptors occurs quite readily after chronic opioid administration. The first sign of this is the need of the patient to take a higher and higher dose of opioid in order to relieve pain or to induce the desired euphoria. Eventually, there may be little room between the dose that causes euphoria and that which produces toxic effects of an overdose. Another sign that dependence has occurred and that opioid receptors have adapted by decreasing their sensitivity to agonist actions is the production of a withdrawal syndrome once the chronically administered opioid wears off. The opioid antagonists, such as naloxone, can precipitate a withdrawal syndrome in opioid-dependent persons. The opioid withdrawal syndrome is characterized by the patient feeling dysphoria, craving another dose of opioid, being irritable, and having signs of autonomic hyperactivity such as tachycardia, tremor, and sweating. Pilo-erection (“goose-bumps”) is often associated with opioid withdrawal, especially when drug is stopped suddenly (“cold turkey”). This is so subjectively horrible that the opioid abuser will often stop at nothing in order to get another dose of opioid to relieve symptoms of withdrawal. Thus, what may have begun as a quest for euphoria may end up as a quest to avoid withdrawal. Clonidine, an α2-adrenergic agonist, can reduce signs of autonomic hyperactivity during withdrawal and aid in the detoxification process.

Opioid receptors can readapt to normal if given a chance to do so in the absence of additional intake of an opioid. This may be too difficult to tolerate, so reinstituting another opioid, such as methadone, which can be taken orally and then slowly tapered, may assist in the detoxification process. A partial µ-opioid agonist, buprenophine, now available in a sublingual dosage formulation combined with naloxone, can also substitute for stronger full agonist opioids, and then be tapered. It is combined with the opioid antagonist naloxone, which does not get absorbed orally or sublingually, but prevents intravenous abuse, since injection of the combination of buprenorphine plus naloxone results in no high and may even precipitate withdrawal. Agonist substitution treatments are best used in the setting of a structured maintenance treatment program that includes random urine drug screening and intensive psychological, medical, and vocational services. For those who can stop taking opioids for at least 7–10 days so that serious withdrawal symptoms do not occur, long-acting injectable naltrexone can be a highly effective therapy for opioid addicts, since this drug blocks any “cheating” for a month, and prevents the pharmacologic actions of abused opioids at the µ-opioid site, allowing detoxification to proceed even if the patient tries to take an opioid. This is the same drug in the same formulation discussed above and approved for treatment of alcohol abuse.

Not all who abuse opioids are the stereotypical addict who smokes or injects drugs, lives on the streets, and supports himself with crime. There also is a current serious epidemic of oral abuse of prescription opioids by those who are employed or students, and who obtain these drugs either from prescribers or from drug dealers who procure them from prescribers, pharmacies, or online.


You can indeed get stoned without inhaling! Actions of marijuana and its active ingredient Δ9-tetrahydrocannabinol (THC) on reward circuits are at cannabinoid receptors, shown in Figure 14-7, which are the sites where endogenous cannabinoids are utilized naturally as retrograde neurotransmitters. Cannabis preparations are smoked in order to deliver cannabinoids that interact with the brain’s own cannabinoid receptors to trigger dopamine release from the mesolimbic reward system (Figure 14-7). There are two known cannabinoid receptors, CB1 (in brain, coupled via G proteins and modulate adenylate cyclase and ion channels) and CB2 (predominantly in the immune system). CB1receptors may mediate not only marijuana’s reinforcing properties, but also those of alcohol and to some extent those of other psychoactive substances (including possibly some foods). Anandamide is one of the endocannabinoids and a member of a chemical class of neurotransmitter that is not a monoamine, not an amino acid, and not a peptide: it is a lipid, specifically a member of a family of fatty acid ethanolamides. Anandamide shares most but not all of the pharmacologic properties of THC, since its actions at brain cannabinoid receptors are mimicked not only by THC but are antagonized in part by the selective brain cannabinoid CB1 receptor antagonist rimonabant.

In usual intoxicating doses, marijuana produces a sense of well-being, relaxation, a sense of friendliness, a loss of temporal awareness, including confusing the past with the present, slowing of thought processes, impairment of short-term memory, and a feeling of achieving special insights. At high doses, marijuana can induce panic, toxic delirium, and, rarely, psychosis. One complication of long-term use is the “amotivational syndrome” in frequent users. This syndrome is seen predominantly in heavy daily users and is characterized by the emergence of decreased drive and ambition, thus “amotivational.” It is also associated with other socially and occupationally impairing symptoms, including a shortened attention span, poor judgment, easy distractibility, impaired communication skills, introversion, and diminished effectiveness in interpersonal situations. Personal habits may deteriorate, and there may be a loss of insight, and even feelings of depersonalization. Another downside to marijuana is that individuals vulnerable to schizophrenia might precipitate this illness, make its onset earlier, or exacerbate established illness when abusing marijuana, and to a greater extent than with any other abusable drug.


The hallucinogens are a group of agents that act at serotonin synapses in the reward system (Figure 14-19). They produce intoxication, sometimes called a “trip” associated with changes in sensory experiences, including visual illusions and hallucinations, an enhanced awareness of external stimuli and an enhanced awareness of internal thoughts and stimuli. These hallucinations are produced with a clear level of consciousness and a lack of confusion and may be both psychedelic and psychotomimeticPsychedelic is the term for the subjective experience that, due to heightened sensory awareness, one’s mind is being expanded or that one is in union with mankind or the universe and having some sort of a religious experience. Psychotomimetic means that the experience mimics a state of psychosis, but the resemblance between a trip and psychosis is superficial at best. The stimulants cocaine and amphetamine and the club drug phencyclidine (PCP) much more genuinely mimic psychosis (see discussion in Chapter 4). Hallucinogen intoxication includes visual illusions; visual “trails” where the image smears into streaks of its image as it moves across a visual trail; macropsia and micropsia; emotional and mood lability; subjective slowing of time; the sense that colors are heard and sounds are seen; intensification of sound perception; depersonalization and derealization; yet retaining a state of full wakefulness and alertness. Other changes may include impaired judgment, fear of losing one’s mind, anxiety, nausea, tachycardia, increased blood pressure, and increased body temperature. Not surprisingly, hallucinogen intoxication can cause what is perceived as a panic attack, often called a “bad trip.” As intoxication escalates, one can experience an acute confusional state called delirium where the abuser is disoriented and agitated. This can evolve further into frank psychosis with delusions and paranoia.

Figure 14-19. Mechanism of hallucinogens at 5HT2A receptors. The primary action of hallucinogenic drugs such as LSD, mescaline, psilocybin, and MDMA are shown here: namely, agonism of 5HT2Areceptors. Hallucinogens may have additional actions at other serotonin receptors (particularly 5HT1A and 5HT2C) and at other neurotransmitter systems, and MDMA in particular also blocks the serotonin transporter (SERT).

Common hallucinogens include two major classes of agents. The first class of agents resemble serotonin (indole-alkylamines) and include the classical hallucinogens D-lysergic acid diethylamide (LSD), psilocybin, and dimethyltryptamine (DMT) (Figure 14-19). The second class of agents resemble norepinephrine and dopamine and are also related to amphetamine (phenylalkylamines) and include mescaline, 2,5-dimethoxy-4-methylamphetamine (DOM) and others. More recently, synthetic chemists have come up with some new “designer drugs” such as 3,4-methylenedioxymethamphetamine (MDMA) and “Foxy” (5-methoxy-diisopropyltryptamine). These are either stimulants or hallucinogens and produce a complex subjective state sometimes referred to as “ecstacy,” which is also what abusers call MDMA itself. MDMA produces euphoria, disorientation, confusion, enhanced sociability, and a sense of increased empathy and personal insight.

Hallucinogens have rather complex interactions at neurotransmitter systems, but one of the most prominent is a common action as agonists at 5HT2A receptor sites (Figure 14-19). Hallucinogens certainly have additional effects at other 5HT receptors (especially 5HT1A somatodendritic autoreceptors and 5HT2C postsynaptic receptors) and also at other neurotransmitter systems, especially norepinephrine and dopamine, but the relative importance of these other actions is less well known. MDMA also appears to be a powerful inhibitor of the serotonin transporter (SERT) and is also a releaser of serotonin. MDMA and several other drugs structurally related to it may even destroy serotonin axon terminals. However, the action that appears to explain a common mechanism for most of the hallucinogens is the stimulation of 5HT2A receptors.

Hallucinogens can produce incredible tolerance, sometimes after a single dose. Desensitization of 5HT2A receptors is hypothesized to underlie this rapid clinical and pharmacological tolerance. Another unique dimension of hallucinogen abuse is the production of “flashbacks,” namely the spontaneous recurrence of some of the symptoms of intoxication that lasts from a few seconds to several hours but in the absence of recent administration of the hallucinogen. This occurs days to months after the last drug experience, and can apparently be precipitated by a number of environmental stimuli. The psychopharmacological mechanism underlying flashbacks is unknown, but its phenomenology suggests the possibility of a neurochemical adaptation of the serotonin system and its receptors related to reverse tolerance that is incredibly long-lasting. Alternatively, flashbacks could be a form of emotional conditioning embedded in the amygdala and then triggered when a later emotional experience while not taking a hallucinogen nevertheless reminds one of experiences that occurred when intoxicated with a hallucinogen. This could precipitate a whole cascade of feelings that occurred while intoxicated with a hallucinogen. This is analogous to the types of re-experiencing flashbacks that occur without drugs in patients with posttraumatic stress disorder.

Club drugs and others

Phencyclidine (PCP) and ketamine both have actions at glutamate synapses within the reward system (Figure 14-7 and Figure 7-91). They both act as antagonists of NMDA receptors, binding to a site in the calcium channel (see discussion in Chapter 4 and Figure 4-28). Both were originally developed as anesthetics. PCP proved to be unacceptable for this use because it induces a unique psychotomimetic/hallucinatory experience very similar to schizophrenia. The NMDA receptor hypoactivity that is caused by PCP has become a model for the same neurotransmitter abnormalities postulated to underlie schizophrenia (see discussion in Chapter 4 and Figure 4-28). PCP causes intense analgesia, amnesia, delirium, stimulant as well as depressant actions, staggering gait, slurred speech, and a unique form of nystagmus (i.e., vertical nystagmus). Higher degrees of intoxication can cause catatonia (excitement alternating with stupor and catalepsy), hallucinations, delusions, paranoia, disorientation, and lack of judgment. Overdose effects can include coma, extremely high temperature, seizures, and muscle breakdown (rhabdomyolysis).

PCP’s structurally related and mechanism-related analog ketamine is still used as an anesthetic, but causes far less of the psychotomimetic/hallucinatory experience. Nevertheless, some people do abuse ketamine, one of the “club drugs,” and it is sometimes called “special K.” Interestingly, subanesthetic infusions of ketamine have been repeatedly shown to reduce symptoms of depression in unipolar treatment-resistant depression and in bipolar depression and to decrease suicidal thoughts (see discussion in Chapter 7 and Figures 7-90 through 7-93).

Gamma-hydroxybutyrate (GHB) is discussed in Chapter 11 as a treatment for narcolepsy/cataplexy. It is sometimes also abused by individuals wanting to get high or by predators to intoxicate their dates (GHB is one of the “date rape” drugs). The mechanism of action of GHB is as an agonist at its own GHB receptors and at GABAB receptors (illustrated in Figure 11-27).

Inhalants such as toluene are thought to be direct releasers of dopamine in the nucleus accumbens. “Bath salts” are synthetic stimulants that commonly include the active ingredient methylenedioxypyrovalerone (MDPV) but may also contain mephedrone or mehylone. They are also called “plant food” and like other stimulants can have reinforcing effects but also cause agitation, paranoia, hallucinations, suicidality, and chest pain.

Obesity as an impulsive–compulsive disorder

Can you become addicted to food? Can your brain circuits make you eat it? Although food addiction is not yet accepted as a formal diagnosis, it does appear that when external stimuli are triggers for maladaptive eating habits that are performed despite apparent satiety and adverse health consequences, this does define a compulsion and a habit, with the formation of aberrant eating behaviors in a manner that parallels drug addiction (Table 14-4). Compulsive eating in obesity, and in binge eating disorder and bulimia, can be mirrored by compulsive rejection of food as in anorexia nervosa. This chapter does not provide a comprehensive explanation of eating disorders but addresses only aspects of obesity that may in some cases fit with the construct of an impulsive–compulsive disorder and how some of the new treatments for obesity may help such cases.

Table 14-4 Food addiction: is obesity an impulsive–compulsive disorder?

Obesity, appetite, eating and the dimensions of impulsivity/compulsivity

Enhanced reward of food/enhanced motivation and drive to consume food

Increasing amounts of food to maintain satiety, tolerance

Lack of control over eating – cannot stop

Great deal of time spent eating

Conditioning and habits to food and food cues

Distress and dysphoria when dieting

Eating too rapidly or too much when not hungry, to the point of being uncomfortably full

Overeating maintained despite knowledge of adverse physical and psychological consequences caused by excessive food consumption

Eating alone, feeling disgusted with oneself, guilty, or depressed

Binge eating can occur with or without purging

Bulimia is binge eating with self-disgust and purging leading to attempts to prevent weight gain by excessive exercise, induced vomiting, abuse of laxatives, enemas, or diuretics

When is the way you eat a lifestyle choice and when is it an impulsive–compulsive disorder? Obesity is defined by one’s body mass index (BMI U+2265 30), and not by any associated behaviors. Not everyone who is obese has an eating compulsion, since obesity is also related to genetic and lifestyle factors such as exercise, caloric intake, and the foods eaten and their specific content (fat, carbohydrates, vitamins, and other components). It is only those forms of obesity apparently driven by excessive motivational drive for food and mediated by reward circuitry that might be considered impulsive–compulsive disorders (Table 14-4). When exposed to food cues, obese individuals exhibit increased brain activation (compared to lean individuals) in anatomic areas that process palatability, and decreased activation of reward circuits during actual food consumption, analogous to what happens in drug addiction.

Appetite/motivation for eating, and the actual amount of food consumed, can both be influenced by centrally active psychopharmacologic agents in many individuals. For example, several known drugs of abuse reduce appetite, especially stimulants and nicotine. Bupropion, naltrexone, topiramate, and zonisamide have all been observed anecdotally to cause weight loss in patients taking these agents for other reasons. By contrast, marijuana and some atypical antipsychotics (see Chapter 5 and Figure 5-41) actually stimulate appetite and cause weight gain. The neurobiological basis of eating and appetite are clearly linked to the hypothalamus (Figures 14-20A through 14-20G), and to the connections that hypothalamic circuits make to reward pathways (Figures 14-2 through 14-4). Current research is attempting to clarify the role of a long list of key hypothalamic regulators in the control of eating: orexin (which also regulates sleep and is discussed in Chapter 11 and illustrated in Figures 11-21 through 11-23), α-melanocyte-stimulating hormone (α-MSH), neuropeptide Y, agouti-related peptide (all illustrated in Figures 14-20Athrough 14-20G), and many more including leptin, ghrelin, adiponectin, melanin-concentrating hormone, cholecystokinin, insulin, glucagon, cytokines, cocaine- and amphetamine-regulated transcript (CART) peptides, galanin, and others.

The hypothalamus thus serves as the brain center that controls appetite by utilizing a complex set of circuits and regulators. One formulation of how the hypothalamus does this is the notion that there is a major appetite-stimulating pathway whose actions are mediated by two peptides (neuropeptide Y and agouti-related protein) (Figure 14-20A). Opposing this is a major appetite-suppressing pathway whose actions are mediated by pro-opiomelanocortin (POMC) neurons that make the peptide POMC; POMC can be broken down into either β-endorphin or α-melanocyte-stimulating hormone (α-MSH). α-MSH interacts with melanocyte 4 receptors (MC4Rs) to suppress appetite (Figure 14-20A). Weight gain could occur either by excessive activity of the appetite-stimulating pathway, by deficient activity of the appetite-suppressing pathway, or both.

A. Peptides regulate appetite in the hypothalamus. Appetite is regulated by the balance between an appetite-stimulating pathway (on the left) that releases agouti-related peptide (AgRP) and neuropeptide Y (NPY), and an appetite-suppressing pathway (on the right) that releases α-melanocyte-stimulating hormone (α-MSH). The appetite-suppressing neurons make the precursor pro-opiomelanocortin (POMC), which is broken down into α-MSH, which in turn binds to melanocortin 4 receptors (MC4R) to suppress appetite. Shown here is no occupancy of MC4R by α-MSH, and thus stimulation of appetite.

B. Actions of phentermine. Phentermine increases dopamine (DA) and norepinephrine (NE) in the hypothalamus by blocking both the norepinephrine and dopamine reuptake transporters (NET and DAT, respectively). The increased input of DA and NE onto pro-opiomelanocortin (POMC) neurons in the appetite-suppressing pathway partially activates the POMC neurons (shown as a hatched color on the right), causing an increase in α-melanocyte-stimulating hormone (α-MSH) release, which binds to melanocortin 4 receptors (MC4R) to suppress appetite partially.

C. Topiramate potentiates the actions of phentermine. Topiramate hypothetically inhibits the appetite-stimulating pathway on the left by reducing excitatory glutamatergic input and by increasing inhibitory GABA-ergic input (shown as faded neurons on the left). Combining this with phentermine’s actions on the right that stimulate the appetite-suppressing pathway (shown as a hatched color and also shown in Figure 14-20B), this results in a synergistic and enhanced effect on appetite and on weight loss, allowing lower, more tolerable doses of both phentermine and topiramate to be used.

D. Actions of bupropion. The antidepressant and smoking cessation aid bupropion is thought to have effects in the appetite center of the hypothalamus as well. Bupropion increases dopamine (DA) and norepinephrine (NE) in the hypothalamus by blocking both the norepinephrine and dopamine reuptake transporters (NET and DAT, respectively), just as shown in Figure 14-20B, but perhaps less robustly. The increased input of DA and NE onto pro-opiomelanocortin (POMC) neurons in the appetite-suppressing pathway partially activates the POMC neurons (shown as a hatched color on the right) causing an increase in α-melanocyte-stimulating hormone (α-MSH) release, which binds to melanocortin 4 receptors (MC4R) to suppress appetite partially (compare with Figure 14-20B). The actions of bupropion on the appetite-suppressing pathway, however, are mitigated because stimulation of POMC neurons also activates an endorphin/endogenous-opioid-mediated negative feedback loop (also with phentermine and shown in Figure 14-20B).

E. Naltrexone potentiates the actions of bupropion. Both naltrexone and bupropion alone can lead to some weight loss by themselves. However, the combination of naltrexone and bupropion has a synergistic effect on weight loss that surpasses monotherapy with either agent by dual pharmacologic actions on the appetite-suppressing pathway. That is, naltrexone blocks the endorphin/endogenous-opioid-mediated negative feedback loop that normally limits the activation of pro-opiomelanocortin (POMC) neurons in the appetite-suppressing pathway (shown in Figure 14-20D with only a hatched neuron on the right). With this negative feedback removed by administration of naltrexone, bupropion can more readily increase firing of POMC neurons (shown here as a red-hot neuron on the right), leading to highly elevated levels of α-melanocyte-stimulating hormone (α-MSH), which binds more robustly to melanocortin 4 receptors (MC4R) to more potently suppress appetite and cause weight loss.

F. Actions of lorcaserin. The serotonin 5HT2C agonist lorcaserin has recently been approved for the treatment of obesity. Lorcaserin hypothetically binds to 5HT2C receptors on pro-opiomelanocortin (POMC) neurons in the appetite-suppressing pathway, activating POMC neurons and leading to release of α-melanocyte-stimulating hormone (α-MSH), which binds to melanocortin 4 receptors (MC4R) to robustly suppress appetite.

G. Naltrexone potentiates the actions of zonisamide. The anticonvulsant zonisamide has actions in the appetite center of the hypothalamus that are similar to those of topiramate (Figure 14-20C). Zonisamide hypothetically both reduces excitatory glutamatergic input and increases inhibitory GABAergic input onto neurons in the appetite-stimulating pathway, leading to less output of neuropeptide Y (NPY) and agouti-related peptide (AgRP), and decreased appetite stimulation. Naltrexone removes the endorphin/endogenous-opioid-mediated negative feedback that limits activation of pro-opiomelanocortin (POMC) neurons in the appetite-suppressing pathway (Figure 14-20E). With this negative feedback removed, α-melanocyte-stimulating hormone (α-MSH) levels are increased (i.e., disinhibited), leading to appetite suppression. The combination of naltrexone and zonisamide is currently under investigation as a potential treatment for obesity and impulsive–compulsive eating disorders.

Figure 14-20

Several new treatments for obesity, defined as BMI U+2265 30 (or for being overweight, with a BMI U+2265 27 plus diabetes, hypertension, or dyslipidemia), and not defined by compulsive/addictive behaviors, are now approved or in late-stage clinical testing. One novel obesity treatment that targets multiple sites within these hypothalamic appetite pathways is the combination of the stimulant phentermine, already approved as a monotherapy for the treatment of obesity, with the anticonvulsant topiramate (phentermine/topiramate ER, or Qsymia). Phentermine acts much like amphetamine, blocking both the dopamine transporter (DAT) and the norepinephrine transporter (NET) and, at high doses, the vesicular monoamine transporter (VMAT) (see discussion of amphetamine’s mechanism of action in Chapter 12 and illustrated in Figures 12-28 through 12-31). When stimulants like phentermine increase dopamine and norepinephrine in the hypothalamus, they reduce appetite and cause weight loss. One hypothesis is that they do this by stimulating POMC neurons to release α-MSH in the hypothalamus (Figure 14-20B). However, when given by itself in doses adequate to suppress appetite (Figure 14-20B), there are limitations to the use of phentermine. For example, tolerance for phentermine usually develops over time, and the weight often returns. Also, phentermine simultaneously targets dopamine in reward circuits and risks causing abuse or addiction. Additional dose-related noradrenergic effects of phentermine can increase pulse rate and blood pressure and cause cardiovascular complications, especially in vulnerable obese patients with cardiovascular disease.

One solution to these limitations of phentermine monotherapy has been to lower the dose of phentermine yet enhance its actions by adding the agent topiramate. In the combination product phentermine/topiramate ER, the phentermine dose is only about a quarter to half of what is usually prescribed when phentermine is given as a monotherapy for the treatment of obesity. This mitigates potential cardiovascular and reinforcing side effects. By combining it with topiramate, the efficacy of low-dose phentermine is not lost, and in fact it is enhanced because of synergy with topiramate’s mechanisms (Figure 14-20C). Topiramate, observed to reduce weight as a “side effect” when prescribed for the approved treatments of epilepsy or migraine, does so by poorly understood mechanisms, probably related to both boosting inhibitory GABA actions and reducing excitatory glutamate actions via more direct actions on various voltage-gated ion channels (Figure 14-20C). Topiramate also inhibits the enzyme carbonic anhydrase, although what contribution this makes to topiramate’s therapeutic actions in obesity remains unclear. Theoretically, topiramate may act to reduce glutamatergic stimulation and to enhance GABAergic inhibition in the appetite-stimulating pathway (Figure 14-20C), resulting in net inhibition of this pathway. Such an action would synergize with simultaneous activation of the appetite suppressing pathway by phentermine (Figures 14-20B and 14-20C), to produce a more robust and long-lasting result on appetite suppression than with either drug alone. So far, that seems to be the case. Also, the tolerability of topiramate is enhanced by lowering its dose below that used for epilepsy or migraine, or even for its off-label use as a monotherapy for weight loss. Furthermore, topiramate is administered in a controlled-release formulation so that peak plasma drug levels and thus sedation are reduced.

Phentermine/topiramate ER in clinical trials showed dose-related weight loss, from 6% to 9% over placebo, with about two-thirds of obese patients losing at least 5% of their weight (only 20% of obese patients lost this much on placebo) by 12 weeks. Some patients do not respond, of course, and the weight loss can be modest, and long-term outcomes are not known. Also the topiramate component is potentially teratogenic to pregnant patients. However, this combination promises to be very useful in the management of obesity.

Another combination product that targets multiple simultaneous psychopharmacological mechanisms for obesity is bupropion/naltrexone (Contrave), in late clinical testing as of this writing. Bupropion alone has long been observed anecdotally to cause weight loss in some patients (Figure 14-20D). Bupropion is not only a proven antidepressant (see Chapter 7 and Figures 7-35 through 7-37) but is also a proven treatment for smoking cessation (see discussion earlier in this chapter and Figure 14-13), suggesting that bupropion’s therapeutic actions are occurring at least in part within reward pathways. It is thus not surprising that bupropion could have therapeutic actions in disorders related to nicotine addiction, including possibly obesity and food addiction.

Bupropion acts as a norepinephrine–dopamine reuptake inhibitor (NDRI, Chapter 7Figures 7-35 through 7-37). This mechanism is similar but less robust compared to amphetamine (Chapter 12Figures 12-28 to 12-31; and also Figure 14-20D) or phentermine. When these NDRI actions of bupropion occur in the hypothalamus, this hypothetically enhances POMC-neuron-mediated appetite suppression. However, it also activates a β-endorphin/endogenous-opioid-mediated negative feedback pathway that mitigates how much bupropion can activate the POMC neuron (Figure 14-20D).

Preclinical studies show that addition of naltrexone can remove this negative opioid feedback and potentiate bupropion’s ability to increase POMC neuron firing. These observations provide the rational explanation for the synergistic pharmacologic actions on appetite suppression and weight loss that have been observed both in animals and in clinical trials of obesity when naltrexone is combined with bupropion (Figure 14-20E). Naltrexone alone causes a small amount of weight loss in patients taking it for its approved uses for alcohol and opioid addiction. However, when a dose of naltrexone lower than that generally used to treat alcohol or opioid addiction is combined in obese subjects with a dose of bupropion in the general range used to treat depression or for smoking cessation, the combination treatment produces greater weight loss than either monotherapy alone. Notably, subjects treated with the combination appeared to exhibit continued weight loss through week 24, in contrast to an earlier plateau in patients receiving bupropion alone. Further investigation of the safety of this combination is under way, with the possibility that it will be approved soon for the treatment of obesity.

Another recently approved treatment for obesity is the serotonin 5HT2C agonist lorcaserin (Belviq) (Figures 14-20F and 14-21). 5HT2C receptors have long been linked to appetite, food intake, and weight, and 5HT2Cantagonists are associated with weight gain, especially if they are given simultaneously with H1 antihistamines (which is the case for many atypical antipsychotics and some antidepressants) (Figure 14-21; see also the discussion in Chapter 5 on antipsychotics and Figures 5-295-395-41 through 5-43, and the discussion in Chapter 7 on antidepressants and Figures 7-457-667-67). Consistent with the formulation that blockade of 5HT2C receptors is associated with weight gain is the observation that experimental animals whose 5HT2C receptors are “knocked out” are also obese. It would follow that the opposite action on 5HT2C receptors, namely stimulation of them with an agonist, would be associated with reduced appetite, reduced food intake, and weight loss. SSRIs that increase serotonin at all its receptors, including the 5HT2C receptor, can be associated with weight loss and can be effective in bulimia. Lorcaserin, a selective and well-characterized 5HT2C agonist, may work by activating the POMC appetite-suppressing pathway (Figures 14-20F and 14-21B). Lorcaserin has robust weight-reducing actions in clinical trials with long-term studies up to 2 years in duration. The average weight loss for obese patients taking lorcaserin was 3–4% over those taking placebo. In obese patients who did not have type 2 diabetes, about half of them lost at least 5% of their weight, compared to about a quarter of those treated with placebo.

Figure 14-21. Serotonin 5HT2C and appetite. The combination of histamine H1 antagonism and serotonin 5HT2C antagonism (A) (present in many atypical antipsychotics) may lead to enhanced appetite and consequential weight gain. Conversely, the actions of a 5HT2C agonist, such as lorcaserin (B), lead to appetite suppression and weight loss.

Future treatments for obesity and impulsive–compulsive eating disorders might include another combination product, namely the anticonvulsant zonisamide plus naltrexone (Figure 14-20G), which combines some of the mechanisms already discussed, as zonisamide has some of the same pharmacologic properties as topiramate. Direct-acting MC4R agonists (Figure 14-20) are also in testing for obesity but may be associated with hypothalamic-related side effects, and the efficacy of this single-mechanism approach, like some other single mechanisms for the treatment of obesity, may not be sufficiently robust. Triple reuptake inhibitors of serotonin, norepinephrine, and dopamine such as tasofensine are associated with weight loss and are in clinical testing for obesity. Some medications approved for diabetes have promise for the treatment of obesity, including metformin.

One available treatment for obesity – orlistat – works peripherally to inhibit fat absorption and not on reward circuitry except to the extent that it causes an aversive response to eating fatty foods (diarrhea and flatus). However, orlistat is not highly utilized nor highly palatable to many patients. Bariatric surgery of various types is also effective, particularly for morbid obesity, and is increasingly utilized in obesity in general, but is risky and costly. Several other treatments for obesity have been withdrawn from the market, including various stimulants, such as ephedrine (hypertension and stroke), and the halogenated amphetamine derivative fenfluramine (and dexfenfluramine). Fenfluramine was originally used as a prescription monotherapy, and then later combined with phentermine and widely used for a while in a combination known as phen-fen. However, fenfluramine was pulled from the market after the discovery of cardiac valvular and pulmonary toxicity. Sibutramine (an SNRI at low doses and a triple reuptake inhibitor at high doses) was also withdrawn from the market because of hypertension and cardiac problems. Some stimulants are still available as controlled substances (phentermine, diethylpropion) but are not associated with sustained weight loss in most individuals as monotherapies, and also have side effects from hypertension to abuse potential in monotherapy doses.

Impulsive–compulsive disorders of behavior

The current conceptualization of impulsivity and compulsivity as dimensions of psychopathology that cut across many psychiatric disorders suggests that behaviors themselves may be reinforcing and addicting. Rewarding behaviors and addictions to certain behaviors hypothetically share the same underlying circuitry as drug addiction (Figures 14-1 through 14-7). Impulsivity/compulsivity cannot explain all the aspects of these various and sundry conditions, and discussing how this construct could apply to each of them risks oversimplying some very complex and very different disorders (Table 14-1Tables 14-4 through 14-8). Furthermore, the discussion in this chapter does not address the many other unique aspects of these disorders, their current diagnostic criteria, the ongoing debates on their evolving diagnostic criteria, or even whether some conditions are disorders at all. Instead we look here at psychiatric conditions which exhibit behaviors that are either impulsive (meaning that they are difficult to prevent because short-term reward is chosen over long-term gain) or compulsive (meaning that an originally rewarding behavior becomes a habit which is difficult to stop because it reduces tension and withdrawal effects).

Many impulses can become an impulsive–compulsive disorder when done to excess, and several are listed in Table 14-5. Some experts believe gambling disorder should be classified along with drug addiction as the only nonsubstance disorder in that category. Gambling disorder is characterized by repeated unsuccessful efforts to stop despite adverse consequences, tolerance (gambling higher and higher dollar amounts), psychological withdrawal when not gambling, and relief when reinitiating gambling. Internet addictionpyromania, and kleptomania are not considered actual disorders by many, but they can involve an inability to stop the designated behavior (i.e., time on the internet, setting fires, stealing upon impulse), show the development of tolerance and withdrawal, and demonstrate relief when reinitiating the behavior. Paraphilias, considered to be psychiatric disorders, and hypersexual disorder, under consideration as a psychiatric disorder, all have these same characteristics of impulsivity transitioning to compulsivity for a variety of sexual behaviors.

Table 14-5 When does an impulse become an impulsive–compulsive disorder?

Gambling disorder

Internet addiction




Hypersexual disorder

Many disorders considered to be neurodevelopmental have impulsivity/compulsivity as a symptom dimension. This includes notably ADHD, discussed extensively in Chapter 12 (see circuits associated with impulsivity in Figures 12-212-512-712-8 and 14-114-214-4). ADHD is an impulsive–compulsive disorder in which treatments may be effective for the impulsivity. Whereas increasing dopamine in the nucleus accumbens of the ventral striatum with high-dose rapidly administered stimulants can enhance impulsive action, increasing dopamine in orbitofrontal cortex (OFC) circuits with low-dose and slow-release stimulants can decrease impulsivity and enhance the individual’s ability to say no to an impulsive temptation. Impulsivity can also occur in mania and is particularly difficult to treat when it accompanies ADHD, especially in children. Autism and related spectrum disorders may be associated with impulsivity but also compulsive, stereotyped behaviors. Tourette’s syndrome and related tic disorders and stereotyped movement disorders may be forms mostly of compulsivity (Table 14-6).

Table 14-6 Are there neurodevelopmental impulsive–compulsive disorders?

Attention deficit hyperactivity disorder (ADHD)

Autism spectrum disorders

Tourette’s syndrome and tic disorders

Stereotyped movement disorders

Aggression and violence have long been controversial issues in psychiatry (Table 14-7). When violence is premeditated, callous, and calculated, it may be criminal, psychopathic, and predatory – and this type of violence would be neither impulsive nor compulsive. However, aggression and violence, both to others and to oneself, are associated with many psychiatric disorders (Table 14-7), and especially when aggression and violence are impulsive, and are readily and inappropriately provoked, these behaviors are increasingly being considered an impulsive dimension of psychopathology. Impulsive violence can occur in psychotic disorders of many types, including drug-induced psychosisschizophrenia and bipolar mania, as well as in borderline personality disorder. Treatment of the underlying condition, often with antipsychotics, can be helpful. Aggression and violence in such disorders can be considered an imbalance between top-down “stop” signals and bottom-up drives and “go” signals, just as in other impulsive–compulsive disorders (Figures 14-114-214-3). Sometimes aggression becomes increasingly compulsive, rather than manipulative and planned, as in some cases of repetitive self-harm in borderline personality disorder, especially in institutional settings. There is renewed interest in the condition known as intermittent explosive disorder as an impulsive–compulsive disorder of aggression. It may be described as repeated reactions to frustration with irritability, temper tantrums, and destructive behavior that is not premeditated and not committed to achieve a tangible objective (e.g., money, power, intimidation), in the absence of another psychiatric disorder that could explain the impulsive aggression. Because individuals with antisocial personality disorderdyssocial personality disorderpsychopathic traits, and conduct disorder can all have a mixture of manipulative and planned aggression as well as impulsive aggression, the cause of a given episode of violence can be very difficult to determine. Oppositional defiant disorder in children is often associated with impulsive acts, including impulsive and oppositional verbalizations.

Table 14-7 Can violence be an impulsive–compulsive disorder?

Intermittent explosive disorder

Impulsive violence in psychosis, mania, and borderline personality disorder

Self-harm and parasuicidal behaviors/violence against self

Oppositional defiant disorder

Conduct disorder

Dyssocial personality disorder

Antisocial personality disorder


Obsessive–compulsive disorder

Obsessive–compulsive disorder (OCD) in many ways is the prototypical impulsive–compulsive disorder, although it has often been considered to be an anxiety disorder (Table 14-8). In OCD, many patients experience an intense urge to perform stereotypic, ritualistic acts despite having full insight into how senseless and excessive these behaviors are, and having no real desire for the outcome of these actions. The most common types of compulsions are checking and cleaning. For OCD, a general propensity towards habit may be expressed solely as avoidance, deriving from the comorbid anxiety that they report. In the context of high anxiety, superstitious avoidance responses may offer relief, which reinforces the behavior. Stress and anxiety may enhance the formation of habits, whether positively or negatively motivated. However, as the habit becomes progressively compulsive, the experience of relief may no longer be the driving force, and instead the behavior comes under external control as a conditioned response.

Excessive inflexible behaviors are often thought to be carried out in order to neutralize anxiety or distress evoked by particular obsessions. Paradoxically, although OCD patients feel compelled to perform these behaviors, they are often aware that they are more disruptive than helpful. So why do they do them? Rather than conceptualizing compulsive behaviors as goal-directed to reduce anxiety, these rituals might be better understood as habits provoked mindlessly from a stimulus in the environment.

Such hypothesized habit learning can be reduced or reversed with exposure and response prevention, involving graded exposure to anxiety-provoking stimuli/situations, and prevention of the associated avoidance compulsions. This type of cognitive behavioral therapy is thought to have its therapeutic effect by breaking the pattern of compulsive avoidance that confers dominant control to the external environment (such that, for example, the sight of a door elicits checking) and also maintains inappropriate anxiety. Instead of considering compulsions as behavioral reactions to abnormal obsessions, the reverse may be true: obsessions in OCD may in fact be post hoc rationalizations of otherwise inexplicable compulsive urges. OCD patients have demonstrated lack of efficient information processing in their OFC and lack of cognitive flexibility, and thus cannot inhibit their compulsive responses/habits.

Table 14-8 OCD or ICD? Are obsessive–compulsive spectrum disorders also impulsive–compulsive disorders?

Obsessive–compulsive disorder (OCD)

Body dysmorphic disorder (BDD)


Trichotillomania (TTM)

Skin picking

Compulsive shopping



First-line treatment of OCD is specifically with one of the SSRIs. Although second-line treatments with one of the tricyclic antidepressants with serotonergic properties, clomipramine, with SNRIs or with MAO inhibitors are all worthy of consideration, the best option for a patient who has failed several SSRIs is often to consider very high doses with an SSRI or augmentation of an SSRI with an atypical antipsychotic. The mechanisms of action of all of these agents are covered in detail in Chapter 7. Augmentation of an SSRI with a benzodiazepine, lithium, or buspirone can also be considered. An experimental treatment for OCD is deep brain stimulation, discussed in Chapter 7 for depression and illustrated in Figure 7-76.

Many conditions related to OCD are listed in Table 14-8. These include hoarding and compulsive shopping (not necessarily considered a disorder). Compulsive hair pulling (trichotillomania), and compulsive skin picking are conditions that are often much more compulsive than impulsive. Body dysmorphic disorder is preoccupation with perceived defects or flaws in appearance that cause repetitive behavior such as looking in the mirror, grooming, and reassurance seeking. Even preoccupations with health, body function, and pain such as exist in hypochondriasis and somatization can be considered types of obsessions.


We have discussed the current conceptualization of impulsivity and compulsivity as dimensions of psychopathology that cut across many psychiatric disorders. Rewarding behaviors and addiction to drugs or behaviors hypothetically share the same underlying circuitry with impulsivity – defined as behaviors that are difficult to prevent because short-term reward is chosen over long-term gain – mapped onto a prefrontal ventral striatal reward circuit, and compulsivity – defined as an originally rewarding behavior becoming a habit which is difficult to stop because it reduces tension and withdrawal effects – mapped onto a prefrontal dorsal motor response inhibition circuit. Hypothetically, failure of the balance between top-down inhibition and bottom-up drives is the common underlying neurobiological mechanism of impulsivity and compulsivity.

Both drugs and behaviors can be associated with impulsivity/compulsivity and are dimensions of psychopathology for a wide range of drug addictions and psychiatric disorders. The chapter discusses the psychopharmacology of reward and the brain circuitry that regulates reward. We have attempted to explain the psychopharmacological mechanisms of actions of various drugs of abuse, from nicotine to alcohol, and also opioids, stimulants, sedative hypnotics, marijuana, hallucinogens, and club drugs. In the case of nicotine and alcohol, various novel psychopharmacological treatments are discussed, including the α4β2 selective nicotine partial agonist (NPA) varenicline for smoking cessation, naltrexone for opioid and alcohol addiction, and acamprosate for alcohol addiction. Obesity and its relationship to food addiction and impulsive–compulsive disorders is discussed along with numerous new treatments, including lorcaserin, phentermine/topiramate ER, and bupropion/naltrexone. Finally, a number of behavioral disorders are discussed as potential impulsive–compulsive disorders or even as behavioral addictions, including gambling, ADHD, impulsive violence, borderline personality disorder, obsessive–compulsive disorder, and many more.