The Thirteenth Step: Addiction in the Age of Brain Science

Let thy desire flourish,

In order to let thy heart forget the beatifications for thee.

Follow thy desire, as long as thou shalt live.

Put myrrh upon thy head and clothing of fine linen upon thee,

Being anointed with genuine marvels of the god’s property.

Set an increase to thy good things;

Let not thy heart flag.

Follow thy desire and thy good.

Fulfill thy needs upon earth, after the command of thy heart,

Until there come for thee that day of mourning.¹

If, like Jefferson, you hold the truth to be self-evident that all people are endowed with the right to pursue happiness, you have probably just revealed yourself to be a somewhat secularized Westerner. A Buddhist certainly would disagree, instead viewing the pursuit of happiness or pleasure as illusory while considering absence of pain and suffering, for oneself as well as for others, to be a more worthy goal. But as we know, even within the West and Christianity there have been, and remain, deep divisions over the virtue of pursuing happiness in this life. Throughout the history of humankind, our relationship to happiness, pleasure, or, to use the Greek word, hedonism has been conflicted. That, as it happens, is for good reason. We need to understand the attractions and dangers of seeking pleasure in order to understand people’s relationship to addictive drugs. Before we dive into that material, here’s a riddle to contemplate: find the word in Jefferson’s account that captures the most important aspect of these hedonic processes as they relate to addictive disorders. Hint: it is not the word “happiness.” Keep thinking while we move on.

The fact that consequences of a behavior can reinforce that behavior is classically modeled by an experimental animal that is able to obtain a food pellet when it presses a lever under laboratory conditions.² This phenomenon has been studied since the early days of behaviorism by scientists such as Edward Lee Thorndyke, John B. Watson, and B. F. Skinner. To understand the sea change that was to come, it is important to recognize a programmatic principle of the behaviorist movement, as articulated by Watson and his “methodological behaviorism.” This school of thought went out of its way to negate the relevance of any internal states for behaviors that could be reinforced by their outcomes. So we have to keep in mind that reinforcement, in those days, was a purely descriptive, statistical term. Any consequence of a behavior that made that behavior happen again, with a higher frequency, was considered reinforcing. If approaching the pantry results in obtaining a cookie, the frequency with which I will venture to open the pantry can be expected to increase. But that is not supposed to say anything important or meaningful about how I feel when seeking out and consuming the cookies. Early in my medical training, this perspective on human behavior seemed to me so dry, sterile, and detached from the passions and pains of real patients I encountered that I shrugged it off and decided it had nothing to offer me. It took me almost two decades to reconsider that position, and then only partly.

It is generally held that the shift of focus from easily observable behaviors to internal mental processes can be attributed to the influential cognitive neuroscience movement.³ In its broadest sense, cognition refers to all mental processes, but somehow cognitive neuroscience historically had little interest in motivation and emotion. Instead it focused on processes such as memory, attention, perception, action, and problem solving. Important though these are, it always seemed odd to me that so much more emphasis was placed on how people and animals executed mental processes than why they did so. From a clinical perspective, the latter seemed a lot more important. Today those divides have to some extent been bridged. Motivational and affective processes are increasingly recognized as the fundamental engines fueling other functions of the brain.

With that, it is clear that a major contribution to cracking open the black box of behaviorism was a chance finding made in 1954. That year Jim Olds was a postdoctoral fellow at McGill University in Montreal and seems to have had little interest in reinforcement or, for that matter, addiction. Instead he was using electrical stimulation of different sites in the rat brain to examine whether stimulating their activity could facilitate learning. In the course of that work, Olds made a classic discovery. In doing so, he showed the hallmark of the truly great scientist. Rather than writing off unexpected observations the way most people do, he paid attention to them and tried to understand what was going on. With the data themselves published in the best scientific journals,⁴ the true excitement of the moment was perhaps best captured in an account by Olds many years later:

This suggested a logic that was then, and remains today, as attractively as it is deceptively simple. Some brain circuits, when activated by electrical stimulation, seemed to generate an internal state that was somehow pleasurable or otherwise desirable and therefore was pursued by the animal by returning to the place where that internal state had first been experienced.⁶ Olds quickly realized the implications and took the observation one critical step further. He built a device in which the electric stimulus generator was connected to a pedal the rat could press, without any need for the experimenter to intervene. When the animal pressed the pedal, the generator was activated and delivered current to the intracranial electrode. Rats invariably went for the pedal and kept stimulating. It is difficult to account for this observation without accepting the notion that an internal state can be reinforcing. We may not be able to ask a rat whether it experiences pleasure or happiness while sending electric current into particular brain areas. But whatever the experience, the animal seems intent on vigorously pursuing it. Presumably the activation of some brain circuits generates internal states that can be said to reward and therefore reinforce behavior. With those observations, intracranial self-stimulation (ICSS) was born as a model to study endogenous brain reward circuits.

Many stories have been told about this famous experiment, but the fact is that the brain of the original rat was lost, so we are left to guess where exactly the electrode tip delivered the current in that first case. Anecdote has it that a student had done the surgery and had gravely misplaced the electrode, adding to the mystique of this serendipitous discovery. It is a good story, but extensive subsequent mapping has rendered it less important. Over time it became clear that stimulating a surprisingly large number of sites within the brain seemed rewarding, as determined by the ability of stimulation to maintain self-stimulation. Stimulation at most other sites did not influence self-stimulation behavior at all. Interestingly, when the electrode targeted yet a third set of sites, animals worked to avoid stimulation, hinting at the presence of circuitry that encoded aversive inner states.

Together these observations made Olds articulate the basic concept that the mammalian brain must be equipped with hard-wired reward systems, the activity of which motivates or reinforces behavior that results in approaching and obtaining “natural rewards,” such as food, drink, or a mate. It was also clear that these systems could potentially be hijacked by processes that short-circuit the connection between engaging in the behavior and obtaining the natural reward. One way, shown by the original experiments, was to directly activate the reward circuitry using electric current. It was not far-fetched to dream up the idea that addictive drugs might accomplish something similar. In fact, although little noted at the time, there was another remarkable similarity between drug seeking and ICSS. Olds showed that a rat would be willing to traverse an electrified grid floor to access the pedal that delivered electric stimulation. More than half a century later, drug seeking despite adverse consequences remains a diagnostic criterion of addiction. In recent years this exact aspect has received renewed interest as an expression of compulsive drug seeking in trying to model addiction in experimental animals.⁷

It took almost three decades from these original discoveries until the critical neurocircuitry that supports ICSS and reward processing was outlined and it became widely accepted that Olds’s original “hijacking” postulate might be largely correct.⁸ One reason is perhaps that it first seemed as if too much of the brain was involved in reward. Another reason for the slow uptake was that at the time Olds published his discovery, the chemical anatomy of the brain was still in its infancy. It would be several years before the Swedish physician and pharmacologist Arvid Carlsson published his claim that dopamine was a neurotransmitter.⁹ Few people believed Carlsson; it took until 2000 before he was rewarded for his discovery with the Nobel Prize. A third factor was that it took decades of perfecting the necessary neuroscience tools before the full picture emerged. One such tool was the development of a particular way to make dopamine neurons shine (“fluoresce”) when exposed to ultraviolet light under the microscope.¹⁰ Another was the development of early antipsychotic medications, which block one kind of dopamine receptors. Once those were discovered, it became possible to microinject them into tiny brain areas. That allowed scientists to map out places in the brain where these injections blocked a behavior. In this way it also became possible to show the role of the brain’s own dopamine transmission in those places for that behavior.

Once all these pieces came together, a fairly consistent picture emerged. A major dopamine pathway of the brain starts out in the midbrain, or mesencephalon, in a tiny structure called the ventral tegmental area. In this structure, about half a million cells hug the midline from both sides,¹¹ about two-thirds of them containing dopamine. From here the dopamine cells send their nerve fibers to several different locations in the forebrain. One of those is the nucleus accumbens, the lower part of the larger basal ganglia system.¹² When this anatomy became well understood and the tools were in place, things came together. Animals would self-stimulate when the electrode tip was somewhere along the mesolimbic dopamine pathway from the ventral tegmental area to the nucleus accumbens. They would also do it with the electrode tip in places from which nerve fibers ran that in turn activated the mesolimbic dopamine pathway. Meanwhile, when tiny amounts of antipsychotics were injected into the nucleus accumbens to block dopamine receptors, animals stopped self-stimulating. And stimulants such as amphetamine and cocaine, which potently boost dopamine transmission in ways I will discuss in a later chapter, were also potently reinforcing when offered to a rat through an intravenous (IV) line. In fact, so strong was the reinforcing value that animals given unlimited access to the drug would lever press for IV cocaine at the expense of food, water, and sleep, to the point that 90 percent of them would die over the course of just one month.¹³

Self-administration as a measure of reward has obvious appeal, but it has also limitations. Most important, animals carry out this behavior with drug in their system. To make matters worse, the amounts of drug they are exposed to vary depending on the rates of self-administration. If the drug itself has an ability to influence behavior, it is easy to see that by looking at self-administration rates, we will be getting a mix of different effects. For instance, cocaine stimulates movement in general. Animals injected with cocaine run around more than before, and if they are placed in a box with levers, they will press those levers often, even if that action has no consequences. In contrast, heroin administration has a calming influence at higher doses. Animals curl up in a corner and become inactive. So comparing self-administration rates involves a mix of at least two different things: motivation, which is what we are interested in, and nonspecific motor effects, which we are not. A powerful approach to measuring drug reward in a drug-free state, called conditioned place preference (CPP), was therefore developed as a complement to self-administration studies.¹⁴

In a typical CPP experiment, a drug is repeatedly given to mice or other experimental animals in a distinctly recognizable environment, perhaps with striped walls and a grid floor. This is alternated with administering a saline solution in an environment that is clearly different—walls without a pattern, an even floor surface. After a few days of this, the animal is allowed to freely explore both environments, now connected with an open passage. No drug or saline injection is given. All the investigator is doing is watching how much time the animal spends on each side. This is much like asking the animal to reveal its preference for the drug memory. In this model, normal mice will show a high preference for an environment associated with drugs that are addictive, and also with natural rewards such food or sex. This has led scientists to conclude that conditioned place preference is a reasonable measure of reward. Recent work shows that it works in humans, too.¹⁵ The most robust place preference is obtained with stimulants and opioids, but, similar to other addictive drugs, mice given alcohol show a preference for the environment associated with this drug.¹⁶ Now, that rings a few bells. I could pull a patient chart almost at random and the pattern would be clear. Admission, detox, and discharge are followed by the patient gravitating back to an environment associated with drug use. This is one of the factors that ultimately set the scene for relapse, as will be discussed in a later chapter.

In the mid-1980s these various lines of research converged. A “psychomotor stimulant theory of addiction” was proposed, which built on the brain reward system theories of Olds and others but added to those the critical postulate that mesolimbic dopamine, through actions on the circuitry outlined by the self-stimulation studies, plays a critical role in the reinforcing properties of addictive drugs.¹⁷ This theory seemed be proven beyond reasonable doubt when Urban Ungerstedt at the Karolinska Institute developed a technique to measure dopamine levels in the living brain while animals engaged in behavior. Using this technique, called brain microdialysis, Gaetano Di Chiara and his student Assunta Imperato in Cagliari, Sardinia, showed that most drugs abused by people potently increased levels of dopamine in the nucleus accumbens of rats.¹⁸ A decade later Nora Volkow, currently director of the National Institute on Drug Abuse, carried out a series of human studies using positron emission tomography (PET). She could show that the subjective high people experienced after using cocaine or other stimulants was also associated with the boost these drugs give to dopamine neurotransmission in the nucleus accumbens.¹⁹

In the late 1990s it all seemed mostly settled, then. Brains had reward systems that use dopamine to drive approach behavior needed to obtain natural rewards. Addictive drugs were able to hijack people’s lives because they hijack these endogenous reward systems and activate dopamine transmission directly. Because the activation they produce is so much higher than that triggered by natural rewards, and because they offer an opportunity to short-circuit the activation of reward systems without requiring the hard work needed for the normal rewards of life, drugs tend to win over natural rewards. For example, mating only doubles dopamine levels in the nucleus accumbens. Meanwhile, amphetamine increases them tenfold. No wonder amphetamine is more attractive!

There were variations on this general theme, mostly dealing with trying to sort out what dopamine was really doing. An important aspect seemed to be learning. Electric recordings in monkeys by Wolfram Schultz, then in Fribourg, Switzerland, indicated that dopamine provides a signal of something that became called reward prediction error:

If this sounds too abstract, let’s try again. Imagine you are strolling through the downtown of a city you recently moved to. You are not particularly familiar with the area. When you start feeling hungry, you don’t know where to get good food. So you keep exploring for a while. Finally, in a side street, there is a little place that serves great falafels and is cheap, too. As you consume your meal, your mesolimbic dopamine neurons fire away. This was clearly better than expected. But a few months later, you know the city well. As you complete your shopping, you find yourself at the same food place almost without noticing it. Clearly you have by now learned what you need to do to get there. As you down the tasty falafel, those DA neurons of yours are largely silent. What would be the point of making any motivational noise? Things are just fine as they are. There is no need to drive your behavior in any other direction. In this context, the role of addictive drugs has sometimes been framed as being agents of pathological learning that somehow maintain their ability to activate dopamine neurotransmission over time.²¹ This would clearly distinguish them from natural rewards, which over time become as good as expected, leading to a decline in their ability to activate the dopamine cells of the ventral tegmental area.

This was all fascinating, but as I was reading about these advances, I was toiling in an intense clinical service. And I could not figure out how any of it related to any patient I had ever had in treatment.

As modern times approached, the model outlined above increasingly became addiction science canon. It also became so engrained that the distinction between different drugs often was lost. What had been established for stimulants was automatically held to be true of other drugs, whether data to support that similarity were available or not.²² Yet we have, for instance, long known that alcohol, the most commonly used addictive drug, has nowhere near the ability of cocaine or amphetamine to activate mesolimbic dopamine transmission. Microdialysis studies showed that self-administration of alcohol resulted in dopamine levels in the rat nucleus accumbens that were about double those at baseline, compared with the tenfold or so increase easily obtained in response to amphetamine. Meanwhile, several studies that no one wanted to remember had consistently shown that when dopamine cells were selectively killed using a toxin, rats happily continued to self-administer alcohol.²³ This was not exactly a ringing endorsement of a critical role for mesolimbic dopamine in alcohol addiction.

These findings clearly made it important to find out if alcohol does activate the classical brain reward circuitry in humans, similar to what had been shown for stimulants. The answer is actually a cautious “yes,” but with an important twist, or several of them. In 2003 a study in eight Canadian men used a PET camera to show that drinking alcohol could indeed produce a measurable dopamine release in the nucleus accumbens. A few years later my late colleague Dan Hommer and his coworkers at the NIAAA used functional magnetic resonance imaging (fMRI) and also found that alcohol activated the nucleus accumbens. In their study, which for a serious scientific study got away with the unusually cute title “Why We Like to Drink,” the degree of activation was also strongly correlated with the subjective feeling of intoxication.²⁴ Together these findings would seem to support an important role of classical brain reward circuitry in the pleasurable, reinforcing properties of alcohol, similar to stimulants. But it is critical to point out that both these studies were carried out in healthy, socially drinking participants. And even in this category of people, to get a robust response, the person should preferably be male²⁵ and have a particular genetic makeup.²⁶ When the same fMRI measures have been obtained from people with an increasing degree of alcohol problems, alcohol-induced nucleus accumbens activation has turned out to be progressively lower.²⁷ In severe, treatment-seeking alcoholics, there is simply nothing left of it, at least at levels of intoxication we can safely measure in the lab. That of course parallels patient reports that the high from alcohol declines over the years of being addicted. Yet these are also the patients that are the hardest to treat.

Systematic data like these are not available for other drugs, such as stimulants, but clinical experience suggests that this observation parallels what happens to people addicted to those drugs as well. Initially even a modest dose of amphetamine or cocaine will reliably produce a high in most people who try it. But for most patients with more advanced addiction to stimulants, the pleasurable effects wear off over the years and are largely gone by the time treatment is sought. That is why people keep pushing the doses ever higher. Yet despite the declining high in the late stages of addiction, the disease by then holds the patient in a much stronger grip than during the initial stages of experimental or recreational use.

What if a mechanism existed by which the appetitive or incentive motivation to obtain a reward became progressively stronger, or “sensitized,” while the hedonic impact of consuming the same reward decreased? People in whom this scenario had played out would more and more vigorously pursue their coveted reward, but the outcome would be increasingly disappointing. It is easy to see how this kind of process could lead to progressive escalation of both the efforts to obtain the reward and the amounts consumed. People would do almost anything in the hope of achieving the pleasure that was once there but now eludes them. Replace the falafel with alcohol and the description surely resembles advanced alcoholism.

The “incentive sensitization theory” of addiction put forward by Kent Berridge and Terry Robinson ²⁸ is as sophisticated as it has been influential and is in principle able to reconcile the seemingly paradoxical observations that “rewarding” properties of addictive drugs tend to decline as people develop progressively more severe addictive disorders. In this conceptualization, there is no unitary brain reward system. Instead mesolimbic dopamine circuits promote incentive motivation that leads to drug seeking, and under the right—or rather wrong—conditions sensitizes over time as the individual uses drug. This motivation is famously called “wanting.” In contrast, the pleasure from ultimately taking the drug is mediated by another mechanism, which may well decline over time. That hedonic impact is called “liking.”

Remember the original meaning of the Greek root of “hedonic”? It means “sweet.” Incidentally, sweet-tasting things make babies smile. It turns out that rats are no different, at least not once scientists learn to interpret their facial expressions. In a fascinating line of research, Kent Berridge has studied the neurobiology of these hedonic rat smiles. This research suggests that when hedonic hotspots in the brain are turned on by sweets, the pleasure or “liking” that follows “pursuit” or “wanting” is generated by its own brain machinery. These brain circuits are thought to be located in brain sites that overlap or are intertwined with those that produce pursuit, or “wanting,” but the “liking” circuits seem to use different neurochemicals. The most important among those are endogenous morphine-like substances, or endorphins.²⁹ Endorphins may well be the oldest reward transmitters on the evolutionary block. They produce the pleasure that comes from the things that really matter in life, such as being held by your mother as a child or being stroked by your mate in adulthood.³⁰

The discovery of endogenous opioids and their receptors has all the drama of the best detective stories. Beginning in the mid-1950s, it was speculated that the body may have a receptor for morphine, but at that time there was simply no way of showing it. Then, in the 1960s, chemists at the pharmaceutical company Sankyo tweaked the morphine molecule. In one of those tweaks, they ended up with something called naloxone, a medicine that was able to block all the actions of morphine, and that is still used in the clinic to treat heroin overdose. By tagging naloxone with a radioactive label that could be precisely measured, in 1973 Solomon Snyder at Johns Hopkins University and his graduate student Candace Pert were finally able to show the presence of opioid receptors in neural tissue.³¹ Meanwhile, in Uppsala, Sweden, Lars Terenius and his colleagues found that the brain seemed to contain endogenous substances that bind to the same receptor as morphine.³² And then Hans Kosterlitz, in Aberdeen, Scotland, together with his colleague John Hughes was able to fish out the first of these endogenous opioids, or endorphins.³³ It took only three years from the latter discovery, to 1978, before Snyder, Hughes, and Kosterlitz shared the Lasker Award, after the Nobel Prize the most prestigious award in medicine. It was widely expected that the Nobel Prize would follow, probably to be shared by Snyder, Kosterlitz, and Terenius. It didn’t. In an unprecedented action, Candace Pert sued the Lasker jury for having left her out of the award. That was not the kind of publicity any prestigious award-granting institution wanted. The discovery of the endogenous opioids and their receptors became perhaps the greatest biomedical advance not to be rewarded with a Nobel Prize.³⁴

The receptor that Snyder and Pert had found was labeled with the Greek letter µ (pronounced “mu”), simply as a shorthand for “morphine-receptor.” As years went by, two other opioid receptors, kappa and delta, were also found, together with a lot of exciting science. For a while, based on pharmacological studies, scientists began to think there could be many more types of opioid receptors. But once molecular biology allowed genes to be cloned, it became clear that there are only three opioid receptor genes, from which the mu-, kappa-, and delta-opioid receptors are produced, respectively. After years of attempts, the delta-opioid receptor was the first of these to be cloned, in 1992, simultaneously and independently by Chris Evans of the University of California, Los Angeles, and Brigitte Kieffer in Strasbourg, France. The way cloning was done in those days, once you had one member of a gene family, it was easier to fish out other members. The cloning of the delta-receptor therefore gave the tools that soon allowed several groups to clone the mu-opioid receptor as well. As expected, it turned out to be central for the addictive properties of morphine and morphine-like drugs that I will discuss in a coming chapter. But in short, without the mu-receptor, mice would not self-administer heroin. And with the mu-receptor blocked, opiate addicts saw no value in taking the drug. They simply did not have the liking for it.

We will revisit if and how that knowledge can be used for treatment purposes, in the chapter on addiction medications. For now, let’s just say that the mu-opioid receptors are in all the right places in the nervous system to produce the known effects of opioids. They are present at several levels along the pathway through which painful stimuli are funneled in from sensory detectors in the body and into the brain. This explains how activation of these receptors can dial down the flow of pain information into the brain and also decrease the “painfulness” value assigned to that information once it does reach the brain. The mu-opioid receptors are also present in the brain stem centers that pace breathing. And, most important for us, they are present at two critical stations along the mesolimbic dopamine pathway, which is thought to be critical for reward and approach behavior. Mu-opioid receptors are present at the beginning of this pathway, where the dopamine-producing nerve cells are located in the ventral tegmental area, and where mu-receptor activation turns on the dopamine neurons. And they are present at the site to which the dopamine cells send their nerve endings, the nucleus accumbens, a structure that opioid receptors can also activate directly.³⁵

As we near the end of our discussion of “wanting” and “liking,” what about the riddle at the beginning of this chapter? Returning to the incentive sensitization theory of addiction, we can perhaps now see that the key word in Jefferson’s expression may not be “happiness” but rather “pursuit.” To initiate the hard work of seeking drugs, the wanting machinery must be turned on. It is less critical how much the addict will like the drug once he or she actually gets it. To be honest, I am personally half fascinated, half skeptical whenever I read this theory. The incentive sensitization theory indeed has intellectual elegance to it when offering its explanation of how pursuit may intensify despite declining rewards. But it remains subject to vigorous debate in the addiction research community how important this theoretical framework is for the clinical realities of addiction. To me, the answer will be provided only once we find out whether the theory will help uncover any useful targets for successful addiction treatments. To date that has not been the case. But then again, the list of successful endeavors based on other approaches is quite short as well.

In recent years the neurobiology of transitioning from goal-oriented, motivated behavior to actions that are habitual and automatic has become better understood. Although not all scientists agree, it has been suggested that dopamine transmission also has an important role for moving the drive to execute approach behavior from the nucleus accumbens to the part of the basal ganglia that lies just above it, the dorsal striatum.³⁶ This process has been proposed to play an important role for continued drug seeking and taking despite declining reward value and adverse consequences.³⁷ This too may have parallels in clinical experience. It is certainly true that heavy, “habitual” smokers who have a pack of cigarettes readily available to them will tend to light up without giving it much thought, occasionally even showing surprise when the cigarette reaches their lips. It is also true that occasionally intravenous drug addicts will seek out needles and syringes, fill them with almost anything, and shoot up in a pattern that simply is not sensitive to the fact that they are unlikely to receive any pleasure from this action. But already in 2003 Robinson and Berridge made a powerful argument against habit formation as a major mechanism behind addictive disorders:

Many aspects of addictive drug pursuit are flexible and not habitual. Human addicts face a situation different from rats that merely lever-press for drugs. We suspect that if animals were required to forage freely in a complex environment for drugs the picture seen in animal neuroscience might look more like the situation in human addiction, and automatic habit hypotheses would be less tempting. An addict who steals, another who scams, another who has the money and simply must negotiate a drug purchase—all face new and unique challenges with each new victim or negotiation. Instrumental ingenuity and variation are central to addictive drug pursuit in real life. When an addict’s drugtaking ritual is interrupted, for example, by lack of available drugs, flexible and compulsive pursuit is brought to the fore…. The strongest … habit in the world does not explain the frantic behavior that ensues.³⁸

Of course, a partial contribution from habit formation in some cases does not necessarily contradict a role for incentive sensitization or other mechanisms. Several of these could be at play simultaneously in the same individual during different stages of the addictive process, or in different individuals. I remain skeptical that habit formation would be a major cause behind continued drug use, mostly for the reasons advanced in the quote above. But in this case, as in the previous one, it is best to keep an open mind. We should let the utility of the model for developing successful treatments decide just how important it is.

But other findings suggest that this “reward deficit syndrome” indeed might predispose individuals to escalation of drug use. Michael Nader and his team at Wake Forest University carried out one particularly elegant piece of work in support of this notion by using social manipulations in monkeys.³⁹ In these experiments, the scientists used a PET camera to measure the numbers of dopamine receptors. Socially isolated monkeys are miserable and have low receptor levels, but being in a group is not necessarily a picnic either. Some monkeys end up at the top of the social hierarchy, others at the bottom. It turns out that the stress of being a subordinate monkey, at the low end of the hierarchy, led to numbers of dopamine receptors that were as low as those in socially isolated monkeys. When the monkeys were then given access to intravenous cocaine, those with the low receptor numbers were the ones that self-administered the largest amounts of cocaine. This nicely ties in with a human study in which people with low numbers of dopamine receptors found the stimulant methylphenidate to be quite pleasurable, while those with high receptor numbers did not.⁴⁰ Maybe the reward deficit syndrome actually makes sense. Maybe an alternative to taking drugs for those who are at the bottom of society’s ladder is to change the social order.