CHAPTER SIX

VIRTUOUS PLEASURES (AND A LITTLE PAIN)

Jeff Tweedy, leader of the roots-inflected rock bands Wilco and Uncle Tupelo, struggled mightily with various drug addictions, most notably to prescription painkillers, alcohol, and cigarettes. These were coincident with, and in some cases triggered by, chronic migraines, major depression, and panic attacks that have plagued him for years. After a successful rehab and several years of drug-free living, he had this to say about his life:

I’ve never felt better. I’ve never been healthier…. I run four or five miles, four or five times a week, but I broke both my legs running too much last summer. I had stress fractures in both my tibias from running too much. You know, once you’re an addict, you’re always an addict, so just because I found something good to do doesn’t mean I’m not going to hurt myself doing it.¹

Yes, as we will discover, exercise can activate the pleasure circuit. And so, like nicotine or orgasm or food or gambling, it can become a substrate for addiction as well. This can indeed be a genuine addiction, not merely one as expressed in a common usage like “I’m addicted to sleeping on 600-thread-count sheets.” Real exercise addicts display all of the hallmarks of substance addicts: tolerance, craving, withdrawal, and the need to exercise “just to feel normal.”² Does this make exercise a virtue, a vice, or a little of both?

 

Sustained physical exercise, whether it be running or swimming or cycling or some other aerobic activity, has well-known health benefits, including improvements in the function of the cardiovascular, pulmonary, and endocrine systems. Voluntary exercise is also associated with long-term improvements in mental function and is the single best thing one can do to slow the cognitive decline that accompanies normal aging. Exercise has a dramatic antidepressive effect. It blunts the brain’s response to physical and emotional stress. A regular exercise program produces a large number of changes in the brain, including the new growth and branching of small blood vessels, and increases in the geometric complexity of some neuronal dendrites. It is associated with a host of interrelated biochemical changes as well, including increases in the level of a key protein called BDNF (brain-derived neurotrophic factor). At present we have little understanding of which of these morphological or biochemical changes underlie the beneficial effects of voluntary exercise on brain function, but this is an area of active research.³

In addition to the beneficial long-term effects of a sustained exercise program, there are also short-term benefits of exercise that wear off after an hour or two. These include an increased pain threshold, reduction of acute anxiety, and “runner’s high.” ⁴ Runner’s high (which can occur following any intense aerobic exercise, not just running) is a short-lasting, deeply euphoric state that’s well beyond the simple relaxation or peacefulness felt by many following intense exercise. Careful surveys have revealed it to be rather rare: The majority of athletes, whether amateur or professional, never experience it at all, and those who do do so only intermittently. Indeed, many distance runners or swimmers feel merely drained or even nauseated at the end of a long race, not blissful. Since the 1970s runner’s high has been attributed in the popular imagination to the exercise-triggered production of endorphins, the brain’s own morphine-like molecules. This idea was initiated by a series of studies in which blood was drawn from subjects before and after intense exercise. Analysis revealed an exercise-associated increase in the level of a particular endorphin, called beta-endorphin, in the blood.

There’s a major problem, however, in trying to link runner’s high with circulating beta-endorphin. Beta-endorphin almost completely fails to cross the cellular barrier that separates the bloodstream from the brain. If beta-endorphin in the bloodstream were indeed responsible for runner’s high, then it would have to increase levels of some other chemical messenger that would then cross into the brain to exert its effects. Alternatively, there are different types of endorphins (and related molecules called enkephalins, which together with endorphins are called endogenous opioids) that are synthesized within the brain and therefore could cause euphoria without crossing the blood-brain barrier.

One way to address this question would be to perform a spinal tap on people before and after exercise to see if opioid levels rose in the cerebrospinal fluid that bathes the brain and spinal cord. However, because a spinal tap is painful and carries a small risk of complications, human subjects review boards at most institutions have ruled that it is not ethical to conduct that type of experiment. Dr. Henning Boecker and his coworkers at the University of Bonn in Germany realized that they could investigate runner’s high without resorting to spinal taps by measuring brain opioid levels with a brain scanner.⁵ They recruited ten amateur distance runners who had previously reported experiencing runner’s high. Each subject received a baseline brain scan using a radioactively tagged drug designed to measure the secretion of all forms of endogenous opioid (it bound to all types of the brain’s many opioid receptors) and completed a survey of mood. After the subjects had a two-hour-long run followed by a thirty-minute cool-down period, the brain scan and mood survey were repeated. The researchers found that this long run was associated with increased opioid release in the runners’ brains, particularly in the prefrontal cortex (a planning and evaluation center) and the anterior cingulate cortex and insula (which serve to interface pain and pleasure with emotions). In addition, those subjects who reported the highest levels of euphoria after running also had the highest levels of opioid release.

This study is an interesting first step, but much more remains to be done in the area. One useful line of work will be to repeat the experiment using more specific opioid receptor drug probes in an attempt to implicate a particular endogenous opioid in runner’s high. Then drugs that block those receptors could be given to see if runner’s high could be attenuated. It’s likely that runner’s high is not entirely mediated by the opioid system: Exercise also increases the levels of endocannabinoids, the brain’s natural cannabis-like molecules, in the bloodstream. Unlike beta-endorphin, which cannot readily pass the blood-brain barrier, endocannabinoids easily move throughout the body. Thus exercise-induced increases in endocannabinoid levels in blood are presumably mirrored in the brain and could also contribute to the euphoria of runner’s high.⁶

Putting some of the pieces together, we know that intense exercise can bring about short-term euphoria, reduction of anxiety, and increases in pain threshold. This is coincident with increases in the levels of brain opioids and, presumably, endocannabinoids, both of which can produce all of these short-term psychoactive effects. We also know that endocannabinoids and opioids can indirectly activate dopamine cells of the VTA and thereby stimulate the medial forebrain pleasure circuit. We know that exercise can be addictive and that other substances and behaviors that are addictive have increased dopamine release in VTA target regions as a common property. In rats, sustained wheel-running can cause dopamine release in the nucleus accumbens and other VTA target regions. Rats also show some signs of exercise addiction. For example, they can be trained to work hard (i.e., perform many lever presses) for access to a running wheel.⁷

All these observations taken together suggest that intense exercise will activate release from VTA dopamine neurons, a process that will underlie at least some portion of runner’s high. Unfortunately, to date there is little evidence to support this theory in humans. Gene-Jack Wang and his colleagues at Brookhaven National Laboratory used a brain scanner to image dopamine release in the nucleus accumbens and dorsal striatum of twelve subjects before and after thirty minutes of vigorous treadmill running, followed by a ten-minute cool-down period.⁸ They found no differences in D2 dopamine receptor occupancy (their measure of dopamine release) associated with this exercise regimen. No mood scale ratings were performed, so we cannot know if these subjects experienced runner’s high. It would be useful to repeat this experiment together with mood scale ratings and more intense exercise, as Boecker and coworkers did for their endogenous opioid measurements.

 

In the late eighteenth century the British philosopher Jeremy Bentham famously proclaimed, “Nature has placed mankind under the governance of two sovereign masters, pain and pleasure… . They govern us in all we do, in all we say, in all we think: every effort we can make to throw off our subjection, will serve but to demonstrate and confirm it.”⁹ The accumulating neurobiological evidence indicates that Bentham was only half correct. Pleasure is indeed one compass of our mental function, guiding us toward both virtues and vices, and pain is another. However, we now have reason to believe that they are not two ends of a continuum. The opposite of pleasure isn’t pain; rather, just as the opposite of love is not hate but indifference, the opposite of pleasure is not pain but ennui—a lack of interest in sensation and experience. You don’t have to be a sadomasochistic sex enthusiast to know that pleasure and pain can be felt simultaneously: Recall Boecker’s aching but blissful long-distance runners, or women in childbirth. In the lexicon of cognitive neuroscience, both pleasure and pain indicate salience, that is, experience that is potentially important and thereby deserving of attention. Emotion is the currency of salience, and both positive emotions like euphoria and love and negative emotions like fear, anger, and disgust signal events that we must not ignore.

You’ll recall our discussion in chapter 4 about how epileptic seizures or brain stimulation with electrodes can produce orgasms that are devoid of pleasure or emotional feeling. While we normally experience orgasm (and everything else, really) as a unified perception, these results revealed that orgasm has in fact dissociable sensory/discriminative and pleasurable/emotional components that are mediated by separate brain regions. The same general theme holds true for pain. There is a sensory/discriminative pathway that runs through the lateral portion of the thalamus, far from the midline, and continues to a region of the cortex involved in touch and muscular sensation (called the primary somatosensory cortex). A parallel pathway involved in the emotional sense of pain runs through the medial thalamus and then contacts two emotion centers, the insula and the anterior cingulate cortex. People who sustain damage limited to the lateral pathway will report an unpleasant emotional reaction to a painful stimulus, but will be unable to describe its specific qualities (dull, stabbing, cold, hot, etc.) or to locate the painful region on their body. Selective damage to the lateral, emotional pathway results in the opposite condition, called pain asymbolia, in which people can report the quality and location of a painful stimulus, but it no longer carries any emotional weight. They have normal reflex-withdrawal when subjected to pain (and a normal reflexive facial grimace), yet the pain just doesn’t seem to bother them very much.

In conversation we often use terms like “emotional pain” or “painful social situations.” Is this just metaphoric language, or do we actually experience certain powerful emotions the way we do physical pain? The accumulating evidence of recent years indicates that emotional pain activates the medial but not the lateral portion of the physical pain pathway. Experiments that have been devised to inflict even mild social pain (like exclusion from a group task or betrayal by a partner in a gambling game) have demonstrated significant activation of the insula and the anterior cingulate cortex. Emotional pain isn’t just a metaphor: In terms of brain activation, it partially overlaps with physical pain.¹⁰

 

Both animal and human studies have recently revealed a rather peculiar finding regarding pleasure and pain: Dopamine release from neurons of the VTA, the central biochemical event of the pleasure circuit, is also engaged by painful stimuli. Jon-Kar Zubieta and his colleagues from the University of Michigan performed brain scanning to measure dopamine release in subjects who received a painful stimulus produced by continuously injecting a concentrated salt solution into the jaw muscle.¹¹ This treatment produced a protracted aching-type pain that lasted for about an hour. The control condition consisted of injecting a normal (isotonic) salt solution that is reported not to be painful. All subjects filled out surveys designed to assess both the emotional and the sensory aspects of pain. The main result was that this long-term painful stimulus was associated with increased dopamine release in both the dorsal striatum and the nucleus accumbens. In the nucleus accumbens, the greatest dopamine release was seen in those subjects who reported the highest emotional pain ratings.

How are we to interpret these findings? Isn’t the VTA–nucleus accumbens dopamine pathway supposed to be the core of the pleasure/reward circuit? In trying to make sense of this finding, it is worthwhile to remember that today’s brain scanning techniques are very crude. Even a single voxel (a three-dimensional pixel) in a brain scanning image is averaging the response of many thousands to millions of neurons spread over the course of many seconds, so that both the spatial and the temporal resolution are very low. What happens when this sort of experiment is repeated in a rat, where electrodes can be inserted into the brain and recordings made from single dopamine neurons in the VTA? When Mark Ungless and his colleagues at Imperial College London recorded activity from single VTA dopamine neurons in response to a brief (four-second-long) painful foot shock, they found a very interesting result.¹² While all of the dopamine neurons in the VTA were activated by rewards (like a syrup droplet), there were two different patterns of response to the painful stimulus. Dopamine neurons in the dorsal portion of the VTA were transiently inhibited by foot shock: Their firing rate decreased below background levels.¹³ However, in the ventral portion of the VTA, dopamine neurons were transiently activated by the foot shock. Thus there appear to be two parallel circuits in the VTA. The first is the classic pleasure circuit that we have spent so much time discussing, which is activated by pleasure/rewards and inhibited by pain. The second is a “salience circuit” that is activated by both pleasurable and painful stimuli and is tightly coupled to emotional responses.

Does that mean that pain itself is somehow rewarding, or is it merely salient? The answer isn’t entirely clear and awaits further research. There are several factors that complicate any interpretations we might now make. For example, all brief painful stimuli eventually end, and the relief from pain we experience as they do end is itself pleasurable. Chronic, long-lasting pain is a somewhat different story, as it is likely to produce long-term changes in the brain’s pleasure circuitry through the action of stress hormones. It is tempting to speculate that the addition of pain to pleasure creates a super-salient response in the medial forebrain and that this somehow contributes to the popularity, in some quarters, of sadomasochistic sex, or even of tasty food loaded with chili peppers.

 

Another virtuous pleasure that is culturally widespread and often linked to spiritual practice is meditation. But what, exactly, is meditation? In her book The Blissful Brain, Shanida Nataraja offers these criteria: (1) it must involve a specific technique that is both clearly defined and taught (spacing out in the shower doesn’t count), (2) it must involve progressive muscle relaxation, (3) it must involve a reduction in logical processing, and (4) it must be self-induced (thereby excluding the use of drugs or hypnosis).¹⁴ The range of techniques that fall within these boundaries is actually quite large. While all meditative practices involve the conscious regulation of attention and emotion, beyond these basic criteria there is substantial variation. As Richard Davidson, who is both a neurobiologist and an experienced practitioner of meditation, has observed, “Meditation refers to a whole family of practices—it’s like using the word ‘sports.’”¹⁵

For example, Yoga Nidra meditation—also known as yogic sleep—is a practice in which the meditator becomes a relaxed, dreamy, neutral observer: She experiences a loss of conscious control of her actions, and her mind withdraws from wishing to act. We can contrast this method with Zen Buddhist meditation, which has the goal of “thinking about not thinking” but prescribes a vigilant mental attitude. This is promoted by a specific seated posture with the eyes open. In Zen meditative practice, mental withdrawal from the sensory world and its attendant dreamy quality are actively discouraged.¹⁶ Yet another meditative practice is Loving Kindness–Compassion meditation from the Buddhist tradition, in which the goals are to counteract self-centered tendencies and ultimately, after long training, to feel a generalized, nonreferential compassion for all beings. The practice does not involve focused attention on particular objects, memories, or images.

In recent years, brain scanners have been employed to identify the regional patterns of activation and deactivation in these meditative states. Not surprisingly, given the wide variety of meditative practices, the results have varied.¹⁷ These studies generally compare the same subject in the meditative state versus some control condition. Herbert Benson and his colleagues at Harvard Medical School used a control condition in which subjects were asked to silently generate a random list of animals as a baseline for measuring changes accompanying a form of Kundalini meditation. In this form of meditation the subjects, who were experienced meditators, monitored their breathing and silently recited “sat nam” while inhaling and “wahe guru” while exhaling. When compared to the control task, meditation was associated with increased activity in a large number of brain regions, including the dorsolateral prefrontal cortex (a judgment and planning center), anterior cingulate cortex (an emotion center), hippocampus, and striatum, possibly including the nucleus accumbens (the resolution of the images in this particular study makes this latter determination difficult). Some changes were also noted in the brain activation pattern from the earlier to the later stages of the meditation session. By contrast, some but not all studies of Zen meditation have reported decreases in the activation of the dorsolateral frontal cortex and the anterior cingulate cortex. Furthermore, within the same meditative tradition, sustained practice of meditation over many years can also influence the pattern of brain activation. Richard Davidson and associates at the University of Wisconsin, Madison, found an inverted U-shaped curve in which expert meditators (average of nineteen thousand hours) had more brain activation than novices, but super-expert meditators (average of forty-four thousand hours) had less. These measurements involved a group of cortical regions that seem to be engaged by attentional processes.¹⁸

While meditation is certainly relaxing and is sometimes described as blissful, does it in fact activate the medial forebrain pleasure circuit? To date, I am aware of only one study that has sought to address the question directly. Hans Lou and his colleagues at the John F. Kennedy Institute in Denmark performed brain scans on experienced practitioners of Yoga Nidra meditation, using listening to speech with eyes closed as a control condition.¹⁹ They found a significant increase in dopamine release in the nucleus accumbens of their meditators. This result is suggestive, but it awaits both confirmation and extension to other forms of meditation.

 

Meditation and prayer lie on a continuum of spiritual practice, but they can share certain relaxed and dissociative qualities. What about the brain activation correlates of more transient, intense phenomena such as ecstatic spiritual or mystical experiences? Mario Beauregard and Vincent Paquette of the University of Montréal have sought to explore this question by performing brain scanning on a group of volunteer Catholic Carmelite nuns.²⁰ The Carmelites are a contemplative order whose members spend most of their days in silent prayer and meditation, the ultimate goal of which is to achieve unio mystica, a state where one feels completely united with God. Because most Carmelites experience unio mystica only once or twice in a lifetime of contemplation, they can’t simply be hooked up to a brain scanner and commanded to summon the state. To get around this problem, the investigators instructed the nuns to remember and relive their most intense mystical experience. As a control condition they were also asked to remember and relive their most intense state of union with another human (while affiliated with the Carmelite Order).

Comparing brain scans in the mystical remembrance versus the control condition showed activity increases in a group of regions including the anterior cingulate cortex, the orbitofrontal cortex, and portions of the parietal cortex, but not core portions of the medial forebrain pleasure circuit. My own feeling about these results is that they actually have little to do with the neural activation pattern during unio mystica. Paquette and Beauregard argue that since trained actors can evoke activation of brain regions associated with particular emotions by recalling past emotional experiences, recalling unio mystica is a valid mirror of the primary experience. I remain unconvinced that this construction holds true for intense, euphoric experiences. Does doing your best to recall falling in love really feel like falling in love? Does recalling the experience of orgasm feel like orgasm? Of course not. And neither memory produces the patterns of brain activation that are seen in actual scans of subjects who are falling in love or having an orgasm. Given these conditions, it remains an open question whether ecstatic religious experience activates the medial forebrain pleasure circuit.

But it gets worse. In addition to claiming that the remembrance of mystical union produced the same pattern of brain activation as the experience itself, Beauregard, together with the journalist Denyse O’Leary, go on to claim that this experiment makes a strong case for “the existence of the soul.” As they explain, “To the extent that spiritual experiences are experiences in which we contact the reality of our universe, we should expect them to be complex.”²¹ Yes, nuns recalling unio mystica did show activation of a large number of brain regions. But if we examined other brain scanning papers in the biomedical literature, we’d find that many of them revealed that numerous brain regions lit up in response to all kinds of stimuli. If I hooked you up to a brain scanner and had you watch a Three Stooges film, you would also show a complex array of activated brain regions. Conversely, if mystical-union-recalling nuns had shown activation of few brain regions, would that count as evidence against the existence of the soul? Of course not. The soul may or may not exist, but this study by Beauregard and Paquette adds nothing to our understanding of the issue.

 

When I was in the first grade, I attended an after-school program at the Jewish Community Center in my hometown of Santa Monica, California. In the lobby was a large banner soliciting donations to the United Jewish Appeal that read “Give ’til It Hurts.” I didn’t understand it and found the whole thing vaguely disturbing, to the point that, whenever possible, I would navigate around the lobby so as to avoid even looking at the banner. Several months later it was replaced with a similar one—same font, same logo—that read, “Give ’til It Feels Good.” Freakin’ adults, I thought. Why does everything have to be so confusing?

All this came back to me when I read a recent paper by William Harbaugh and colleagues at the University of Oregon.²² The goal of their study was to measure activation of the nucleus accumbens, a neural substrate of pleasure/reward, to test economic models related to taxation and charitable giving. One theory holds that some individuals give to charity out of “pure altruism.” They feel satisfaction from providing a public good, like assistance to the needy, and they care only about how much benefit is offered and not the process by which it occurs. This model implies that these individuals should get some pleasure even when such a transfer of wealth is mandatory, as in taxation. A second theory, called “warm glow,” holds that people like making their own decision to give. They derive pleasure from the sense of agency, in much the same way that people like to roll their own dice while playing craps and pick their own lottery numbers. In this model, mandatory taxation is not expected to produce a “warm glow.” A third theory proposes that some people take pleasure in charitable giving because of its enhancement of their social status. They enjoy being regarded as wealthy or generous by their peers. Of course, these theories are not mutually exclusive. Someone could be motivated by pure altruism and the warm glow of agency and the desire for social approval.

Harbaugh and his team designed their experiment to address the first two theories, but not the third. They recruited nineteen young women from the area around Eugene, Oregon, and had them perform various economic transactions in a brain scanner. They were instructed that no one, not even the experimenters, would know their choices. (This was true: Their decisions were written directly to computer disk and machine-coded prior to analysis.) Presumably, the design of this experiment removes enhancement of social status as a motivator. Each subject received $100 in an account, which would then be allocated in various amounts to a local food bank. In some of the trials, the subjects had the option to donate; in others they had no choice—they were “taxed.” In other trials, they received money with no conditions. The way the study was carried out was as follows: The subjects first were presented an amount of money on a video screen, say $15 or $30. A few seconds later, they learned the status of the trial: This sum was either a gift to them, an involuntary tax on their account, or an offer to donate to charity, which they could either accept or decline by pushing one of two buttons. The brain scanning results showed over the entire population that, just like receiving money, both taxation and charitable giving activated nearly overlapping regions of the nucleus accumbens. However, on average, charitable giving produced a stronger activation of this pleasure center than did taxation (Figure 6.1). These results support both “pure altruism” and “warm glow” models as motivators of charitable giving.

Figure 6.1 The nucleus accumbens, a key part of the medial forebrain pleasure circuit, is activated by both mandatory payouts (taxes; top) and anonymous charitable giving (bottom). These activation patterns show substantial overlap. Adapted from W. T. Harbaugh, U. Mayr, and D. R. Burghart, “Neural responses to taxation and voluntary giving reveal motives for charitable donations,” Science 316 (2007): 1622–25, with permission from AAAS.

Of course, this doesn’t mean that these same subjects are smiling as they write their checks to the IRS, which supports many programs that may be less appealing than a food bank. It also doesn’t mean that everyone’s brain responds precisely the same way in such conditions. About half of the subjects in the study had more pleasure center activation from receiving money than from giving it, while the other half showed the opposite results. Not surprisingly, those who got more pleasure from giving did indeed choose to give significantly more to charity than the other group. A philosophical question arises from these findings: If giving—even mandatory, anonymous giving—is pleasurable, does that mean that “pure altruism” doesn’t really exist? In other words, if we catch a pleasure buzz from our noblest instincts, does that make them less noble?²³

 

While the experiments of Harbaugh and his colleagues masked the subjects’ choices in an attempt to eliminate issues of social status and approval, this obviously doesn’t reflect conditions in the real world. All of our behavior is embedded in a social context, and this social context powerfully influences our feelings and decisions. We’ve already mentioned how even mild social rejection can activate the emotional pain centers in the anterior cingulate cortex. Does this mean that positive social interactions can activate pleasure centers as well?

One form of positive social interaction is acceptance—a positive evaluation of the self by others. Norihiro Sadato and his coworkers at the National Institute for Physiological Sciences in Japan have sought to delineate the brain regions activated by one’s “good reputation” and compare that to the activation pattern produced by monetary reward. The monetary reward task was like many we have discussed before: Subjects in a brain scanner chose one card of three on a video screen and received various monetary rewards. This produced a pattern of activation similar to that seen in previous studies. The strongest activations were produced by the largest rewards, and these occurred in a number of regions, including the orbitofrontal cortex, the insula, the dorsal striatum, and the nucleus accumbens.

When the same subjects returned for a second day of testing, they took an extensive written personality survey and recorded a short video interview. Then they entered the scanner, where they received social feedback in the form of evaluations of their personality that had supposedly been prepared by a panel of four male and four female observers. To further the deception, they were shown photos of these observers and were told that they would meet them at the end of the experiment. The feedback took the form of a photo of the subject’s own face with a single-word descriptor underneath. Some of the descriptors were positive, such as “trustworthy” and “sincere,” while others were rather neutral, like “patient.” (All of these terms were actually in Japanese, not English.) Of course, these descriptors were all generated by the experimenters and presented in a randomized order. The main finding was that the most positive social reward descriptors activated portions of the reward circuit—most notably the nucleus accumbens and the dorsal striatum—that substantially overlapped with those activated in the monetary reward task. This finding suggests that there is, quite literally, a common neural currency for social and monetary reward.²⁴

 

In recent years, many social scientists have come to realize that social comparison can be an important factor in driving economic (and other) decisions of individuals. We evaluate our own economic circumstances and prospects not on some absolute scale, but rather in comparison to those of people around us. We already know that monetary reward can activate certain pleasure centers. If social comparison is indeed encoded in brain function, then it is reasonable to hypothesize that activation of pleasure centers should reflect socially relative rather than absolute levels of payout? To address this issue, Armin Falk and his colleagues at the University of Bonn conducted an experiment in which nineteen pairs of subjects were monitored in side-by-side brain scanners.²⁵ Each subject simultaneously performed a simple perceptual task: A field of dots was displayed on a video screen for 1.5 seconds, and immediately afterward a number (such as “24”) was shown. The subject had to choose, using a rapid button press, whether the dot count was higher or lower than the number. After a short delay, a feedback screen informed the subject about his performance and that of the other subject, along with respective monetary rewards given (e.g., He: 60 euros, You: 120 euros). The subjects received a payout only if they solved the task correctly: If both subjects failed, no payout was given. If only one was correct, he received either 30 or 60 euros, while the other got nothing. However, when both solved the task correctly (which occurred about 66 percent of the time), the computer would assign rewards randomly, from 30 to 120 euros. Sometimes the subjects would get the same reward, sometimes the rewards would be mildly disparate, and other times they would be highly disparate. This experiment showed that social comparison strongly influences activation of reward centers in the brain. The nucleus accumbens was most strongly activated in those trials in which there was a significant difference between one subject’s monetary reward and that of the adjacent subject. In other words, despite the biblical injunction of the Tenth Commandment, “Thou shalt not covet thy neighbor’s house/wife/slaves/ox/ass/plasma TV/Porsche/etc.,” we seem to be hardwired to compare our own experiences and circumstances to those around us.²⁶

 

We humans thrive on information. We love news, gossip, rumors, and, most important, information about our own future. What’s more, a number of studies by economists and psychologists have confirmed what we already know from experience: We want that information now, not later. Do monkeys also share this desire to know what the future holds? And if so, does this information activate the same VTA dopamine neurons that are stimulated by intrinsically pleasurable stimuli like food and water? In other words, is information about the future pleasurable in and of itself?

These interrelated questions were addressed in a series of experiments performed by Ethan Bromberg-Martin and Okihide Hikosaka at the National Eye Institute in Bethesda, Maryland.²⁷ They trained two thirsty monkeys to perform a simple decision task: Two targets appeared on the left and right sides of a video screen, and the monkey would choose one by flicking its eyes to a given target. Then, following a delay of a few seconds, the monkey would receive either a large or a small water reward. It didn’t matter which target the monkey chose—the rewards were delivered randomly, and at the same overall frequency. The twist to this experiment was that choosing one of the visual targets produced an informative cue during the delay period—a symbol whose shape indicated the size of the upcoming reward—while choosing the other target produced a random cue in the delay period that had no meaning, no predictive value. So in this design, it doesn’t matter whether the monkey chooses to receive advance information or a meaningless symbol: It still has the same chance of getting the large water droplet, and it will get it with the same delay.

Just like humans, when given the choice, the monkeys opted to receive information about the future. Within about ten trials, both monkeys were choosing the information-yielding target almost every time. Recordings made from individual dopamine neurons in the VTA and substantia nigra of the subjects revealed that these neurons briefly increased their firing rate when the monkeys saw the symbol that predicted a large amount of water, while the symbol that predicted a small amount of water briefly attenuated their ongoing firing rate. Crucially, after training, these same neurons were excited during probe trials in which the monkey saw only the target that indicated upcoming information and were inhibited when they saw the target that indicated upcoming random, uninformative symbols. The same dopamine neurons that signal the expected amount of pleasure from water also signal the expectation of information, even when that information cannot be put to any use. The monkeys (and presumably humans as well) are getting a pleasure buzz from the information itself.

To my thinking, this experiment is revolutionary. It suggests that something utterly useless and abstract—knowing merely for the sake of knowing—can engage the pleasure/reward circuitry. This is not pleasure obtained from essential things like food or water or sex, which we need to propagate our genes. Nor is it the pleasure of monetary reward, which, while abstract, still represents some real-world benefit, as it can be exchanged for useful things. Nor is it even like the pleasure of charitable giving or the pleasure of receiving positive social feedback, which can also be evolutionarily beneficial for animals living in certain types of social group.

This experiment suggests that ideas are like addictive drugs. As we have seen, certain psychoactive drugs co-opt the pleasure circuit to engage pleasurable feelings normally triggered by food, sex, and so on. In our recent evolutionary lineage (including primates and probably cetaceans), abstract mental constructs have become able to engage the pleasure circuitry as well, a phenomenon that has reached its fullest expression in our own species. The neuroscientist Read Montague, weaving together several strands of thought in cognitive neuroscience from a number of investigators, calls the human ability to take pleasure in abstract ideas a “superpower”²⁸ and I’m inclined to agree with him. From this perspective, human ideas can even directly oppose our most basic pleasure drives. For example, some people, acting on their religious principles, can forgo sexual activity in service to what they perceive as a more important goal. Likewise, the politically or spiritually motivated hunger-striker is activating her pleasure/reward center by furthering her own ideas, even when this requires acting in precise opposition to one of our most basic and ancient drives.

How does this superpower, in which ideas engage the pleasure circuit, develop on a cellular level? The short answer is that we really don’t know. The longer and more speculative answer is that it’s just the most recent and elaborate manifestation of the modification of the pleasure circuit by experience (or use-dependent plasticity of neurons, as we like to say in the cellular world). When sensory experience or internal states are represented in our brains by particular patterns of neuronal activity, these patterns of activity can produce changes in neuronal function, particularly electrical function. You’ll recall the discussion in chapter 2 of how certain patterns of stimulation can give rise to persistent increases or decreases in the strength of synaptic communication—LTP and LTD, respectively—but LTP and LTD are only one aspect of how experience can change the electrical signaling function of neurons. Furthermore, experience can be written into neuronal memory on different time scales. Some changes can be engaged by a single experience, while others require repetition. Some changes come on rapidly, within seconds, while others require many days. Some of these changes persist only briefly, and others can last a lifetime.

These experience-triggered changes in neuronal function occur throughout the brain, but for the purposes of this discussion, it is critical to emphasize that they take place in the neurons of the medial forebrain pleasure circuit and their immediate connections (the neurons that drive the pleasure circuit and those that are in turn driven by it). What this means is that the simple, hardwired pleasure engaged by ancient stimuli like sex and food can be transformed by experience into much more complex phenomena (Figure 6.2). When Schultz’s monkeys learn to associate a green light with an upcoming syrup droplet reward, they rapidly show increased firing of dopamine neurons time-locked to the presentation of this cue. Presumably, there are excitatory axons conveying green-light signals to the dopamine cells of the VTA, and the synapses between these axons and the dopamine cells undergo rapid-onset LTP to create the pleasurable association with the green light.²⁹

The same basic model could underlie the association of arbitrary stimuli (like money) or even abstract ideas with pleasure. If one imagines that an abstract idea is represented in the brain by particular patterns of neural activity, then those patterns could be conveyed to the pleasure circuit and drive changes within it. It is likely that this form of association would develop more slowly and also be more persistent, like a long-term memory. Long-term memory storage in the brain appears to be associated with microstructural changes in neuronal wiring, and so this form of experience-driven neuronal change is likely to be required for linking abstract ideas to pleasure. Finally, we have seen in chapter 2 how drug addiction can slowly and persistently change the function of the pleasure circuit and thereby transform pleasure and liking into wanting and craving. This is also a long-term process and has been shown to involve changes to neuronal structure, such as increases in dendritic spine density (Figure 2.5). While it is likely that similar changes in the pleasure circuit accompany the development of behavioral addictions, there has been little work to date to test that hypothesis.

Figure 6.2 The transformation of simple, hardwired pleasure by experience. Pleasure can be transformed through associative learning processes in the medial forebrain circuitry to yield all sorts of phenomena, both beneficial and detrimental. Rapid-onset associative learning in the pleasure circuit can give rise to reward prediction, as with Schultz’s monkeys. Repeated association and the attendant slower-onset but longer-lasting changes in circuit function can give rise to arbitrary rewards (like money) and even pleasure driven by ideas, Read Montague’s human “superpower.” Finally, in some cases, repeated activation of the pleasure circuit by certain drugs or behaviors can give rise to addiction, in which liking is transformed into wanting, and tolerance, withdrawal, and cravings emerge. Illustration by Joan M. K. Tycko.

In sum, the interaction of pleasure and associative learning in our human brains is the classic two-edged sword: The ability of experience to produce long-term changes in the pleasure circuit has enabled arbitrary rewards and abstract ideas to be felt as pleasurable, a phenomenon that ultimately underlies much of human behavior and culture. Unfortunately, that same process allows pleasure to be transformed into addiction.