فصل 3

3

THE CHEMISTRY OF SEDUCTION

You have just emerged from your mother’s womb into a hospital room full of strangers, bright lights, and machines. Bewildered by the multitude of new sights and sensations, you begin to cry. At this point, crying is one of the few things you know how to do, part of a small repertoire of instinctive behaviors like suckling. Over the course of your life, however, you’ll develop the desire and ability to play with blocks, read written words, hit a baseball, kiss another person, hold a job, and acquire and eat everyday foods. This striking behavioral transformation is due to a phenomenon, often taken for granted, called learning. Learning is the process of acquiring new knowledge, skills, movement patterns, motivations, and preferences, or reinforcing those that already exist. As it turns out, learning—particularly the effects of learning on our motivation to seek certain foods—is one of the key reasons we overeat, despite our better judgment.

In order to learn, you must start with a goal. If you don’t have a goal, you can’t determine which behaviors are more valuable than others and, therefore, which should be cultivated. In the evolutionary sense, the ultimate goal of any organism is to maximize its reproductive success: have as many high-quality offspring as possible, which in turn will produce as many high-quality offspring as possible. But that’s not the goal we’re thinking about when we dig into a bowl of cereal; in fact, we’re rarely, if ever, aware of it. What we are aware of is a variety of proximate goals that natural selection has hardwired into our brains as a shorthand for the ultimate goal of reproductive success. For most animals, these goals include obtaining food and water, mating, seeking safety from danger, and seeking physical comfort. Humans, being more complex and social than most other animals, also seek social status and material wealth (although we can’t lay exclusive claim to this; many social animals, including chimpanzees, use favors, sex, and violence to climb the social ladder). These goals—to eat, to drink, to have sex, to be safe and comfortable, to be liked—are the fundamental drivers of motivation and learning. Because food is so important for survival and reproduction, it tends to be a very powerful teacher.

When we hear the word learning, we tend to imagine ourselves poring over a textbook absorbing facts, but nearly everything we do—and think and feel—was learned at some point, whether intentionally or not. Roy Wise, a motivation and addiction researcher at the National Institute on Drug Abuse in Baltimore, Maryland, made this point in a 2004 review paper: Most goal-directed motivation—even the seeking of food or water when hungry or thirsty—is learned. It is largely through selective reinforcement of initially random movements that the behavior of the [newborn] comes to be both directed and motivated by appropriate stimuli in the environment.

To give you an example, imagine an infant trying to grab the tail of a cat sitting in front of him. He’s not coordinated enough, and he mostly flails around, sometimes touching the tail but having a hard time grasping it. Suddenly, by chance, his arm and hand move in just the right way, and he grabs the cat’s tail for a brief moment. Realizing that something good just happened, his brain increases the likelihood that the same movement pattern will recur the next time he wants to grab the cat’s tail. With practice, his brain refines the movement, until he can terrorize the cat at will. More generally, when something good happens, the brain increases the likelihood that the pattern of brain activity that immediately preceded the good event will recur in the future. To put this in terms of what we covered in the last chapter, the successful option generators get a stronger bid.

From the outside, we observe that when a behavior meets a goal, it is more likely to recur in the future—it’s reinforced. The noted American psychologist Edward Thorndike described the phenomenon of reinforcement as early as 1905, stating that “any act which in a given situation produces satisfaction becomes associated with that situation so that when the situation recurs the act is more likely than before to recur also.” Over the course of our lives, with experience, we refine our ability to achieve our goals, and reinforcement is one of the simplest and most powerful means of doing so.

To further illustrate this, let’s return to the example of getting yourself something to eat. To satisfy your hunger, you activated option generators representing the restaurant up the street, getting on your bike, and pedaling. This is the pattern of motivations, thoughts, and movements that got you to the restaurant. Now, let’s say you ate at the restaurant and the food was really good—unexpectedly good. You achieved your food goal very effectively. The option generators that got you to the restaurant will be strengthened, and the next time you’re hungry, you’ll be more likely to crave the restaurant up the street, and perhaps even hop on your bike to get there. You’ll come to enjoy the thought, appearance, and smell of the restaurant. Your behavior of going to the restaurant up the street will be reinforced.

Learning shapes all three levels of our decision-making process: motivational, cognitive, and motor. Reinforcement strengthens them all, because they’re all required for effective goal-directed behaviors. The process of reinforcement operates completely outside our conscious awareness, and it has existed since before the time of our common ancestor with lampreys.

Learning also operates in the opposite direction. When something bad happens as the result of a behavior, the likelihood of that behavior recurring decreases. For example, if you develop food poisoning after eating at the restaurant up the street, you’ll be less likely to eat there again. You won’t crave the restaurant when you’re hungry, and the thought, appearance, and smell of the restaurant may even provoke a feeling of disgust. This is called negative reinforcement.

For reinforcement to occur, there has to be a teaching signal that changes the activity of basal ganglia loops based on experience, such that good responses are reinforced and bad responses are discarded. Most researchers believe the brain’s teaching signal is the fascinating molecule dopamine.

THE LEARNING CHEMICAL

Ross McDevitt, a postdoctoral researcher at the National Institutes of Health in Baltimore, Maryland, gently places a mouse into a clear plastic cage and presses a thin fiber-optic cable into a tiny connector on the animal’s head. McDevitt is using a cutting-edge technique called optogenetics to specifically stimulate cells in the mouse’s ventral tegmental area (VTA). As we discussed in the last chapter, the VTA is a brain region that sends dopamine-laden fibers to the primary motivation center of the brain: the ventral striatum. When these fibers release dopamine, they change the activity of cells in the ventral striatum and related brain regions, with profound consequences for behavior. Previously, we saw that high overall levels of dopamine can increase the likelihood that any option generator will grab the reins of behavior—yet it has other effects that are far more elegant. Dopamine, as a matter of fact, is the essence of reinforcement.

The end result of McDevitt’s experimental setup is that he can send bursts of dopamine into the ventral striatum with the flick of a switch, which will illustrate the remarkable power of this pathway in learning and motivation.

The mouse’s cage has a little box built into it, and each time the mouse pokes its nose into the box, light travels down the fiber-optic cable on its head, shines onto its VTA neurons, and causes them to release bursts of dopamine into the ventral striatum and related brain regions. Yet the mouse doesn’t know any of this at the beginning of the experiment. When the mouse is first placed into its cage, the little box is utterly meaningless. The mouse has no particular affinity for the box, and it pokes its nose into it only occasionally, out of curiosity. Each time the mouse nose-pokes, however, it experiences the mouse equivalent of eating chocolate, having sex, and winning the lottery all at the same time.

It doesn’t take long for the nose-poking behavior to become more frequent. And more frequent. “What we find,” explains McDevitt, “is that the mice go crazy for it. They love it.” Whereas initially the mouse nose-pokes only out of curiosity, eventually it comes to understand the remarkable significance of the little box. McDevitt’s mice end up nose-poking eight hundred times per hour—ignoring practically everything else in their cages. Other labs have shown that rats will nose-poke up to five thousand times per hour for VTA stimulation, which is more than once per second! In other words, dopamine in the ventral striatum is highly reinforcing.

On a cellular level, this happens because dopamine acts on basal ganglia loops that were recently active, increasing the likelihood that they will be activated again in the future. So whatever you’re doing when the dopamine hits, you’re more likely to repeat it when the same situation arises again. The VTA basically says, “I like what just happened; I’m going to sprinkle some dopamine into the ventral striatum to make sure it happens again next time.” Although McDevitt evoked an exaggerated form of reinforcement by directly stimulating the VTA, this process happens naturally in our brains every day. When you accomplish a goal like eating a triple bacon cheeseburger, dopamine is released in short bursts that reinforce your “successful” behavior. This is how dopamine teaches us how to feel, think, and behave in ways that help us achieve our hardwired goals—whether or not our conscious, rational brains support them. Dopamine in the ventral striatum is particularly important for learning motivations: for example, learning which foods to crave and which to avoid.

Although dopamine wasn’t discovered until a half century after his initial experiments, the Russian physiologist Ivan Pavlov was one of the first researchers to describe how animals learn to associate neutral cues with food. Pavlov’s team studied digestion in dogs, and he quickly learned that his experimental animals drooled when they saw food—an observation that many dog owners can corroborate. Pavlov also noticed, with some irritation, that the dogs would salivate even if he didn’t have any food. They had learned to associate Pavlov’s presence with food.

Later, Pavlov’s team found that when they consistently rang a bell prior to feeding the dogs, the dogs eventually salivated in response to the bell alone. The dogs had learned to associate the sound of the bell with the delivery of food, and because of this, the previously neutral sound cue had acquired importance. This is the same process that caused McDevitt’s mice to learn the importance of the initially boring nose-poke box as it was repeatedly associated with a powerful reward (dopamine in the ventral striatum).

DOPAMINE: THE PLEASURE CHEMICAL?

You may have heard that dopamine is the “pleasure chemical,” responsible for causing the neurochemical rush that makes us feel really good as we win a race, have sex, eat chocolate, or smoke crack cocaine. Although that idea is common in popular science writing, it has long been out of date within the scientific community. As a matter of fact, dopamine release doesn’t correspond very well with the experience of pleasure. Experiments have shown that animals appear to experience pleasure without dopamine, and studies in humans back this up. Pleasure is more related to a class of chemicals called endorphins that are often released in the striatum simultaneously with dopamine, although these are probably only one component of the experience of pleasure. Dopamine is the “learning chemical” rather than the “pleasure chemical.” Today, we call this process Pavlovian conditioning, and it’s the reason why the sight of cola on television, a little taste of ice cream, or the smell of french fries can cause us to crave—and also to salivate. After you’ve learned that the sight and smell of french fries predicts a fatty, starchy reward in your belly (with the help of dopamine), these sensory cues acquire importance and stimulate your motivation to eat french fries when you encounter them. Yet not all foods have this seductive effect on us. Why not?

THE CALORIE-SEEKING BRAIN

At this point, we know food can be a powerful reinforcer, tugging on the reins of our behavior. But some foods do this more than others. Brussels sprouts, for example, are a lot less seductive than ice cream. To understand why we overeat, we need to first answer the fundamental question: What is it about food, exactly, that’s reinforcing? Anthony Sclafani, the researcher who invented the “cafeteria diet” we encountered in chapter 1, has dedicated the majority of his career to this question, and he’s made remarkable progress.

Your average lab rat likes cherry-flavored water about as much as grape-flavored water. So if you place a bottle of each into its cage, it will drink about the same amount from each. However, in a groundbreaking study published in 1988, Sclafani’s team showed that when they infused partially digested starch directly into the rats’ stomachs while they drank cherry-flavored water, the rats developed a preference for that flavor over the grape flavor. And the opposite preference developed when they repeated it with the grape flavor. Even though the starch never entered their mouths, after four days, the rats displayed a near-total preference for the starch-paired flavor. Sclafani called this phenomenon conditioned flavor preference. Conditioned flavor preferences are a remarkable phenomenon. The rats were totally unaware that researchers were infusing starch directly into their stomachs, yet somehow this starch sent a signal to the brain that caused the rats to increase their preference for a flavor they detected concurrently. In essence, they learned to prefer that previously neutral flavor just like Pavlov’s dogs learned to salivate at the sound of a bell. How did this happen?

Further experiments showed that the rats weren’t detecting the starch itself but the sugar glucose that’s released as starch is broken up in the digestive tract. And the critical location for detection was the upper small intestine. Somehow, the intestine was sensing the glucose and sending a signal to the brain that said, “Something good just happened. Do that again!” Sclafani still isn’t certain how the signal gets from the gut to the brain, but he and other researchers have already figured out how the signal influences flavor preferences once it gets there. The obvious suspect is—you guessed it—dopamine in the ventral striatum. Ivan de Araujo, associate professor of psychiatry at the Yale School of Medicine, and colleagues demonstrated independently of Sclafani that calories infused into the small intestine increase dopamine levels in the ventral striatum, and the more calories they infuse, the more dopamine spikes. Consistent with de Araujo’s findings, Sclafani’s group found that blocking dopamine action in the ventral striatum prevents the development of conditioned flavor preferences. “This suggests,” says Sclafani, “that dopamine may be a central component of the process.” This research paints a nearly complete picture of how carbohydrate conditions flavor preferences. As a rat eats food, its mouth and nose detect the flavors and aromas associated with the food. After the rat swallows the food, it enters the rat’s stomach and then its small intestine. The small intestine detects glucose and sends an unknown signal to the brain that causes dopamine to spike in the ventral striatum. If the food is rich in starch or sugar, a large spike in dopamine causes the rat to increase its preference for the flavors and aromas of the food it just ate—and become more motivated to seek foods with those flavors and aromas in the future. In this way, the rat becomes better at identifying and seeking foods that contain carbohydrate.

Sclafani’s team has also been able to condition flavor preferences using fat and protein, demonstrating that rats respond to all three major classes of calorie-containing nutrients: carbohydrate (starches and sugars), fat, and protein. Sclafani’s work also revealed that the more concentrated the caloric load of a food—or the higher its calorie density—the more reinforcing it is. Apparently, the rat brain evolved not just to seek carbohydrate but to seek all types of calories—particularly foods that deliver the most calories per bite. Sound familiar?

Flavors and smells are a quick way for the brain to gather information about the nutritional quality of a food before it enters the digestive tract. Sclafani and other researchers have shown that certain flavors on the tongue can enhance conditioned flavor preferences. For example, rats form stronger conditioned flavor preferences if they can taste the sugar as they drink it, rather than having it infused directly into the stomach. The sugar’s effects on the tongue combined with its effects in the small intestine cooperate to reinforce behavior. Other flavors work similarly. The meaty umami flavor associated with the amino acid glutamate (the principal component of monosodium glutamate, or MSG) increases food desirability in rats just as in humans, and glutamate also conditions flavor preferences when administered into the stomach. Sweetness indicates ripe fruit, and glutamate indicates protein-rich food like meat—both of which are important sources of calories and other nutrients in the wild. Conversely, bitter flavors, odors of decay, and any food that has previously caused digestive distress are aversive. Calories don’t just drive flavor preferences; they also drive preferences for the aromas, sights, sounds, and even locations that predict the availability of calories. It turns out that rats like to hang out in places where good things happen, and calories in the belly is definitely a good thing. This is how we learn to respond to our surroundings in a way that gets us what we want.

Together, this shows that animals don’t seek food indiscriminately—they seek specific food properties that the brain instinctively recognizes as valuable—and most of these properties indicate high-calorie foods. These are presumably the nutrients that are critical for survival and reproduction in a natural environment, explaining why the rat brain responds to them by releasing dopamine and reinforcing the behaviors of seeking and eating them. The rat brain instinctively values nutrients that keep wild rats healthy and fertile, and it gradually learns how to obtain them efficiently, under the tutelage of dopamine.

CONDITIONING CULTURE

If you’re starting to feel like you have a lot in common with rats, you’re right. Humans share most of our innate food preferences with rats, which makes sense if you think about it: We’re both omnivores that have been eating human food for many generations. Humans and rats are born liking sweet and disliking bitter flavors, suggesting that these are deeply wired adaptations that may have evolved prior to the divergence of our species seventy-five million years ago. In addition, people of all cultures enjoy the meaty flavor of glutamate, dislike odors of decay, and are repulsed by foods that have previously caused digestive distress. Humans also have a strong preference for salt (sodium chloride)—one inclination we don’t share with rats.

INNATE PREFERENCES

INNATE AVERSIONS

Calorie density

Fat

Carbohydrate

Protein

Sweet flavor

Salty flavor

Meaty flavor (umami)

Bitter flavor

Odor of decay

Anything that causes digestive distress

Our gastronomic similarity to rats is supported by the research of Leann Birch, a childhood obesity researcher at the University of Georgia, whose research team set out to see if they could extend Sclafani’s results to humans. Their findings, and those of other researchers, show that certain nutrients, especially fat and carbohydrate, can indeed condition flavor preferences in our own species. In other words, they reinforce behavior.

In the box above, I’ve listed the innate food preferences and aversions that all humans share (known or suspected). These are the food properties that shape our eating behavior by causing dopamine to spike in our brains, and thereby driving us to seek the flavors, smells, textures, appearances, and places we’ve learned to associate with food that contains those properties.

People in China and in France enjoy sugar, salt, fat, and meaty flavors. Of course, a French person may not have much appreciation for authentic Chinese food, and a Chinese person may find the odor of strong French cheese repulsive. This is because each culture has its own set of unique flavor, smell, texture, and appearance preferences that develop as the result of reinforcement learning. Just as rats aren’t born preferring cherry flavor over grape flavor, we aren’t born appreciating the flavors and smells unique to our culture: We develop them as conditioned preferences. These food properties become enjoyable and motivating by virtue of their repeated association with innately reinforcing properties, such as fat and carbohydrate, particularly in childhood.

When you examine this list, you’ll find it apparent that the human brain is extremely preoccupied with calories. Other than salt, every innate preference is a signal that indicates a concentrated calorie source. This is presumably because calorie shortage was a key threat to our ancestors’ reproductive success, so we’ve been wired by evolution to value high-calorie foods above all others. Consistent with this, modern-day hunter-gatherers living in a manner similar to how our distant ancestors would have lived don’t spend much time or effort gathering wild brussels sprouts. They primarily spend their time looking for calorie-rich foods that can sustain their energy needs, like nuts, meats, tubers, honey, and fruit—a topic we’ll return to later.

Our innate food preferences readily explain why children like ice cream and not brussels sprouts. As far as the ventral striatum is concerned, the fact that brussels sprouts are loaded with vitamins and minerals counts for approximately nothing, because they scarcely deliver any calories. In contrast, we crave ice cream because our brains know that its flavor, texture, and appearance predict the delivery of a truckload of easily digested fat and sugar. Having evolved in an era of relative food scarcity, the human brain interprets this as highly desirable and draws us toward the freezer.

Nonconscious parts of the brain perceive certain foods as so valuable that they drive us to seek and eat them, even if we aren’t hungry and even in the face of a sincere desire to eat a healthy diet and stay lean. We crave dessert even after a large meal. We crave a soda with lunch. We crave that extra slice of pizza. Willpower often bows before the power of dopamine-reinforced sensory cues, such as the dessert menu on the table, the sight of the soda machine, or the aroma of pizza. And sometimes, the reinforcement process spirals out of control.

VERY STRONG HABITS

Although reinforcement is a natural process that evolved to guide us through life successfully, it can sometimes get out of hand. As McDevitt’s experiment demonstrated, when dopamine reaches very high levels in the ventral striatum, it reinforces behaviors so strongly that they can take priority over natural, constructive behaviors. This is the essence of addiction. It makes sense, then, that every known addictive drug either increases dopamine levels in the ventral striatum or stimulates the same signaling pathway in a different manner. Even relatively benign habit-forming drugs, such as caffeine, seem to act on the same pathway. As a result of repeated dopamine stimulation of the ventral striatum, obtaining and using highly addictive drugs like crack cocaine can become such a high priority that it outweighs food, safety, comfort, and social relationships. The option generators responsible for seeking and smoking crack put in pathologically strong bids to the striatum, overwhelming most other bids.

According to Roy Wise, the addiction researcher we encountered earlier in this chapter, addiction is simply an exaggerated version of the same reinforcement process that occurs in everyday life. “Addiction just happens to be a very strong habit,” says Wise, “because addictive drugs are very strong reinforcers.” So is it possible to become addicted to food? Food triggers dopamine release in the ventral striatum, just like drugs of abuse, and we know food can be strongly reinforcing. Yet this idea remains extremely controversial among researchers. How is it possible to become addicted to a substance that’s essential to the body? Are we addicted to water and oxygen as well? Everything else that’s good in our lives, from sex to a new car to a job well done, also presumably provokes dopamine release in the ventral striatum. Are we addicted to everything?

We aren’t, of course, addicted to everything. Most people have a constructive relationship with most of the good things in their lives—a relationship that can hardly be described as addiction. Yet research conducted by Ashley Gearhardt and Kelly Brownell, formerly at Yale University, suggests that some people may indeed be addicted to food. Brownell’s team examined the standard diagnostic criteria for addiction to nonfood reinforcers, such as drugs, sex, and gambling. They used these criteria to design a questionnaire that identifies addiction-like eating behaviors, particularly focusing on loss of control, eating certain foods despite negative consequences, and withdrawal symptoms. The questionnaire includes such items as “There have been times when I consumed foods so often or in such large quantities that I started to eat food instead of working, spending time with my family or friends, or engaging in other important activities and recreational activities I enjoy” and “I eat to the point where I feel physically ill.” Among their first sample of people, most of whom were lean, 11 percent met the criteria for food addiction. Further research revealed that people who meet food addiction criteria are more likely to have obesity and more likely to exhibit binge-eating behavior. This supports the idea that the reinforcing effect of food can lead to addiction-like behavior, overeating, and weight gain in susceptible people. However, not all people with obesity meet food addiction criteria, and not all people who meet food addiction criteria have obesity, so it only offers a partial explanation for the obesity epidemic.

To understand food addiction, we need to examine the types of foods that trigger addiction-like behavior. As it turns out, people don’t become addicted to celery and lentils. What foods are they drawn to instead? The following quote by Gearhardt and Brownell sheds light on the question: High concentrations of sugar, refined carbohydrates (bread, white rice, pasta made with white flour), fats (butter, lard, margarine), salt, and caffeine are addictive substances and the foods containing these ingredients may be consumed in a manner consistent with addictive behavior. Just like drugs of abuse, these food substances may not be addictive until they are processed, extracted, highly refined and concentrated by modern industrial processes; meanwhile, combinations of these look-like-food substances may greatly enhance their addictive qualities.

Although it’s controversial to call these foods addictive, it is fair to say that they provoke dopamine release in the ventral striatum. And the more concentrated they are, the more dopamine they release. The more dopamine they release, the more they reinforce behavior, and the more they reinforce behavior, the closer they bring us to addiction.

What makes this alarming is that modern food technology has allowed us to maximize the reinforcing qualities of food, making it far more seductive than ever before in human history. We now have extremely calorie-dense, carefully engineered combinations of sugar, fat, salt, and starch that would have been inconceivable to our hunter-gatherer ancestors, who had no choice but to eat simple wild foods. Certain modern foods likely provoke a larger release of dopamine than the human brain evolved to expect, leading to destructive addiction-like behaviors in susceptible people. (Yet not everyone is susceptible—a topic we’ll return to shortly.) As Gearhardt and Brownell alluded to, this is similar to how drugs of abuse are often concentrated versions of less-addictive naturally occurring substances. For example, the leaves of the coca plant are widely chewed in South America as a mild stimulant and appetite suppressant reminiscent of caffeine. However, when we extract and concentrate the active ingredient of the coca leaf, this results in a much more addictive substance: cocaine. A secondary chemical process called freebasing transforms cocaine into the extremely addictive drug crack cocaine. Human technology has allowed us to concentrate and enhance the property of the coca plant that increases dopamine release and reinforces behavior, transforming it from a useful herb into a life-destroying drug. Similarly, modern food technology has allowed us to concentrate the reinforcing “active ingredients” in food to an unprecedented degree, and addiction-like behavior in a subset of people is the predictable result.

Chocolate is a prime example of the riot of reinforcing properties that characterizes modern foods. The seeds of the cacao tree, a plant native to tropical South America, are naturally extremely calorie dense, due to their high fat content. When fermented, roasted, and ground into a paste, these seeds become chocolate: a magical substance that’s solid at room temperature and melts in the mouth. To mask the naturally bitter flavor of chocolate, we add a generous dose of refined sugar, and sometimes dairy. The calorie density, fat content, carbohydrate content, and sweet taste of chocolate are a powerfully reinforcing combination, but chocolate has another trick up its sleeve that makes it the king of cravings: a habit-forming drug called theobromine. Theobromine is a mild stimulant that’s moderately reinforcing, like its cousin caffeine. Although theobromine on its own may not be the bee’s knees, when added to a substance that’s already highly reinforcing, it puts many of us over the edge. It may come as no surprise that chocolate addiction is a legitimate topic of scientific research. Even those of us who aren’t “chocoholics” may experience chocolate cravings, and research indicates that chocolate is the most frequently craved food among women.

Although most of us aren’t literally addicted to food, recall that addiction is simply an exaggerated version of the same reinforcement process that happens in all of us. We may not be so seduced by food that it interferes with our work or family lives, but most of us are driven to eat more calories than we should, acting against our own self-interest despite our better judgment.

CONTROLLING CRAVINGS

How can we fight this instinctive force that compels us to eat too much? Drug addiction research gives us important clues. One of the best-established treatment strategies for drug addiction is simply to avoid drug-associated cues. We know from the work of Ivan Pavlov and others that sensory cues that are repeatedly associated with positive outcomes become motivational triggers. When an addicted person sees a crack pipe, smells crack smoke, or walks by the street where he usually buys crack cocaine, it triggers his motivation to smoke crack—and the urge can be overwhelming. Whether or not you’re addicted to food, when you walk by a bakery and see freshly baked pastries and smell the aroma wafting into your nose, it triggers your motivation to eat pastries (or whatever food you prefer). That’s simply the nature of reinforcement. Yet if you don’t walk by the bakery and don’t experience its sensory cues, your motivation to eat pastries will be much lower, and you won’t have to fight yourself to avoid eating a fattening food.

We may have a hard time fighting our nonconscious urges to eat tasty, calorie-rich foods when they’re right in front of us, but with a little bit of advance planning, we can prevail without having to exert too much of our limited willpower. The key is to control food cues in your personal environment. Ultimately, a little bit of wise planning can go a long way.

Reinforcement is a nonconscious process, explaining why it isn’t very intuitive, but it associates with a process that we know quite well: pleasure.

PLEASURES OF THE PALATE

Food that brings us pleasure when we eat it is described as palatable. Palatable food tastes good. It’s a sign that the brain values a food, either as a result of instinct or reinforcement learning.

The brain presumably values certain food properties above others because they would have increased the reproductive success of our ancestors. The most highly palatable foods tend to be dense in easily digested calories and combine multiple innately preferred food properties in highly concentrated form: ice cream, cookies, pizza, potato chips, french fries, chocolate, bacon, and many others. These are the foods that are most likely to cause cravings and a loss of control over eating, because their physical properties make them exceptionally reinforcing, motivating, and palatable. Researchers have an umbrella term for this combination of effects on the brain: food reward. Highly rewarding foods are those that seduce us.

Not surprisingly, research suggests that people eat larger quantities of foods they like. John de Castro, a professor of psychology at Sam Houston State University, and his research team demonstrated that people living their everyday lives eat 44 percent more calories at meals they describe as highly palatable than at meals they describe as bland. The brain perceives these foods as so valuable that it motivates us to keep eating them even if we have no particular need for energy—in fact, even if we’re already drowning in energy.

What happens to food intake and adiposity when researchers dramatically restrict food reward? In 1965, the Annals of the New York Academy of Sciences published a very unusual study that unintentionally addressed this question. Here’s the stated goal of the study: The study of food intake in man is fraught with difficulties which result from the enormously complex nature of human eating behavior. In man, in contrast to lower animals, the eating process involves an intricate mixture of physiologic, psychologic, cultural and esthetic considerations. People eat not only to assuage hunger, but because of the enjoyment of the meal ceremony, the pleasures of the palate and often to gratify unconscious needs that are hard to identify. Because of inherent difficulties in studying human food intake in the usual setting, we have attempted to develop a system that would minimize the variables involved and thereby improve the chances of obtaining more reliable and reproducible data.

The “system” in question was a machine that dispensed liquid food through a straw at the press of a button—7.4 milliliters per press, to be exact. Volunteers were given access to the machine and allowed to consume as much of the liquid diet as they wanted, but no other food. Since they were in a hospital setting, the researchers could be confident that the volunteers ate nothing else. The liquid food supplied adequate levels of all nutrients, yet it was bland, completely lacking in variety, and almost totally devoid of all normal food cues.

The researchers first fed two lean people using the machine—one for sixteen days and the other for nine. Without requiring any guidance, both lean volunteers consumed their typical calorie intake and maintained a stable weight during this period.

Next, the researchers did the same experiment with two “grossly obese” volunteers weighing approximately four hundred pounds. Again, they were asked to “obtain food from the machine whenever hungry.” Over the course of the first eighteen days, the first (male) volunteer consumed a meager 275 calories per day—less than 10 percent of his usual calorie intake. The second (female) volunteer consumed a ridiculously low 144 calories per day over the course of twelve days, losing twenty-three pounds. The investigators remarked that an additional three volunteers with obesity “showed a similar inhibition of calorie intake when fed by machine.” The first volunteer continued eating bland food from the machine for a total of seventy days, losing approximately seventy pounds. After that, he was sent home with the formula and instructed to drink 400 calories of it per day, which he did for an additional 185 days, after which he had lost two hundred pounds—precisely half his body weight. The researchers remarked that “during all this time weight was steadily lost and the patient never complained of hunger.” This is truly a starvation-level calorie intake, and to eat it continuously for 255 days without hunger suggests that something rather interesting was happening in this man’s body. Further studies from the same group and others supported the idea that a bland liquid diet leads people to eat fewer calories and lose excess fat. This machine-feeding regimen was just about as close as one can get to a diet with zero reward value and zero variety. Although the food contained sugar, fat, and protein, it contained little odor or texture with which to associate them. In people with obesity, this diet caused an impressive spontaneous reduction of calorie intake and rapid fat loss, without hunger. Yet, strangely, lean people maintained weight on this regimen rather than becoming underweight. This suggests that people with obesity may be more sensitive to the impact of food reward on calorie intake. Is this because heightened food reward sensitivity causes obesity or because obesity increases food reward sensitivity? This question would require further research to disentangle. In 2010, Chris Voigt, the director of the Washington State Potato Commission, decided to eat nothing but potatoes and a small amount of cooking oil for sixty days. Voigt was protesting a decision by the federal Women, Infants, and Children food assistance program to remove potatoes from the list of vegetables it will pay for. Voigt contended, correctly, that potatoes are actually quite nutritious—in fact, one of the few foods that provide a broad enough complement of nutrients to sustain a human in good health for months at a time. He documented his journey on a Web site titled 20 Potatoes a Day, which refers to the number of potatoes he would have to eat to maintain his weight. Voigt signed himself up for two months of a starchy, bland, repetitive diet.

Despite Voigt’s goal not to lose weight, the pounds melted off. He lost twenty-one pounds over the sixty-day period, much of it from his waistline. According to physical examinations before and after, his blood glucose, blood pressure, and cholesterol levels improved considerably. He had trouble eating enough to meet his energy needs because he simply wasn’t hungry. While we may be tempted to question the veracity of a person whose job involves promoting potatoes, Voigt’s experiment triggered an avalanche of Internet copycats who used the “potato diet” for rapid weight loss. Although anecdotal, their reports suggest that eating this bland, repetitive diet does indeed reduce spontaneous calorie intake without provoking hunger. Yet this isn’t just because of the blandness of an all-potato diet—there’s more to the story.

THE BUFFET EFFECT

“Eat a varied diet” is a maxim that lies at the foundation of our modern approach to health. If we eat a large variety of different foods, we’re likely to meet our overall nutritional needs. While this principle is sound, it also has a dark side: Food variety has a powerful influence on our calorie intake, and the more variety we encounter at a meal, the more we eat.

The effect of food variety on food intake relates to a fundamental property of the nervous system called habituation. Habituation is the simplest form of learning—one shared by all animals with a nervous system. It probably evolved along with the first nervous systems nearly seven hundred million years ago, since it exists in our truly ancient relatives the jellyfish.

Habituation is one of the key tricks we use to distinguish important events from unimportant noise, and it’s simply this: The more we’re exposed to a stimulus within a short period of time, the less we respond to it. This was demonstrated in a classic series of experiments in human infants. An infant was seated in his mother’s lap, while a screen in front of him intermittently displayed a black-and-white checkerboard pattern. Researchers recorded the amount of time the infant’s gaze remained fixated on the pattern each time it appeared. As any parent might have guessed, the infants paid a lot of attention to the pattern at first and then spent less time fixating on it with each exposure. When a stimulus is new, we tend to be very interested in it because it might be important. Once we’ve seen it many times in a short period of time, it’s less likely to be important, and we stop paying attention.

As it turns out, this habituation process operates each time we sit down to a meal. In a pioneering 1981 study by Barbara Rolls and colleagues, volunteers rated the palatability of eight different foods by tasting a small amount of each, and then were provided one of the foods for lunch. After lunch, they once again rated the palatability of the same eight foods by tasting them. Rolls found that the palatability rating of the food the volunteers had eaten for lunch decreased much more than the palatability rating of the other seven foods they hadn’t eaten. When the volunteers were presented with an unexpected second course containing all eight foods, they tended to eat less of the food they had just eaten for lunch. This shows that we can eat our fill of a specific food and feel totally satisfied, but that doesn’t mean we won’t eat other foods if they’re available. Rolls called this phenomenon sensory-specific satiety. Satiety is the sensation of fullness we get after we eat food, and sensory-specific means this fullness only applies to foods that have similar sensory properties (sweet, salty, sour, fatty) to the ones we just ate.

BEATING THE BUFFET EFFECT

The fact that sensory-specific satiety drives us to overeat suggests a simple solution to the problem: Limit yourself to a few foods. If you find yourself at a buffet, tapas restaurant, or similar situation in which high food variety may cause you to overeat, simply choose three items you think would make a satisfying meal, and stick to them. You’ll probably feel just as full on fewer calories.

Several independent researchers using various methods have confirmed that we tend to eat more total food—and gain weight—when we’re presented with a large variety of foods. This goes a long way toward explaining what researchers call the buffet effect. We tend to overeat spectacularly at buffets, despite the fact that the food isn’t always the crème de la crème. At a buffet, we don’t have the opportunity to habituate to any particular food, because every few bites, we’re eating something new. The brain’s satiety system eventually throws the emergency brake, but not before we’ve eaten far too much.

Sensory-specific satiety also helps explain why we’re happy to eat dessert even after a large meal. We’re no longer hungry for savory food at all, yet when the dessert menu appears, we suddenly grow a “second stomach.” We’re satiated of savory foods, but we aren’t satiated of sweets. A novel sensory stimulus with an extremely high reward value makes it easy to pack away an additional 200 Calories of dessert. So it makes sense that the converse is also true, as we saw with the potato diet: When food reward and variety decrease, so does food intake.

SMOKING JOINTS FOR SCIENCE

“Numerous anecdotal accounts indicate that marijuana increases appetite and food intake in humans,” begins a 1988 paper by substance abuse researcher Richard Foltin and colleagues, referring to a phenomenon known to smokers as “the munchies.” But could Foltin’s team replicate this effect scientifically, or was it simply stoner lore? For thirteen days, Foltin and his colleagues confined six men to a laboratory setting where all food was provided and accurately measured. Each day, volunteers smoked either “two cigarettes containing active marijuana” or two placebo joints containing no marijuana. The primary psychoactive ingredient of marijuana, Δ9-tetrahydrocannabinol (THC), activates the cannabinoid receptor type 1 (CB1), which plays a key role in the brain circuitry that regulates food reward. If these circuits really do have a major impact on food intake and adiposity, then activating them using marijuana should have a clear effect.

The outcome of Foltin’s study was unequivocal: The men ate 40 percent more calories while they were stoned than while they were sober, and their body weights also climbed rapidly. Interestingly, they didn’t overeat at meals, but instead ate more highly palatable sweet snacks, such as candy bars, between meals. A number of other studies have confirmed that marijuana increases food intake, including my favorite one: “Effects of Marihuana on the Solution of Anagrams, Memory and Appetite.” I suppose even getting stoned, playing games, and pigging out is worthy of scientific study.

If THC activates the CB1 receptor and this increases food intake and adiposity, then it stands to reason that blocking the CB1 receptor should reduce food intake and cause weight loss. This is exactly the rationale for the CB1-blocking drug rimonabant, or “reverse marijuana” as I like to call it. As predicted, rimonabant reduces food intake and causes weight loss in a variety of animals, including humans.

Although the drug has demonstrated its effectiveness in a research setting and it was briefly approved as a weight-loss drug in Europe, it’s not currently approved for the treatment of any condition due to ongoing concerns about its negative side effects. Shockingly, “reverse marijuana” seems to increase the risk of depression, anxiety, and suicidal thoughts. Nonetheless, marijuana and rimonabant illustrate the powerful pull the reward system exerts on our behavior, including how much we choose to eat.

Yet, if food reward causes us to overeat and we’re all surrounded by it, then why do some people develop obesity while others don’t?

LEAD FOOT, WORN BRAKES

A young woman stares intently at a computer screen in Leonard Epstein’s lab at the University at Buffalo in New York. She’s playing a slot machine computer game. Each time she clicks the mouse button, three columns of shapes spin on the screen and then settle. If the shapes are different from one another, she gets nothing—but if they line up, she gets a point. While it may sound like she’s goofing off, in fact she’s participating in a series of fascinating studies that are beginning to shed light on why some people become obese and others don’t.

Once she earns two points, she gets a small piece of a candy bar. While she only needs two points to receive a candy bar the first time, the next time, she must earn four points to receive the same candy bar, and then eight. “We keep increasing the work requirements,” explains Epstein, “until eventually the person says, ‘Jeez, it’s just not worth it.’” The number of responses she’s willing to make before quitting quantifies how hard she’s willing to work for food. To compare this to how hard she’s willing to work for a nonfood reward, the researchers also give her simultaneous access to a second computer in the room. This computer is programmed with the exact same game, except it earns the woman access to an appealing magazine for a few minutes, rather than candy. The woman can switch between the two computers anytime she wants, and once she decides neither reward is worth the effort, the experiment is over.

This simple technique allows Epstein’s team to calculate a personal characteristic called the relative reinforcing value of food (RRVfood). RRVfood is a measure of how hard a person is willing to work for food, relative to a nonfood reward such as reading material—and people differ greatly in this regard. “There are huge individual differences in that some people will work really, really hard to get access to food, and other people will only work a little bit,” explains Epstein. The fact that RRVfood measures the motivational value of food relative to nonfoods is important, because we often have the choice to do something other than eat. RRVfood asks: When faced with a choice, are you more likely to eat or to do something else?

These studies have produced very provocative results: first, that sweet foods are exceptionally motivating, especially to youths. “If you let teenagers work for sweet soda,” relates Epstein, “they will work really hard … People will make thousands of responses for a small piece of candy.” A second provocative conclusion is that people who are overweight or obese tend to have a higher RRVfood than people who are lean. In particular, children who are overweight or obese are much more willing to work for highly rewarding foods like pizza or candy than lean children, even if their baseline level of hunger is the same. Consistent with their heightened food motivation, people with a high RRVfood eat more food both in the lab and at home. People who are overweight or obese find food more motivating than lean people do, and this leads them to eat more.

However, these studies don’t tell us whether a high RRVfood causes people to gain weight, or whether something about the overweight state causes RRVfood to increase. All they tell us is that the two are associated with one another. To begin to explore the question of whether a high RRVfood actually causes weight gain, Epstein and other researchers went back in time.

Knowing that a substantial fraction of lean children go on to become overweight, they examined whether RRVfood could predict who would go on to gain weight and who wouldn’t. Their results were remarkably consistent: RRVfood not only predicts weight gain in children, but in every age group they examined. In one study, adults with a high RRVfood gained more than five pounds over the course of a year, whereas adults with a low RRVfood only gained half a pound. “If you look at a group of lean people and you measure the reinforcing value of food,” explains Epstein, “you can predict who will gain weight.” These findings suggest that people differ in their motivation to eat food, particularly highly rewarding foods, and that this is a stable personality trait that influences each person’s susceptibility to weight gain over time. This offers a partial answer to the question posed earlier: Heightened food reward sensitivity does seem to contribute to overeating and fat gain over time.

Working hard for food makes a lot of sense when it’s the only way to survive. Throughout most of human history, and long before, our ancestors spent the bulk of their lives collecting, hunting, growing, and eating food—and it was often hard work. Without a powerful instinctive drive to obtain and eat food, we wouldn’t have survived at a time when securing food demanded so much effort. We still carry that instinct today, but in the modern world, where food is easy to get and highly rewarding, that powerful drive can often lead us to overeat. Yet, as with most traits, people differ widely in their level of food motivation.

But that’s not the end of the story. Drug abuse research suggests that a person’s susceptibility to addiction depends not only on how reinforcing the drug is for him but also on his ability to control his behavior in response to a craving—in other words, his impulsivity. Impulsivity describes a person’s ability—or lack thereof—to suppress or ignore basic urges that are beyond conscious control. It’s the opposite of what we commonly call self-control. A person who finds crack cocaine highly reinforcing will crave the drug intensely after he’s smoked it a few times, but if he’s able to prevent himself from acting on those cravings, he won’t be addicted. A second person who finds crack cocaine equally reinforcing but who is highly impulsive will readily become addicted. As Epstein puts it, “If you find something really rewarding and you have really poor impulse control, you’re in a lot of trouble.” Epstein coined the term reinforcement pathology to describe the dangerous combination of high reinforcement sensitivity and high impulsivity. He explains that it’s like having a “lead foot and worn brakes.” This may identify why some people are more susceptible to food addiction than others, despite the fact that we’re all exposed to potentially addictive foods.

On the other hand, people who have a high RRVfood but who aren’t impulsive (lead foot and good brakes) aren’t at an increased risk of overeating or weight gain. “If you have really good self-control,” explains Epstein, “you can overcome the reward value, and you can be a foodie: someone who loves food, who’s a gourmet cook, but who is lean because they can regulate the amount of food.” Does reinforcement pathology actually predict eating behavior and weight gain in the real world? While most people aren’t literally addicted to food, including those who are overweight and obese, the same principles of reinforcement and impulsivity should still apply to nonaddicted people. Even if you aren’t actually addicted to potato chips, you may still be drawn to eat them when you aren’t hungry, and your ability to suppress that urge when they’re available will influence how much of them you eat. Research from Epstein’s group and others supports this concept: People who exhibit reinforcement pathology are highly susceptible to overeating, and they’re also highly susceptible to weight gain.

Epstein is quick to point out that there’s a third important factor in addition to RRVfood and impulsivity: the presence of highly rewarding food in your personal environment. “Obviously, if you have something with very low reward value, you don’t need very good brakes. If you have a liver Popsicle, you don’t need self-control to not eat a liver Popsicle. If you have a grilled steak and you love meat, all of a sudden you need a lot of self-control to regulate that.” The deadliest combination, therefore, occurs when an impulsive person with a high food reward sensitivity lives in an environment that’s bursting at the seams with highly rewarding foods. And as we will soon see, the United States qualifies as such an environment.