فصل 2

2

THE SELECTION PROBLEM

In the basement of Sten Grillner’s laboratory at the Karolinska Institute in Stockholm, Sweden, dozens of foot-long, wormlike creatures adhere to the glass wall of a large fish tank using round, suction-cup-like mouths filled with needle-sharp teeth. These nightmarish creatures are lampreys, one of our truly ancient relatives. Lampreys and their cousins the hagfish are considered the most primitive living vertebrates, meaning animals that have evolved a backbone, spinal cord, and brain. The ancestors of today’s lampreys diverged from our own ancestors approximately 560 million years ago—before the evolution of mammals, dinosaurs, reptiles, amphibians, and even most types of fishes—and long before our ancestors laid a fin on land.

Because lampreys are our most distant vertebrate relatives, comparing their brains to the brains of mammals reveals the common elements of all vertebrate brains: the core processing circuits that form the foundation of the human mind. Grillner’s research has shown that within the pea-sized brain of these primitive creatures lie the seeds of the human decision-making apparatus. If we are to fathom our own eating behavior, we have to understand the fundamentals of how the brain makes decisions, and the lamprey is an excellent place to start.

THE SELECTION PROBLEM: HOW DECISIONS ARE MADE IN A COMPLEX WORLD

Imagine two robots on a car assembly line. As each car door rolls by, Robot 1 paints it green. Door after door, Robot 1 performs the exact same action, and that single action is all it can do. This kind of robot doesn’t require much processing power because it only has one job, one capability, and therefore no decisions to make. Now, imagine that Robot 2 can do two different things: It can paint a door green, or it can paint it red. Robot 2 only has one paint nozzle, so it can’t paint a door both colors at once. Rather, it has to choose which color to use. So how does Robot 2 decide? This fundamental challenge is called the selection problem, and it occurs any time there are multiple options (green versus red paint) competing for the same shared resource (one paint nozzle). To solve the selection problem, Robot 2 needs a selector—some sort of function that decides which color is most appropriate to apply to each door.

Our very earliest ancestors were probably like Robot 1—simple creatures that didn’t have to make any decisions about what to do. But this didn’t last long. As soon as they evolved the ability to do more than one thing with the same set of resources, they had to begin making decisions about what to do—and those that made the best decisions passed on their genes to the next generation. Lampreys, for example, can do a number of different things: They can suction onto a rock, track prey, flee predators, mate, build a nest, feed, and swim in a nearly infinite number of directions. Many of these options are mutually exclusive because they require the same muscles inside the same body. So like Robot 2, the lamprey has a selection problem, and it needs a selector to resolve it.

According to researchers in computational neuroscience and artificial intelligence, an effective selector must have certain key properties, whether it exists in a computer or in a brain:

1- The selector must be able to choose one option. If there are incompatible options, such as fleeing a predator or mating, the selector must be able to pick only one and allow it access to the resources necessary to execute its program.

2- The selector must be able to choose the best option in any given situation. For example, if a lamprey sees a dangerous predator, it should flee. A lamprey that tries to mate when it sees a dangerous predator won’t survive and won’t pass on its genes to the next generation of lampreys.

3- The selector must be able to select decisively between options. If one option is only slightly better than the others, it still must win definitively, shutting off all incompatible options completely. A lamprey that tries to mate and flee at the same time is probably not going to leave many offspring.

In a seminal research paper published in 1999, University of Sheffield researchers brought together evidence from neuroscience and computer modeling to argue that selection is precisely the function of an ancient group of structures deep within the human brain called the basal ganglia. Today, this idea is accepted by most neuroscientists. To understand how the human selector works, let’s start with a simpler version of it: the lamprey selector.

THE LAMPREY SOLUTION TO THE SELECTION PROBLEM

How does the lamprey decide what to do? Within the lamprey basal ganglia lies a key structure called the striatum, which is the portion of the basal ganglia that receives most of the incoming signals from other parts of the brain. The striatum receives “bids” from other brain regions, each of which represents a specific action. A little piece of the lamprey’s brain is whispering, “Mate,” to the striatum, while another piece is shouting, “Flee the predator!” and so on. It would be a very bad idea for these movements to occur simultaneously—because a lamprey can’t do them all at the same time—so to prevent simultaneous activation of many different movements, all these regions are held in check by powerful inhibitory connections from the basal ganglia. This means that the basal ganglia keep all behaviors in “off” mode by default. Only once a specific action’s bid has been selected do the basal ganglia turn off this inhibitory control, allowing the behavior to occur. You can think of the basal ganglia as a bouncer that chooses which behavior gets access to the muscles and turns away the rest. This fulfills the first key property of a selector: It must be able to pick one option and allow it access to the muscles.

Many of these action bids originate from a region of the lamprey brain called the pallium, which is thought to be involved in planning behavior. Each little region of the pallium is responsible for a particular behavior, such as tracking prey, suctioning onto a rock, or fleeing predators.

These regions are thought to have two basic functions. The first is to execute the behavior in which it specializes, once it has received permission from the basal ganglia. For example, the “track prey” region activates downstream pathways that contract the lamprey’s muscles in a pattern that causes the animal to track its prey.

The second basic function of these regions is to collect relevant information about the lamprey’s surroundings and internal state, which determines how strong of a bid it will put in to the striatum. For example, if there’s a predator nearby, the “flee predator” region will put in a very strong bid to the striatum, while the “build a nest” bid will be weak. If the lamprey is hungry and it sees prey, the “track prey” bid will be strong, but the “suction onto a rock” bid will be weak.

Each little region of the pallium is attempting to execute its specific behavior and competing against all other regions that are incompatible with it. The strength of each bid represents how valuable that specific behavior appears to the organism at that particular moment, and the striatum’s job is simple: Select the strongest bid. This fulfills the second key property of a selector—that it must be able to choose the best option for a given situation.

At the same time that the striatum selects the strongest bid, it shuts down competing bids. So once the “flee predator” bid has won the competition, bids like “suction onto a rock” and “track prey” get bounced. This fulfills the third key property of a selector—that it must be able to select decisively, picking one option and shutting down all competitors.

Each region of the pallium sends a connection to a particular region of the striatum, which (via other parts of the basal ganglia) returns a connection back to the same starting location in the pallium. This means that each region of the pallium is reciprocally connected with the striatum via a specific loop that regulates a particular action. For example, there’s a loop for tracking prey, a loop for fleeing predators, a loop for anchoring to a rock, and so on. Each region of the pallium is constantly whispering to the striatum to let it trigger its behavior, and the striatum always says “no!” by default. In the appropriate situation, the region’s whisper becomes a shout, and the striatum allows it to use the muscles to execute its action. This is how the lamprey is able to react appropriately to its surroundings and internal state.

With all this in mind, it’s helpful to think of each individual region of the lamprey pallium as an option generator that’s responsible for a specific behavior. Each option generator is constantly competing with all other incompatible option generators for access to the muscles, and the option generator with the strongest bid at any particular moment wins the competition. The basal ganglia evaluate the bids, decide which one is the strongest, give the winning option generator access to the muscles, and shut down all competing option generators. The lamprey flees, avoids the predator, and lives to pass on its genes to the next generation of lampreys.

THE MAMMALIAN SOLUTION TO THE SELECTION PROBLEM

Most people would agree that the human brain is a bit more sophisticated than the brain of a lamprey. All right, a lot more sophisticated. One of the things that sets mammals apart from most other creatures on earth is the tremendous complexity of our nervous systems, which allows us to make remarkably smart decisions. To understand just how useful our high-powered model is, consider how much energy it guzzles. In humans, the brain eats up one-fifth of our total energy usage, even though it accounts for only 2 percent of our body weight. The fact that evolution allowed us to bear this energy-hogging ball and chain is a testament to its importance. Making smart decisions is a good evolutionary strategy, and no animal does it better than humans.

So what does the lamprey brain have to do with the human brain? This is what Karolinska Institute researcher Sten Grillner and his former graduate student Marcus Stephenson-Jones set out to answer. Building on the work of previous researchers, they compared the anatomy and function of the basal ganglia in lampreys and mammals. Their findings are nothing short of remarkable: Despite being separated by a 560-million-year evolutionary chasm, the basal ganglia of lampreys and mammals (including humans) are strikingly similar. They contain the same regions, organized and connected in the same way. Within these regions lie neurons with the same electrical properties, communicating with one another using the same chemical messengers. These findings led Grillner and Stephenson-Jones to the stunning conclusion that “practically all details of the basal ganglia circuitry had developed some 560 million years ago.” Stephenson-Jones adds: “This is really a fundamental part of the vertebrate brain that’s been used across evolution as a common mechanism for how lampreys, fish, birds, mammals, and even humans make decisions.” Our ancestors hit an evolutionary home run 560 million years ago, and we still carry the hardware they developed in the ancient seas.

A lamprey is capable of making many different types of decisions, but certainly far fewer than humans are. We have to make up our minds about things a lamprey cannot fathom, like what to cook for dinner, how to pay off the mortgage, and whether or not to believe in God. Clearly, there are important differences in the brain hardware that allows us to understand the world and make choices. But if the decision-making capacity of a human and a lamprey are so different, why are the basal ganglia of lampreys and humans so strikingly similar? Grillner and Stephenson-Jones propose an explanation: an evolutionary process called exaptation. As opposed to adaptation, which is the process of developing new traits—such as air-breathing lungs or a four-chambered heart—exaptation takes something that already exists and finds a new function for it; for example, expanding the basal ganglia’s decision-making jurisdiction to govern other, more advanced types of decisions. Grillner and Stephenson-Jones propose that the early vertebrate basal ganglia were already very good at making decisions, and there was no need for evolution to fix what wasn’t broken. We only needed to build on it.

In humans, the most numerous inputs to the striatum come from the cerebral cortex, which evolved from the pallium (similar to the one found in today’s lampreys). The cortex is critical for advanced decision-making. You can still do a lot of basic things without it, things that are regulated by deeper, older parts of the brain, but you can’t decide about the mortgage or about God. The cortex is comically enlarged in humans relative to other animals, and it plays a key role in our exceptional intelligence. The lamprey pallium is rudimentary by comparison. This is part of the reason why lampreys don’t have mortgages.

These major inputs from the cortex to the striatum suggest that the role of the basal ganglia has expanded considerably since our divergence from the ancestors of lampreys. As it turns out, the cortex doesn’t just send inputs to the basal ganglia—it also receives input back, just like the lamprey pallium. These reciprocal connections also form loops that travel to and from specific regions of the cortex, each of which is an option generator. In fact, there are similar loops connecting the basal ganglia to many parts of the mammalian brain—parts that regulate not only movements but motivations and emotions, thoughts and associations, and numerous other processes.

Over the course of evolutionary history, the process of exaptation multiplied the basic decision-making units of the basal ganglia and connected them to fancy new option generators capable of proposing much more sophisticated options and computing value in more advanced ways. In addition to deciding how to move, the human basal ganglia can decide how to feel, what to think, what to say, and—to return to the matter at hand—what to eat.

THE BASAL GANGLIA GO TO A RESTAURANT

Broken down to its fundamental elements, behavior begins to appear quite complex because it involves the coordination of many interacting parts. To achieve a seemingly simple goal such as eating at a restaurant, you must first become motivated to eat, then you must figure out where you want to eat and how to get there, and then you must control your musculature in just the right way to get there and place the food into your mouth. This is vastly more challenging than the task that Robot 2 faced, because each of these processes involves a decision. These motivational, cognitive, and motor tasks are all processed in different parts of the brain, yet they work together so seamlessly that we’re scarcely aware they’re distinct from one another. How does the brain make all these decisions in such a coordinated manner?

It’s impossible to be certain, because we can’t do detailed invasive studies on the human brain as we do in other species, but researchers have a compelling hypothesis that draws from a variety of scientific clues. To understand this hypothesis, I spoke with University of Sheffield researchers Peter Redgrave and Kevin Gurney, who have played key roles in uncovering the function of the basal ganglia in decision-making. Here’s what they explained to me: Let’s say you haven’t eaten in a while. From a survival standpoint, your body wants energy, so eating would be a valuable action. How does it get you to do it? The first step is to set your motivation for food. The ventral (bottom) part of the striatum is responsible for selecting between competing motivations and emotions. “Those are the motivational channels that select high-level goals,” explains Redgrave. “Are you hungry, thirsty, frightened, sexy, cold, or hot?” The hungry, thirsty, frightened, sexy, cold, and hot option generators compete for expression by sending bids to the ventral striatum. At the moment, because you’re low on energy, the hunger option generator is putting in a very strong bid (a potential pitfall we’ll explore later on). It wins the competition and is allowed to express itself. You begin to feel hungry.

Once the hunger option generator has won and you feel motivated to eat, it begins to activate other option generators in the cortex responsible for figuring out how to get food, or generally, making a plan. Option generators representing the refrigerator, pizza delivery, the restaurant down the street, and the really good restaurant across town enter into competition in the dorsal (upper) striatum. Just as the strength of the hunger bid was determined by information about your body’s energy status, the strength of each eating plan bid is determined by relevant information about that option: how good the food was last time you had it, what other people have said about it, how much effort it requires, and how expensive it is. As it turns out, the restaurant across town is really good, but you don’t feel like driving. The food in your fridge is cheapest, but it requires cooking. The restaurant up the street is close and cheap, and so it puts in the strongest bid and wins the competition.

Now you have a plan, but how do you execute it? Do you walk, ride your bike, drive, or take the bus? The “restaurant up the street” option generator initiates another competition between walk, bike, and bus option generators in the cortex, which again provide competing inputs to the dorsal striatum. You feel like getting some fresh air, but you also want to get there quickly, so the bike option wins. Once you get onto your bike, how do you make it move forward? Do you wave your hands, wiggle your toes, shake your head, or pedal? The answer is obvious, but it still requires a decision between competing options in motor regions of the brain. Pedaling puts in the strongest bid, and you jump on your bike and head for the restaurant. This process is illustrated schematically in figure 11, and the underlying brain circuits are illustrated in figure 12.

I’ll spare you the description of all the decisions you have to make while riding to the restaurant, choosing from the menu, and eating your food. The point is that many of our behaviors are thought to result from a cascading series of competitions within motivational, cognitive, and motor brain regions. The winning motivation initiates subsequent competitions in cognitive areas that are relevant to fulfilling that motivation, and then the cognitive areas initiate competitions in motor areas that are relevant to physically executing the plan of action. The strength of each bid is determined by experience, internal cues, and external cues, and the basal ganglia only allow the strongest bids to express themselves. This process occurs beyond our conscious awareness—we only become aware of bids after they’re selected. This is consistent with Daniel Kahneman’s idea (discussed in the introduction) that most of what happens in the brain, including many decision-making processes, is nonconscious.

Many behaviors we view as trivial, such as pumping gasoline or washing the dishes, are actually tremendously complex. Artificial intelligence researchers are keenly aware of the difficulties of reproducing even elementary goal-directed behaviors, which explains why today’s computers are good at computing but not so good at making complex decisions without human guidance. This is a testament to how much of the functioning of our own brains we take for granted.

THE MAN WHO HAD NO THOUGHTS

To illustrate the crucial importance of the basal ganglia in decision-making processes, let’s consider what happens when they don’t work.

As it turns out, several disorders affect the basal ganglia. The most common is Parkinson’s disease, which results from the progressive loss of cells in a part of the basal ganglia called the substantia nigra. These cells send connections to the dorsal striatum, where they produce dopamine, a chemical messenger that plays a very important role in the function of the striatum. Dopamine is a fascinating and widely misunderstood molecule that we’ll discuss further in the next chapter, but for now, its most relevant function is to increase the likelihood of engaging in any behavior.

When dopamine levels in the striatum are increased—for example, by cocaine or amphetamine—mice (and humans) tend to move around a lot. High levels of dopamine essentially make the basal ganglia more sensitive to incoming bids, lowering the threshold for activating movements. Figure 13 illustrates the striking effect of dopamine-boosting cocaine on mouse locomotion (walking and running).

Conversely, when dopamine levels are low, the basal ganglia become less sensitive to incoming bids and the threshold for activating movements is high. In this scenario, animals tend to stay put. The most extreme example of this is the dopamine-deficient mice created by Richard Palmiter, a neuroscience researcher at the University of Washington. These animals sit in their cages nearly motionless all day due to a complete absence of dopamine. “If you set a dopamine-deficient mouse on a table,” explains Palmiter, “it will just sit there and look at you. It’s totally apathetic.” When Palmiter’s team chemically replaces the mice’s dopamine, they eat, drink, and run around like mad until the dopamine is gone.

In Parkinson’s disease, the gradual loss of substantia nigra neurons causes dopamine levels to decline in the areas of the dorsal striatum that select movements, particularly well-worn movement patterns. This makes the dorsal striatum progressively less sensitive to incoming bids from motor regions, rendering it increasingly difficult for any motor option generator to gain access to the body’s muscles. People with Parkinson’s disease develop difficulty initiating and executing movements and have a particularly hard time performing sequences of movements. In severe cases, Parkinson’s disease patients are scarcely able to initiate movements at all, a phenomenon called akinesia, Greek for “no movement.” Fortunately, modern medicine has developed drugs that help relieve the debilitating motor impairments of Parkinson’s disease. Most of these are designed to increase dopamine signaling in the brain. The most effective and commonly used drug is the dopamine precursor L-dopa. When taken orally, L-dopa enters the circulation, and some of it crosses into the brain. Once inside the brain, it’s taken up by dopamine-producing neurons and converted into dopamine. Although there’s currently no way to regenerate lost cells in the substantia nigra, L-dopa makes the remaining cells, and perhaps even other cell types that don’t normally contain dopamine, produce more dopamine to compensate for the deficit. Higher levels of dopamine make the dorsal striatum more sensitive to incoming bids from motor regions and enables patients to move more normally once again.

As with many drugs, L-dopa is a blunt tool. In Parkinson’s disease, parts of the dorsal striatum need more dopamine—but the rest of the brain doesn’t. When a person takes L-dopa, dopamine-producing neurons throughout the brain—including those located in the area that provides dopamine to the ventral striatum, the ventral tegmental area (VTA)—sponge it up and convert it to dopamine. This can lead to abnormally elevated levels of dopamine in the ventral striatum.

As I mentioned earlier, the ventral striatum primarily regulates motivations and emotional states. Analogous to what happens in the dorsal striatum, elevating dopamine in the ventral striatum makes it more sensitive to incoming bids, increasing the likelihood that it will activate motivational and emotional states. In fact, common side effects of L-dopa treatment include heightened emotional states, hypersexuality, and compulsive and addictive behaviors, such as gambling, shopping, drug abuse, and binge eating. These are called impulse control disorders because people lose the ability to keep their basic impulses in check. The ventral striatum is so sensitive to incoming bids that inappropriate option generators are able to grab the reins. In addition, higher levels of dopamine in the striatum may cause the activity of certain loops to become abnormally strong over time, resulting in addictive and compulsive behaviors, a topic we’ll discuss in the next chapter.

Other basal ganglia disorders are even more interesting. Consider Jim, a former miner who was admitted to a psychiatric hospital at the age of fifty-seven with a cluster of unusual symptoms. As recorded in his case report: During the preceding three years he had become increasingly withdrawn and unspontaneous. In the month before admission he had deteriorated to the point where he was doubly incontinent, answered only yes or no to questions, and would sit or stand unmoving if not prompted. He only ate with prompting, and would sometimes continue putting spoon to mouth, sometimes for as long as two minutes after his plate was empty. Similarly he would flush the toilet repeatedly until asked to stop.

Jim was suffering from a rare disorder called abulia, which is Greek for “an absence of will.” Patients who suffer from abulia can respond to questions and perform specific tasks if prompted, but they have difficulty spontaneously initiating motivations, emotions, and thoughts. A severely abulic patient seated in a bare room by himself will remain immobile until someone enters the room. If asked what he was thinking or feeling, he’ll reply, “Nothing.” Needless to say, abulic patients have little motivation to eat.

Abulia is typically associated with damage to the basal ganglia and related circuits, and it often responds well to drugs that increase dopamine signaling. One of these is bromocriptine, the drug used to treat Jim: He was started on bromocriptine 5 mg per day, and increased in 5 mg stages to a maximum of 55 mg in divided doses. His first spontaneous activity, dressing without prompting, occurred at 20 mg. At 30 mg he started to initiate conversations with other residents, but there was considerable day to day variation. As the dose increased, he washed, dressed and ate his meals without prompting and was not perseverative. On some days though, he reverted to his pre-treatment state. On the maximum dosage such days were rare and he was generally independent in his activities of daily living … Researchers believe that the brain damage associated with abulia causes the basal ganglia to become insensitive to incoming bids, such that even the most appropriate feelings, thoughts, and motivations aren’t able to be expressed (or even to enter consciousness). Drugs that increase dopamine signaling make the striatum more sensitive to bids, allowing some abulic patients to recover the ability to feel, think, and move spontaneously.

WHAT DOES THIS HAVE TO DO WITH OVEREATING?

Now that we know something about how the brain makes decisions, we can get more specific about how it decides what, and how much, to eat. Eating is a complex behavior that requires coordinated decision-making on motivational, cognitive, and motor levels. The fundamental spark that sets the whole behavioral cascade in motion, however, is motivation. This motivation to eat can come from several different brain regions, in response to different cues. For example, the option generator that causes hunger is presumably different from the one that causes you to eat dessert after a big meal, and is also presumably different from the one that caused Joey Chestnut to eat sixty-nine hot dogs in ten minutes to win Nathan’s Hot Dog Eating Contest. Yet without motivation, there is no eating.

In the next chapter, we’ll delve into the circuits that set our motivation for eating food—particularly those that drive us to overeat. Which brain circuits motivate us to overeat, what cues drive that motivation, and what can we do about it? We’ll start off by sticking with the basal ganglia, exploring how these structures cause us to learn about food, crave it, and perhaps even become addicted to it.