Finally, the PFC mediates fear extinction. Yesterday the rat learned, “That tone is followed by a shock,” so the sound of the tone began to trigger freezing. Today there are no shocks, and the rat has acquired another truth that takes precedence—“but not today.” The first truth is still there; as proof, start coupling tone with shock again, and freezing to tone is “reinstated” faster than the association was initially learned.
Where is “but not today” consolidated? In the PFC, after receiving information from the hippocampus.67 The medial PFC activates inhibitory circuits in the BLA, and the rat stops freezing to the tone. In a similar vein but reflecting cognition specific to humans, condition people to associate a blue square on a screen with a shock, and the amygdala will activate when seeing that square—but less so in subjects who reappraise the situation, activating the medial PFC by thinking of, say, a beautiful blue sky.
This segues into the subject of regulating emotion through thought.68 It’s hard to regulate thought (try not thinking about a hippo) but even tougher with emotion; research by my Stanford colleague and close friend James Gross has explored this. First off, “thinking differently” about something emotional differs from simply suppressing the expression of the emotions. For example, show someone graphic footage of, say, an amputation. Subjects cringe, activate the amygdala and sympathetic nervous system. Now one group is instructed to hide their emotions (“I’m going to show you another film clip, and I want you to hide your emotional reactions”). How to do so most effectively? Gross distinguishes between “antecedent” and “response”-focused strategies. Response-focused is dragging the emotional horse back to the barn after it’s fled—you’re watching the next horrific footage, feeling queasy, and you think, “Okay, sit still, breathe slowly.” Typically this causes even greater activation of the amygdala and sympathetic nervous system.
Antecedent strategies generally work better, as they keep the barn door closed from the start. These are about thinking/feeling about something else (e.g., that great vacation), or thinking/feeling differently about what you’re seeing (reappraisals such as “That isn’t real; those are just actors”). And when done right, the PFC, particularly the dlPFC, activates, the amygdala and sympathetic nervous system are damped, and subjective distress decreases.*
Antecedent reappraisal is why placebos work.69 Thinking, “My finger is about to be pricked by a pin,” activates the amygdala along with a circuit of pain-responsive brain regions, and the pin hurts. Be told beforehand that the hand cream being slathered on your finger is a powerful analgesic cream, and you think, “My finger is about to be pricked by a pin, but this cream will block the pain.” The PFC activates, blunting activity in the amygdala and pain circuitry, as well as pain perception.
Thought processes like these, writ large, are the core of a particularly effective type of psychotherapy—cognitive behavioral therapy (CBT)—for the treatment of disorders of emotion regulation.70 Consider someone with a social anxiety disorder caused by a horrible early experience with trauma. To simplify, CBT is about providing the tools to reappraise circumstances that evoke the anxiety—remember that in this social situation those awful feelings you’re having are about what happened back then, not what is happening now.*
Controlling emotional responses with thought like this is very top down; the frontal cortex calms the overwrought amygdala. But the PFC/limbic relationship can be bottom up as well, when a decision involves a gut feeling. This is the backbone of Damasio’s somatic marker hypothesis. Choosing among options can involve a cerebral cost-benefit analysis. But it also involves “somatic markers,” internal simulations of what each outcome would feel like, run in the limbic system and reported to the vmPFC. The process is not a thought experiment; it’s an emotion experiment, in effect an emotional memory of a possible future.
A mild somatic marker activates only the limbic system.71 “Should I do behavior A? Maybe not—the possibility of outcome B feels scary.” A more vivid somatic marker activates the sympathetic nervous system as well. “Should I do behavior A? Definitely not—I can feel my skin getting clammy at the possibility of outcome B.” Experimentally boosting the strength of that sympathetic signal strengthens the aversion.
This is a picture of normal collaboration between the limbic system and frontal cortex.72 Naturally, things are not always balanced. Anger, for example, makes people less analytical and more reflexive in decisions about punishment. Stressed people often make hideously bad decisions, marinated in emotion; chapter 4 examines what stress does to the amygdala and frontal cortex.*
The effects of stress on the frontal cortex are dissected by the late Harvard psychologist Daniel Wegner in an aptly titled paper, “How to Think, Say or Do Precisely the Worst Thing on Any Occasion.”73 He considers what Edgar Allan Poe called the “imp of the perverse”:
We see a rut coming up in the road ahead and proceed to steer our bike right into it. We make a mental note not to mention a sore point in conversation and then cringe in horror as we blurt out exactly that thing. We carefully cradle the glass of red wine as we cross the room, all the while thinking “don’t spill,” and then juggle it onto the carpet under the gaze of our host.
Wegner demonstrated a two-step process of frontocortical regulation: (A) one stream identifies X as being very important; (B) the other stream tracks whether the conclusion is “Do X” or “Never do X.” And during stress, distraction, or heavy cognitive load, the two streams can dissociate; the A stream exerts its presence without the B stream saying which fork in the road to take. The chance that you will do precisely the wrong thing rises not despite your best efforts but because of a stress-boggled version of them.
—
This concludes our overview of the frontal cortex; the mantra is that it makes you do the harder thing when that is the right thing. Five final points:
“Doing the harder thing” effectively is not an argument for valuing either emotion or cognition more than the other. For example, as discussed in chapter 11, we are our most prosocial concerning in-group morality when our rapid, implicit emotions and intuitions dominate, but are most prosocial concerning out-group morality when cognition holds sway.
It’s easy to conclude that the PFC is about preventing imprudent behaviors (“Don’t do it; you’ll regret it”). But that isn’t always the case. For example, in chapter 17 we’ll consider the surprising amount of frontocortical effort it can take to pull a trigger.
Like everything about the brain, the structure and function of the frontal cortex vary enormously among individuals; for example, resting metabolic rate in the PFC varies approximately thirtyfold among people.* What causes such individual differences? See the rest of this book.74
“Doing the harder thing when it’s the right thing to do.” “Right” in this case is used in a neurobiological and instrumental sense, rather than a moral one.
Consider lying. Obviously, the frontal cortex aids the hard job of resisting the temptation. But it is also a major frontocortical task, particularly a dlPFC task, to lie competently, to control the emotional content of a signal, to generate an abstract distance between message and meaning. Interestingly, pathological liars have atypically large amounts of white matter in the PFC, indicating more complex wiring.75
But again, the “right thing,” in the setting of the frontal cortically assisted lying, is amoral. An actor lies to an audience about having the feelings of a morose Danish prince. A situationally ethical child lies, telling Grandma how excited she is about her present, concealing the fact that she already has that toy. A leader tells bold-faced lies, starting a war. A financier with Ponzi in his blood defrauds investors. A peasant woman lies to a uniformed thug, telling him she does not know the whereabouts of the refugees she knows are hiding in her attic. As with much about the frontal cortex, it’s context, context, context.
—
Where does the frontal cortex get the metaphorical
motivation to do the harder thing? For this we now look at our final branch, the dopaminergic “reward” system in the brain.
THE MESOLIMBIC/MESOCORTICAL DOPAMINE SYSTEM
Reward, pleasure, and happiness are complex, and the motivated pursuit of them occurs in at least a rudimentary form in many species. The neurotransmitter dopamine is central to understanding this.
Nuclei, Inputs, and Outputs
Dopamine is synthesized in multiple brain regions. One such region helps initiate movement; damage there produces Parkinson’s disease. Another regulates the release of a pituitary hormone. But the dopaminergic system that concerns us arises from an ancient, evolutionarily conserved region near the brain stem called the ventral tegmental area (henceforth the “tegmentum”).
A key target of these dopaminergic neurons is the last multisyllabic brain region to be introduced in this chapter, the nucleus accumbens (henceforth the “accumbens”). There’s debate as to whether the accumbens should count as part of the limbic system, but at the least it’s highly limbic-ish.
Here’s our first pass at the organization of this circuitry:76
The tegmentum sends projections to the accumbens and (other) limbic areas such as the amygdala and hippocampus. This is collectively called the “mesolimbic dopamine pathway.”
The tegmentum also projects to the PFC (but, significantly, not other cortical areas). This is called the “mesocortical dopamine pathway.” I’ll be lumping the mesolimbic plus mesocortical pathways together as the “dopaminergic system,” ignoring their not always being activated simultaneously.*
The accumbens projects to regions associated with movement.
Naturally, most areas getting projections from the tegmentum and/or accumbens project back to them. Most interesting will be the projections from the amygdala and PFC.
Reward
As a first pass, the dopaminergic system is about reward—various pleasurable stimuli activate tegmental neurons, triggering their release of dopamine.77 Some supporting evidence: (a) drugs like cocaine, heroin, and alcohol release dopamine in the accumbens; (b) if tegmental release of dopamine is blocked, previously rewarding stimuli become aversive; (c) chronic stress or pain depletes dopamine and decreases the sensitivity of dopamine neurons to stimulation, producing the defining symptom of depression—“anhedonia,” the inability to feel pleasure.
Some rewards, such as sex, release dopamine in every species examined.78 For humans, just thinking about sex suffices.*79 Food evokes dopamine release in hungry individuals of all species, with an added twist in humans. Show a picture of a milkshake to someone after they’ve consumed one, and there’s rarely dopaminergic activation—there’s satiation. But with subjects who have been dieting, there’s further activation. If you’re working to restrict your food intake, a milkshake just makes you want another one.
The mesolimbic dopamine system also responds to pleasurable aesthetics.80 In one study people listened to new music; the more accumbens activation, the more likely subjects were to buy the music afterward. And then there is dopaminergic activation for artificial cultural inventions—for example, when typical males look at pictures of sports cars.
Patterns of dopamine release are most interesting when concerning social interactions.81 Some findings are downright heartwarming. In one study a subject would play an economic game with someone, where a player is rewarded under two circumstances: (a) if both players cooperate, each receives a moderate reward, and (b) stabbing the other person in the back gets the subject a big reward, while the other person gets nothing. While both outcomes increased dopaminergic activity, the bigger increase occurred after cooperation.*
Other research examined the economic behavior of punishing jerks.82 In one study subjects played a game where player B could screw over player A for a profit. Depending on the round, player A could either (a) do nothing, (b) punish player B by having some of player B’s money taken (at no cost to player B), or (c) pay one unit of money to have two units taken from player B. Punishment activated the dopamine system, especially when subjects had to pay to punish; the greater the dopamine increase during no-cost punishment, the more willing someone was to pay to punish. Punishing norm violations is satisfying.
Another great study, carried out by Elizabeth Phelps of New York University, concerns “overbidding” in auctions, where people bid more money than anticipated.83 This is interpreted as reflecting the additional reward of besting someone in the competitive aspect of bidding. Thus, “winning” an auction is intrinsically socially competitive, unlike “winning” a lottery. Winning a lottery and winning a bid both activated dopaminergic signaling in subjects; losing a lottery had no effect, while losing a bidding war inhibited dopamine release. Not winning the lottery is bad luck; not winning an auction is social subordination.
This raises the specter of envy. In one neuroimaging study subjects read about a hypothetical person’s academic record, popularity, attractiveness, and wealth.84 Descriptions that evoked self-reported envy activated cortical regions involved in pain perception. Then the hypothetical individual was described as experiencing a misfortune (e.g., they were demoted). More activation of pain pathways at the news of the person’s good fortune predicted more dopaminergic activation after learning of their misfortune. Thus there’s dopaminergic activation during schadenfreude—gloating over an envied person’s fall from grace.
The dopamine system gives insights into jealousy, resentment, and invidiousness, leading to another depressing finding.85 A monkey has learned that when he presses a lever ten times, he gets a raisin as a reward. That’s just happened, and as a result, ten units of dopamine are released in the accumbens. Now—surprise!—the monkey presses the lever ten times and gets two raisins. Whoa: twenty units of dopamine are released. And as the monkey continues to get paychecks of two raisins, the size of the dopamine response returns to ten units. Now reward the monkey with only a single raisin, and dopamine levels decline.
Why? This is our world of habituation, where nothing is ever as good as that first time.
Unfortunately, things have to work this way because of our range of rewards.86 After all, reward coding must accommodate the rewarding properties of both solving a math problem and having an orgasm. Dopaminergic responses to reward, rather than being absolute, are relative to the reward value of alternative outcomes. In order to accommodate the pleasures of both mathematics and orgasms, the system must constantly rescale to accommodate the range of intensity offered by particular stimuli. The response to any reward must habituate with repetition, so that the system can respond over its full range to the next new thing.
This was shown in a beautiful study by Wolfram Schultz of Cambridge University.87 Depending on the circumstance, monkeys were trained to expect either two or twenty units of reward. If they unexpectedly got either four or forty units, respectively, there’d be an identical burst of dopamine release; giving one or ten units produced an identical decrease. It was the relative, not absolute, size of the surprise that mattered over a tenfold range of reward.
These studies show that the dopamine system is bidirectional.88 It responds with scale-free increases for unexpected good news and decreases for bad. Schultz demonstrated that following a reward, the dopamine system codes for discrepancy from expectation—get what you expected, and there’s a steady-state dribble of dopamine. Get more reward and/or get it sooner than expected, and there’s a big burst; less and/or later, a decrease. Some tegmental neurons respond to positive discrepancy from expectation, others to negative; appropriately, the latter are local neurons that release the inhibitory neurotransmitter GABA. Those same neurons participate in habituation, where the reward that once elicited a big dopamine response becomes less exciting.*
Logically, these different types of coding neurons in the tegmentum (as well as the accumbens) get projections from the frontal cortex—that’s where all the expectancy/discrepancy calculations take pl
ace—“Okay, I thought I was going to get 5.0 but got 4.9. How big of a bummer is that?”
Additional cortical regions weigh in. In one study subjects were shown an item to purchase, with the degree of accumbens activation predicting how much a person would pay.89 Then they were told the price; if it was less than what they were willing to spend, there was activation of the emotional vmPFC; more expensive, and there’d be activation of that disgust-related insular cortex. Combine all the neuroimaging data, and you could predict whether the person would buy the item.
Thus, in typical mammals the dopamine system codes in a scale-free manner over a wide range of experience for both good and bad surprises and is constantly habituating to yesterday’s news. But humans have something in addition, namely that we invent pleasures far more intense than anything offered by the natural world.
Once, during a concert of cathedral organ music, as I sat getting gooseflesh amid that tsunami of sound, I was struck with a thought: for a medieval peasant, this must have been the loudest human-made sound they ever experienced, awe-inspiring in now-unimaginable ways. No wonder they signed up for the religion being proffered. And now we are constantly pummeled with sounds that dwarf quaint organs. Once, hunter-gatherers might chance upon honey from a beehive and thus briefly satisfy a hardwired food craving. And now we have hundreds of carefully designed commercial foods that supply a burst of sensation unmatched by some lowly natural food. Once, we had lives that, amid considerable privation, also offered numerous subtle, hard-won pleasures. And now we have drugs that cause spasms of pleasure and dopamine release a thousandfold higher than anything stimulated in our old drug-free world.
Behave: The Biology of Humans at Our Best and Worst Page 7