SOME INPUTS TO THE AMYGDALA
Sensory inputs. For starters, the amygdala, specifically the BLA, gets projections from all the sensory systems.26 How else can you get terrified by the shark’s theme music in Jaws? Normally, sensory information from various modalities (eyes, ears, skin . . .) courses into the brain, reaching the appropriate cortical region (visual cortex, auditory cortex, tactile cortex . . .) for processing. For example, the visual cortex would engage layers and layers of neurons to turn pixels of retinal stimulation into recognizable images before it can scream to the amygdala, “It’s a gun!” Importantly, some sensory information entering the brain takes a shortcut, bypassing the cortex and going directly to the amygdala. Thus the amygdala can be informed about something scary before the cortex has a clue. Moreover, thanks to the extreme excitability of this pathway, the amygdala can respond to stimuli that are too fleeting or faint for the cortex to note. Additionally, the shortcut projections form stronger, more excitable synapses in the BLA than do the ones from the sensory cortex; emotional arousal enhances fear conditioning through this pathway. This shortcut’s power is shown in the case of a man with stroke damage to his visual cortex, producing “cortical blindness.” While unable to process most visual information, he still recognized emotional facial expressions via the shortcut.*
Crucially, while sensory information reaches the amygdala rapidly by this shortcut, it isn’t terribly accurate (since, after all, accuracy is what the cortex supplies). As we’ll see in the next chapter, this produces tragic circumstances where, say, the amygdala decides it’s seeing a handgun before the visual cortex can report that it’s actually a cell phone.
Information about pain. The amygdala receives news of that reliable trigger of fear and aggression, namely pain.27 This is mediated by projections from an ancient, core brain structure, the “periaqueductal gray” (PAG); stimulation of the PAG can evoke panic attacks, and it is enlarged in people with chronic panic attacks. Reflecting the amygdala’s roles in vigilance, uncertainty, anxiety, and fear, it’s unpredictable pain, rather than pain itself, that activates the amygdala. Pain (and the amygdala’s response to it) is all about context.
Disgust of all stripes. The amygdala also receives a hugely interesting projection from the “insular cortex,” an honorary part of the prefrontal cortex, which we will consider at length in later chapters.28 If you (or any other mammal) bite into rancid food, the insular cortex lights up, causing you to spit it out, gag, feel nauseated, make a revolted facial expression—the insular cortex processes gustatory disgust. Ditto for disgusting smells.
Remarkably, humans also activate it by thinking about something morally disgusting—social norm violations or individuals who are typically stigmatized in society. And in that circumstance its activation drives that of the amygdala. Someone does something lousy and selfish to you in a game, and the extent of insular and amygdaloid activation predicts how much outrage you feel and how much revenge you take. This is all about sociality—the insula and amygdala don’t activate if it’s a computer that has stabbed you in the back.
The insula activates when we eat a cockroach or imagine doing so. And the insula and amygdala activate when we think of the neighboring tribe as loathsome cockroaches. As we’ll see, this is central to how our brains process “us and them.”
And finally, the amygdala gets tons of inputs from the frontal cortex. Much more to come.
SOME OUTPUTS FROM THE AMYGDALA
Bidirectional connections. As we’ll see, the amygdala talks to many of the regions that talk to it, including the frontal cortex, insula, periaqueductal gray, and sensory projections, modulating their sensitivity.
The amygdala/hippocampus interface. Naturally, the amygdala talks to other limbic structures, including the hippocampus. As reviewed, typically the amygdala learns fear and the hippocampus learns detached, dispassionate facts. But at times of extreme fear, the amygdala pulls the hippocampus into a type of fear learning.29
Back to the rat undergoing fear conditioning. When it’s in cage A, a tone is followed by a shock. But in cage B, the tone isn’t. This produces context-dependent conditioning—the tone causes fearful freezing in cage A but not in cage B. The amygdala learns the stimulus cue—the tone—while the hippocampus learns about the contexts of cage A versus B. The coupled learning between amygdala and hippocampus is very focalized—we all remember the view of the plane hitting the second World Trade Center tower, but not whether there were clouds in the background. The hippocampus decides whether a factoid is worth filing away, depending on whether the amygdala has gotten worked up over it. Moreover, the coupling can rescale. Suppose someone robs you at gunpoint in an alley in a bad part of town. Afterward, depending on the circumstance, the gun can be the cue and the alley the context, or the alley is the cue and the bad part of town the context.
Motor outputs. There’s a second shortcut regarding the amygdala, specifically when it’s talking to motor neurons that command movement.30 Logically, when the amygdala wants to mobilize a behavior—say, fleeing—it talks to the frontal cortex, seeking its executive approval. But if sufficiently aroused, the amygdala talks directly to subcortical, reflexive motor pathways. Again, there’s a trade-off—increased speed by bypassing the cortex, but decreased accuracy. Thus the input shortcut may prompt you to see the cell phone as a gun. And the output shortcut may prompt you to pull a trigger before you consciously mean to.
Arousal. Ultimately, amygdala outputs are mostly about setting off alarms throughout the brain and body. As we saw, the core of the amygdala is the central amygdala.31 Axonal projections from there go to an amygdala-ish structure nearby called the bed nucleus of the stria terminalis (BNST). The BNST, in turn, projects to parts of the hypothalamus that initiate the hormonal stress response (see chapter 4), as well as to midbrain and brain-stem sites that activate the sympathetic nervous system and inhibit the parasympathetic nervous system. Something emotionally arousing occurs, layer 2 limbic amygdala signals layer 1 regions, and heart rate and blood pressure soar.*
The amygdala also activates a brain-stem structure called the locus coeruleus, akin to the brain’s own sympathetic nervous system.32 It sends norepinephrine-releasing projections throughout the brain, particularly the cortex. If the locus coeruleus is drowsy and silent, so are you. If it’s moderately activated, you’re alert. And if it’s firing like gangbusters, thanks to inputs from an aroused amygdala, all neuronal hands are on deck.
The amygdala’s projection pattern raises an important point.33 When is the sympathetic nervous system going full blast? During fear, flight, fight, and sex. Or if you’ve won the lottery, are happily sprinting down a soccer field, or have just solved Fermat’s theorem (if you’re that kind of person). Reflecting this, about a quarter of neurons in one hypothalamic nucleus are involved in both sexual behavior and, when stimulated at a higher intensity, aggressive behavior in male mice.
This has two implications. Both sex and aggression activate the sympathetic nervous system, which in turn can influence behavior—people feel differently about things if, say, their heart is racing versus beating slowly. Does this mean that the pattern of your autonomic arousal influences what you feel? Not really. But autonomic feedback influences the intensity of what is felt. More on this in the next chapter.
The second consequence reflects a core idea of this book. Your heart does roughly the same thing whether you are in a murderous rage or having an orgasm. Again, the opposite of love is not hate, it’s indifference.
—
This concludes our overview of the amygdala. Amid the jargon and complexity, the most important theme is the amygdala’s dual role in both aggression and facets of fear and anxiety. Fear and aggression are not inevitably intertwined—not all fear causes aggression, and not all aggression is rooted in fear. Fear typically increases aggression only in those already prone to it; among the subordinate who lack the option of expressing aggression safely, fear does the oppos
ite.
The dissociation between fear and aggression is evident in violent psychopaths, who are the antithesis of fearful—both physiologically and subjectively they are less reactive to pain; their amygdalae are relatively unresponsive to typical fear-evoking stimuli and are smaller than normal.34 This fits with the picture of psychopathic violence; it is not done in aroused reaction to provocation. Instead, it is purely instrumental, using others as a means to an end with emotionless, remorseless, reptilian indifference.
Thus, fear and violence are not always connected at the hip. But a connection is likely when the aggression evoked is reactive, frenzied, and flecked with spittle. In a world in which no amygdaloid neuron need be afraid and instead can sit under its vine and fig tree, the world is very likely to be a more peaceful place.*
—
We now move to the second of the three brain regions we’re considering in detail.
THE FRONTAL CORTEX
I’ve spent decades studying the hippocampus. It’s been good to me; I’d like to think I’ve been the same in return. Yet I think I might have made the wrong choice back then—maybe I should have studied the frontal cortex all these years. Because it’s the most interesting part of the brain.
What does the frontal cortex do? Its list of expertise includes working memory, executive function (organizing knowledge strategically, and then initiating an action based on an executive decision), gratification postponement, long-term planning, regulation of emotions, and reining in impulsivity.35
This is a sprawling portfolio. I will group these varied functions under a single definition, pertinent to every page of this book: the frontal cortex makes you do the harder thing when it’s the right thing to do.
To start, here are some important features of the frontal cortex:
It’s the most recently evolved brain region, not approaching full splendor until the emergence of primates; a disproportionate percentage of genes unique to primates are active in the frontal cortex. Moreover, such gene expression patterns are highly individuated, with greater interindividual variability than average levels of whole-brain differences between humans and chimps.
The human frontal cortex is more complexly wired than in other apes and, by some definitions as to its boundaries, proportionately bigger as well.36
The frontal cortex is the last brain region to fully mature, with the most evolutionarily recent subparts the very last. Amazingly, it’s not fully online until people are in their midtwenties. You’d better bet this factoid will be relevant to the chapter about adolescence.
Finally, the frontal cortex has a unique cell type. In general, the human brain isn’t unique because we’ve evolved unique types of neurons, neurotransmitters, enzymes, and so on. Human and fly neurons are remarkably similar; the uniqueness is quantitative—for every fly neuron, we have a gazillion more neurons and a bazillion more connections.37
The sole exception is an obscure type of neuron with a distinctive shape and pattern of wiring, called von Economo neurons (aka spindle neurons). At first they seemed to be unique to humans, but we’ve now found them in other primates, whales, dolphins, and elephants.* That’s an all-star team of socially complex species.
Moreover, the few von Economo neurons occur only in two subregions of the frontal cortex, as shown by John Allman at Caltech. One we’ve heard about already—the insula, with its role in gustatory and moral disgust. The second is an equally interesting area called the anterior cingulate. To give a hint (with more to come), it’s central to empathy.
So from the standpoint of evolution, size, complexity, development, genetics, and neuron type, the frontal cortex is distinctive, with the human version the most unique.
The Subregions of the Frontal Cortex
Frontal cortical anatomy is hellishly complicated, and there are debates as to whether some parts of the primate frontal cortex even exist in “simpler” species. Nonetheless, there are some useful broad themes.
In the very front is the prefrontal cortex (PFC), the newest part of the frontal cortex. As noted, the frontal cortex is central to executive function. To quote George W. Bush, within the frontal cortex, it’s the PFC that is “the decider.” Most broadly, the PFC chooses between conflicting options—Coke or Pepsi; blurting out what you really think or restraining yourself; pulling the trigger or not. And often the conflict being resolved is between a decision heavily driven by cognition and one driven by emotions.
Once it has decided, the PFC sends orders via projections to the rest of the frontal cortex, sitting just behind it. Those neurons then talk to the “premotor cortex,” sitting just behind it, which then passes it to the “motor cortex,” which talks to your muscles. And a behavior ensues.*
Before considering how the frontal cortex influences social behavior, let’s start with a simpler domain of its function.
The Frontal Cortex and Cognition
What does “doing the harder thing when it’s the right thing to do” look like in the realm of cognition (defined by Princeton’s Jonathan Cohen as “the ability to orchestrate thought and action in accordance with internal goals”)?38 Suppose you’ve looked up a phone number in a city where you once lived. The frontal cortex not only remembers it long enough to dial but also considers it strategically. Just before dialing, you consciously recall that it is in that other city and retrieve your memory of the city’s area code. And then you remember to dial “1” before the area code.*
The frontal cortex is also concerned with focusing on a task. If you step off the curb planning to jaywalk, you look at traffic, paying attention to motion, calculating whether you can cross safely. If you step off looking for a taxi, you pay attention to whether a car has one of those lit taxicab thingies on top. In a great study, monkeys were trained to look at a screen of dots of various colors moving in particular directions; depending on a signal, a monkey had to pay attention to either color or movement. Each signal indicating a shift in tasks triggered a burst of PFC activity and, coupled with that, suppression of the stream of information (color or movement) that was now irrelevant. This is the PFC getting you to do the harder thing; remembering that the rule has changed, don’t do the previous habitual response.39
The frontal cortex also mediates “executive function”—considering bits of information, looking for patterns, and then choosing a strategic action.40 Consider this truly frontally demanding test. The experimenter tells a masochistic volunteer, “I’m going to the market and I’m going to buy peaches, cornflakes, laundry detergent, cinnamon . . .” Sixteen items recited, the volunteer is asked to repeat the list. Maybe they correctly recall the first few, the last few, list some near misses—say, nutmeg instead of cinnamon. Then the experimenter repeats the same list. This time the volunteer remembers a few more, avoids repeating the nutmeg incident. Now do it again and again.
This is more than a simple memory test. With repetition, subjects notice that four of the items are fruits, four for cleaning, four spices, four carbs. They come in categories. And this changes subjects’ encoding strategy as they start clumping by semantic group—“Peaches. Apples. Blueberries—no, I mean blackberries. There was another fruit, can’t remember what. Okay, cornflakes, bread, doughnuts, muffins. Cumin, nutmeg—argh, again!—I mean cinnamon, oregano . . .” And throughout, the PFC imposes an overarching executive strategy for remembering these sixteen factoids.*
The PFC is essential for categorical thinking, for organizing and thinking about bits of information with different labels. The PFC groups apples and peaches as closer to each other in a conceptual map than are apples and toilet plungers. In a relevant study, monkeys were trained to differentiate between pictures of a dog and of a cat. The PFC contained individual neurons that responded to “dog” and others that responded to “cat.” Now the scientists morphed the pictures together, creating hybrids with varying percentages of dog and cat. “Dog” PFC neurons responded about as much to hybrids that were 80
percent dog and 20 percent cat, or 60:40, as to 100 percent dog. But not to 40:60—“cat” neurons would kick in there.41
The frontal cortex aids the underdog outcome, fueled by thoughts supplied from influences that fill the rest of this book—stop, those aren’t your cookies; you’ll go to hell; self-discipline is good; you’re happier when you’re thinner—all giving some lone inhibitory motor neuron more of a fighting chance.
Frontal Metabolism and an Implicit Vulnerability
This raises an important point, pertinent to the social as well as cognitive functions of the frontal cortex.42 All this “I wouldn’t do that if I were you”–ing by the frontal cortex is taxing. Other brain regions respond to instances of some contingency; the frontal cortex tracks rules. Just think how around age three, our frontal cortices learned a rule followed for the rest of our lives—don’t pee whenever you feel like it—and gained the means to enact that rule by increasing their influence over neurons regulating the bladder.
Moreover, the frontal mantra of “self-discipline is good” when cookies beckon is also invoked when economizing to increase retirement savings. Frontal cortical neurons are generalists, with broad patterns of projections, which makes for more work.43
All this takes energy, and when it is working hard, the frontal cortex has an extremely high metabolic rate and rates of activation of genes related to energy production.44 Willpower is more than just a metaphor; self-control is a finite resource. Frontal neurons are expensive cells, and expensive cells are vulnerable cells. Consistent with that, the frontal cortex is atypically vulnerable to various neurological insults.
Pertinent to this is the concept of “cognitive load.” Make the frontal cortex work hard—a tough working-memory task, regulating social behavior, or making numerous decisions while shopping. Immediately afterward performance on a different frontally dependent task declines.45 Likewise during multitasking, where PFC neurons simultaneously participate in multiple activated circuits.
Behave: The Biology of Humans at Our Best and Worst Page 5