We’re now ready to examine the brain regions most central to this book, namely the amygdala, the frontal cortex, and the mesolimbic/mesocortical dopamine system (discussion of other bit-player regions will be subsumed under the headings for these three). We start with the one arguably most central to our worst behaviors.
THE AMYGDALA
The amygdala* is the archetypal limbic structure, sitting under the cortex in the temporal lobe. It is central to mediating aggression, along with other behaviors that tell us tons about aggression.
A First Pass at the Amygdala and Aggression
The evidence for the amygdala’s role in aggression is extensive, based on research approaches that will become familiar.
First there’s the correlative “recording” approach. Stick recording electrodes into numerous species’ amygdalae* and see when neurons there have action potentials; this turns out to be when the animal is being aggressive.* In a related approach, determine which brain regions consume extra oxygen or glucose, or synthesize certain activity-related proteins, during aggression—the amygdala tops the list.
Moving beyond mere correlation, if you lesion the amygdala in an animal, rates of aggression decline. The same occurs transiently when you temporarily silence the amygdala by injecting Novocain into it. Conversely, implanting electrodes that stimulate neurons there, or spritzing in excitatory neurotransmitters (stay tuned), triggers aggression.4
Show human subjects pictures that provoke anger, and the amygdala activates (as shown with neuroimaging). Sticking an electrode in someone’s amygdala and stimulating it (as is done before certain types of neurosurgery) produces rage.
The most convincing data concern rare humans with damage restricted to the amygdala, either due to a type of encephalitis or a congenital disorder called Urbach-Wiethe disease, or where the amygdala was surgically destroyed to control severe, drug-resistant seizures originating there.5 Such individuals are impaired in detecting angry facial expressions (while being fine at recognizing other emotional states—stay tuned).
And what does amygdala damage do to aggressive behavior? This was studied in humans where amygdalotomies were done not to control seizures but to control aggression. Such psychosurgery provoked fiery controversy in the 1970s. And I don’t mean scientists not saying hello to each other at conferences. I mean a major public shit storm.
The issue raised bioethical lightning rods: What counted as pathological aggression? Who decided? What other interventions had been tried unsuccessfully? Were some types of hyperaggressive individuals more likely to go under the knife than others? What constituted a cure?6
Most of these cases concerned rare epileptics where seizure onset was associated with uncontrollable aggression, and where the goal was to contain that behavior (these papers had titles such as “Clinical and physiological effects of stereotaxic bilateral amygdalotomy for intractable aggression”). The fecal hurricane concerned the involuntary lopping out of the amygdala in people without epilepsy but with a history of severe aggression. Well, doing this could be profoundly helpful. Or Orwellian. This is a long, dark story and I will save it for another time.
Did destruction of the human amygdala lessen aggression? Pretty clearly so, when violence was a reflexive, inchoate outburst preceding a seizure. But with surgery done solely to control behavior, the answer is, er, maybe—the heterogeneity of patients and surgical approaches, the lack of modern neuroimaging to pinpoint exactly which parts of the amygdala were destroyed in each individual, and the imprecision in the behavioral data (with papers reporting from 33 to 100 percent “success” rates) make things inconclusive. The procedure has almost entirely fallen out of practice.
The amygdala/aggression link pops up in two notorious cases of violence. The first concerns Ulrike Meinhof, a founder in 1968 of the Red Army Faction (aka the Baader-Meinhof Gang), a terrorist group responsible for bombings and bank robberies in West Germany. Meinhof had a conventional earlier life as a journalist before becoming violently radicalized. During her 1976 murder trial, she was found hanged in her jail cell (suicide or murder? still unclear). In 1962 Meinhof had had a benign brain tumor surgically removed; the 1976 autopsy showed that remnants of the tumor and surgical scar tissue impinged on her amygdala.7
A second case concerns Charles Whitman, the 1966 “Texas Tower” sniper who, after killing his wife and mother, opened fire atop a tower at the University of Texas in Austin, killing sixteen and wounding thirty-two, one of the first school massacres. Whitman was literally an Eagle Scout and childhood choirboy, a happily married engineering major with an IQ in the 99th percentile. In the prior year he had seen doctors, complaining of severe headaches and violent impulses (e.g., to shoot people from the campus tower). He left notes by the bodies of his wife and his mother, proclaiming love and puzzlement at his actions: “I cannot rationaly [sic] pinpoint any specific reason for [killing her],” and “let there be no doubt in your mind that I loved this woman with all my heart.” His suicide note requested an autopsy of his brain, and that any money he had be given to a mental health foundation. The autopsy proved his intuition correct—Whitman had a glioblastoma tumor pressing on his amygdala. Did Whitman’s tumor “cause” his violence? Probably not in a strict “amygdaloid tumor = murderer” sense, as he had risk factors that interacted with his neurological issues. Whitman grew up being beaten by his father and watching his mother and siblings experience the same. This choirboy Eagle Scout had repeatedly physically abused his wife and had been court-martialed as a Marine for physically threatening another soldier.* And, perhaps indicative of a thread running through the family, his brother was murdered at age twenty-four during a bar fight.8
A Whole Other Domain of Amygdaloid Function to the Center Stage
Thus considerable evidence implicates the amygdala in aggression. But if you asked amygdala experts what behavior their favorite brain structure brings to mind, “aggression” wouldn’t top their list. It would be fear and anxiety.9 Crucially, the brain region most involved in feeling afraid and anxious is most involved in generating aggression.
The amygdala/fear link is based on evidence similar to that supporting the amygdala/aggression link.10 In lab animals this has involved lesioning the structure, detecting activity in its neurons with “recording electrodes,” electrically stimulating it, or manipulating genes in it. All suggest a key role for the amygdala in perceiving fear-provoking stimuli and in expressing fear. Moreover, fear activates the amygdala in humans, with more activation predicting more behavioral signs of fear.
In one study subjects in a brain scanner played a Ms. Pac-Man–from–hell video game where they were pursued in a maze by a dot; if caught, they’d be shocked.11 When people were evading the dot, the amygdala was silent. However, its activity increased as the dot approached; the stronger the shocks, the farther away the dot would be when first activating the amygdala, the stronger the activation, and the larger the self-reported feeling of panic.
In another study subjects waited an unknown length of time to receive a shock.12 This lack of predictability and control was so aversive that many chose to receive a stronger shock immediately. And in the others the period of anticipatory dread increasingly activated the amygdala.
Thus the human amygdala preferentially responds to fear-evoking stimuli, even stimuli so fleeting as to be below conscious detection.
Powerful support for an amygdaloid role in fear processing comes from post-traumatic stress disorder (PTSD). In PTSD sufferers the amygdala is overreactive to mildly fearful stimuli and is slow in calming down after being activated.13 Moreover, the amygdala expands in size with long-term PTSD. This role of stress in this expansion will be covered in chapter 4.
The amygdala is also involved in the expression of anxiety.14 Take a deck of cards—half are black, half are red; how much would you wager that the top card is red? That’s about risk. Here’s a deck of cards—at least one is black, at least one is
red; how much would you wager that the top card is red? That’s about ambiguity. The circumstances carry identical probabilities, but people are made more anxious by the second scenario and activate the amygdala more. The amygdala is particularly sensitive to unsettling circumstances that are social. A high-ranking male rhesus monkey is in a sexual consortship with a female; in one condition the female is placed in another room, where the male can see her. In the second she’s in the other room along with a rival of the male. No surprise, that situation activates the amygdala. Is that about aggression or anxiety? Seemingly the latter—the extent of activation did not correlate with the amount of aggressive behaviors and vocalizations the male made, or the amount of testosterone secreted. Instead, it correlated with the extent of anxiety displayed (e.g., teeth chattering, or self-scratching).
The amygdala is linked to social uncertainty in other ways. In one neuroimaging study, a subject would participate in a competitive game against a group of other players; outcomes were rigged so that the subject would wind up in the middle of the rankings.15 Experimenters then manipulated game outcomes so that subjects’ rankings either remained stable or fluctuated wildly. Stable rankings activated parts of the frontal cortex that we’ll soon consider. Instability activated the frontal cortex plus the amygdala. Being unsure of your place is unsettling.
Another study explored the neurobiology of conforming.16 To simplify, a subject is part of a group (where, secretly, the rest are confederates); they are shown “X,” then asked, “What did you see?” Everyone else says “Y.” Does the subject lie and say “Y” also? Often. Subjects who stuck to their guns with “X” showed amygdala activation.
Finally, activating specific circuits within the amygdala in mice turns anxiety on and off; activating others made mice unable to distinguish between safe and anxiety-producing settings.*17
The amygdala also helps mediate both innate and learned fear.18 The core of innate fear (aka a phobia) is that you don’t have to learn by trial and error that something is aversive. For example, a rat born in a lab, who has interacted only with other rats and grad students, instinctually fears and avoids the smell of cats. While different phobias activate somewhat different brain circuitry (for example, dentist phobia involves the cortex more than does snake phobia), they all activate the amygdala.
Such innate fear contrasts with things we learn to fear—a bad neighborhood, a letter from the IRS. The dichotomy between innate and learned fear is actually a bit fuzzy.19 Everyone knows that humans are innately afraid of snakes and spiders. But some people keep them as pets, give them cute names.* Instead of inevitable fear, we show “prepared learning”—learning to be afraid of snakes and spiders more readily than of pandas or beagles.
The same occurs in other primates. For example, lab monkeys who have never encountered snakes (or artificial flowers) can be conditioned to fear the former more readily than the latter. As we’ll see in the next chapter, humans show prepared learning, being predisposed to be conditioned to fear people with a certain type of appearance.
The fuzzy distinction between innate and learned fear maps nicely onto the amygdala’s structure. The evolutionarily ancient central amygdala plays a key role in innate fears. Surrounding it is the basolateral amygdala (BLA), which is more recently evolved and somewhat resembles the fancy, modern cortex. It’s the BLA that learns fear and then sends the news to the central amygdala.
Joseph LeDoux at New York University has shown how the BLA learns fear.*20 Expose a rat to an innate trigger of fear—a shock. When this “unconditioned stimulus” occurs, the central amygdala activates, stress hormones are secreted, the sympathetic nervous system mobilizes, and, as a clear end point, the rat freezes in place—“What was that? What do I do?” Now do some conditioning. Before each shock, expose the rat to a stimulus that normally does not evoke fear, such as a tone. And with repeated coupling of the tone (the conditioned stimulus) with the shock (the unconditioned one), fear conditioning occurs—the sound of the tone alone elicits freezing, stress hormone release, and so on.*
LeDoux and others have shown how auditory information about the tone stimulates BLA neurons. At first, activation of those neurons is irrelevant to the central amygdala (whose neurons are destined to activate following the shock). But with repeated coupling of tone with shock, there is remapping and those BLA neurons acquire the means to activate the central amygdala.*
BLA neurons that respond to the tone only once conditioning has occurred would also have responded if conditioning instead had been to a light. In other words, these neurons respond to the meaning of the stimulus, rather than to its specific modality. Moreover, if you electrically stimulate them, rats are easier to fear-condition; you’ve lowered the threshold for this association to be made. And if you electrically stimulate the auditory sensory input at the same time as shocks (i.e., there’s no tone, just activation of the pathway that normally carries news of the tone to the amygdala), you cause fear conditioning to a tone. You’ve engineered the learning of a false fear.
There are synaptic changes as well. Once conditioning to a tone has occurred, the synapses coupling the BLA and central nucleus neurons have become more excitable; how this occurs is understood at the level of changes in the amount of receptors for excitatory neurotransmitters in dendritic spines in these circuits.* Furthermore, conditioning increases levels of “growth factors,” which prompt the growth of new connections between BLA and central amygdala neurons; some of the genes involved have even been identified.
We’ve now got learning to be afraid under our belts.*21 Now conditions change—the tone still occurs now and then, but no more shock. Gradually the conditioned fear response abates. How does “fear extinction” occur? How do we learn that this person wasn’t so scary after all, that different doesn’t necessarily equal frightening? Recall how a subset of BLA neurons respond to the tone only once conditioning has occurred. Another population does the opposite, responding to the tone once it’s no longer signaling shock (logically, the two populations of neurons inhibit each other). Where do these “Ohhh, the tone isn’t scary anymore” neurons get inputs from? The frontal cortex. When we stop fearing something, it isn’t because some amygdaloid neurons have lost their excitability. We don’t passively forget that something is scary. We actively learn that it isn’t anymore.*
The amygdala also plays a logical role in social and emotional decision making. In the Ultimatum Game, an economic game involving two players, the first makes an offer as to how to divide a pot of money, which the other player either accepts or rejects.22 If the latter, neither gets anything. Research shows that rejecting an offer is an emotional decision, triggered by anger at a lousy offer and the desire to punish. The more the amygdala activation in the second player after an offer, the more likely the rejection. People with damaged amygdalae are atypically generous in the Ultimatum Game and don’t increase rejection rates if they start receiving unfair offers.
Why? These individuals understand the rules and can give sound, strategic advice to other players. Moreover, they use the same strategies as control subjects in a nonsocial version of the game, when believing the other player is a computer. And they don’t have a particularly long view, undistracted by the amygdala’s emotional tumult, reasoning that their noncontingent generosity will induce reciprocity and pay off in the long run. When asked, they anticipate the same levels of reciprocity as do controls.
Instead, these findings suggest that the amygdala injects implicit distrust and vigilance into social decision making.23 All thanks to learning. In the words of the authors of the study, “The generosity in the trust game of our BLA-damaged subjects might be considered pathological altruism, in the sense that inborn altruistic behaviors have not, due to BLA damage, been un-learned through negative social experience.” In other words, the default state is to trust, and what the amygdala does is learn vigilance and distrust.
Unexpectedly, the amygdala and one o
f its hypothalamic targets also play a role in male sexual motivation (other hypothalamic nuclei are central to male sexual performance)* but not female.* What’s that about? One neuroimaging study sheds some light. “Young heterosexual men” looked at pictures of attractive women (versus, as a control, of attractive men). Passively observing the pictures activated the reward circuitry just alluded to. In contrast, working to see the pictures—by repeatedly pressing a button—also activated the amygdala. Similarly, other studies show that the amygdala is most responsive to positive stimuli when the value of the reward is shifting. Moreover, some BLA neurons that respond in that circumstance also respond when the severity of something aversive is shifting—these neurons are paying attention to change, independent of direction. For them, “the amount of reward is changing” and “the amount of punishment is changing” are the same. Studies like these clarify that the amygdala isn’t about the pleasure of experiencing pleasure. It’s about the uncertain, unsettled yearning for a potential pleasure, the anxiety and fear and anger that the reward may be smaller than anticipated, or may not even happen. It’s about how many of our pleasures and our pursuits of them contain a corrosive vein of disease.*24
The Amygdala as Part of Networks in the Brain
Now that we know about the subparts of the amygdala, it’s informative to consider its extrinsic connections—i.e., what parts of the brain send projection to it, and what parts does it project to?25
Behave: The Biology of Humans at Our Best and Worst Page 4