Behave: The Biology of Humans at Our Best and Worst

by Robert M. Sapolsky

  Given its more traditional roles in mammals, the ACC’s empathy connection is unexpected. Broadly, those roles are:

  Processing interoceptive information. As introduced in chapter 3, our brains monitor sensory information not only from outside us but from our internal world as well—interoceptive information about achy muscles, dry mouths, bowels in an uproar. If you unconsciously sense that your heart is racing and that makes you experience some emotion more intensely, thank the ACC. The ACC funnels literal gut feelings into intuitions and metaphorical gut feelings influencing frontal function. Pain is a key type of interoceptive information that catches the ACC’s attention.13

  Conflict monitoring. The ACC responds to “conflict” in the sense of a discrepancy from what is expected. If you associate doing some behavior with a particular outcome, when that outcome doesn’t occur, the ACC takes notice. This monitoring of discrepancy from expectation is asymmetrical—do some task that pays two brownie points and today, unexpectedly, you get three instead, and the ACC perks up and takes notice; do the task and instead of two brownie points you only get one, and the ACC activates like mad. In the words of Kevin Ochsner of Columbia University and colleagues, the ACC is an “all-purpose alarm that signals when ongoing behavior has hit a snag.”14

  Unexpected pain is at the intersection of those two roles of the ACC, a sure sign that things are amiss with your schema about the world. Even with anticipated pain, you monitor whether it turns out to be of the quality and quantity expected. As noted, the ACC doesn’t concern itself with pedestrian concerns about pain (is it my finger or my toe that hurts?); that’s the purview of less refined, more ancient brain circuitry. What the ACC cares about is the meaning of the pain. Good news or bad, and of what nature? Thus the ACC’s perception of pain can be manipulated. Poke your finger with a pin and the ACC activates, along with those brain regions telling you which finger and what parameters of pain. Make someone believe that the inert cream you just smeared on his finger is a powerful painkiller, and when you poke his finger, the “it’s my finger, not my toe” circuitry still activates. But the ACC falls for the placebo effect and stays silent.

  Obviously the ACC receives inputs from interoceptive and exteroceptive outposts. Equally logically, it sends lots of projections into the sensorimotor cortex, making you very aware of and focused on the body part that hurts.

  But the sophistication of the ACC, the reason it sits up there in the frontal cortex, is apparent when considering another type of pain. Back to chapter 6 and the Cyberball game where subjects in brain scanners play catch with a virtual ball on a computer screen, tossing it back and forth, and the other two players stop throwing the ball to you. You’re being left out, and the ACC activates. Insofar as the ACC cares about the meaning of pain, it’s just as concerned with the abstractions of social and emotional pain—social exclusion, anxiety, disgust, embarrassment—as with physical pain. Intriguingly, major depression is associated with various abnormalities in the ACC.* And the ACC is also involved during positive resonance—when their pleasure is your pleasure.15

  All this makes the ACC sound pretty self-oriented, mighty concerned with your well-being. Which makes its empathy role initially surprising. Nonetheless, numerous studies consistently show that if someone else’s pain—a poked finger, a sad face, a tale of misfortune—is evoking an empathic state in you, the ACC is involved.16 Moreover, the more painful the other person’s situation seems to be, the more ACC activation. The ACC is also central to doing something to alleviate someone else’s distress.

  The neuropeptide/hormone oxytocin gets into the mix. Recall from chapter 4 how it promotes bonding and affiliative behaviors, trust, and generosity.* Recall the study in which prairie voles are observed consoling their stressed partner. And we’d expect, the effect depends on the actions of oxytocin. Remarkably, the oxytocin works in the ACC—selectively block oxytocin effects in the ACC, and voles don’t console.

  So how do we go from the ACC as this outpost of self-interest, monitoring your pain and whether you are getting what you think you deserve, to the ACC allowing you to feel the pain of the wretched of the earth? I think the link is a key issue of this chapter—how much is an empathic state actually about yourself?17 “Ouch, that hurt” is a good way to learn not to repeat whatever you just did. But often, even better is to monitor someone else’s misfortune—“That sure seems to have hurt her; I’m staying away from doing that.” Crucially, the ACC is essential for learning fear and conditioned avoidance by observation alone. Going from “She seems to be having a miserable time” to “Thus I should avoid that” requires an intervening step of shared representation of self: “Like her, I wouldn’t enjoy feeling that way.” Feeling someone else’s pain can be more effective for learning than just knowing that they’re in pain. At its core the ACC is about self-interest, with caring about that other person in pain as an add-on.

  Other brain regions are pertinent as well. As we saw, maturation of the circuitry of empathy involves bringing into the mix not only the ACC but the insula as well.18 By adulthood the insula (and to a lesser degree the amygdala) is nearly as intertwined with experiencing empathy as is the ACC. The three regions are highly interconnected, and a big chunk of the amygdala texting the frontal cortex is funneled through the ACC. Numerous circumstances that evoke a sense of empathy, particularly physical pain, activate the insula along with the ACC, with the magnitude of the response correlating with the subject’s basic proclivity toward empathy, or the subjective sense of empathy they are feeling in the situation.

  This makes sense, given the workings of the insula and amygdala. As we saw, their involvement in empathic states emerges developmentally as kids first embed empathy in context and causality—why is this person in pain, and whose fault is it? This is obvious when pain is rooted in injustice, when disgust, indignation, and anger sweep in because we know that this pain could have been prevented, that someone profited from it. Even when it is unclear that a cause of pain lies in injustice, we seek attribution—the intertwining of the ACC with the insula and amygdala is our world of scapegoating. And that pattern is so often there even when pain is random, without human agency or villainy—literal or metaphorical tectonic plates shift, the earth opens up and swallows someone innocent, and we rail against the people who deprived that victim of a happier life before the tragedy struck, against the God behind this act of God, against the mechanistic indifference of the universe. And as we will see, the more the purity of empathy is clouded with the anger, disgust, and indignation of blame, the harder it is to actually help.

  The Cognitive Side of Things

  When do the more cognitive components of an empathic state—the PFC, the dlPFC in particular, along with Theory of Mind networks such as the temporoparietal juncture (TPJ) and superior central sulcus—come more to the forefront? Obviously, and uninterestingly, when it’s challenging to even figure out what’s going on—“Wait, who won the game?” “Do I want my pieces to surround or be surrounded by the other person’s?”

  More interesting is when more cognitive brain circuitry is recruited by issues of causation and intentionality: “Wait, does he have a horrible headache because he’s a migrant farm worker who was sprayed with pesticide, or because he’s been binge drinking with his frat bros?” “Did this AIDS patient get his HIV from a blood transfusion or drug use? (People show more activation of the ACC for the former.) This is precisely what chimps have thought through when comforting an innocent victim of aggression but not an instigator. As we saw in chapter 7, the more cognitive profile of activation emerges when kids start distinguishing between self- and other-inflicted pain. In the words of Jean Decety, who did such research, this demonstrates that “empathic arousal [was] moderated early in information processing by a priori attitudes toward other people.”19 In other words, cognitive processes serve as a gatekeeper, deciding whether a particular misfortune is worthy of empathy.

  It is also a cognit
ive task to resonate with pain that is less overt—for example, there is more engagement of the dmPFC when observing someone in emotional pain than physical pain. Likewise when the pain is presented more abstractly—a signal on a screen indicating that someone’s hand has been stuck with a needle versus the act itself being shown. Resonating with someone else’s pain is also a cognitive task when it is a type of pain that you haven’t experienced. “Well, I suppose I can understand the disappointment of this militia leader when he was passed over for the chance to carry out the ethnic cleansing—kinda like when I lost the election in kindergarten to be president of the random-act-of-kindness club.” Now, that takes cognitive work. In one study subjects considered people suffering from a neurological disorder involving a novel type of pain sensitivity; empathizing for that novel pain involved more frontal cortical activation than for more conventional pain.20

  As we saw, the rudimentary “empathy” of rodents is contingent, depending on whether the other individual is a cagemate or a stranger.21 It is an enormous cognitive task for humans to overcome that, to reach an empathic state for someone who is different, unappealing. A hospital chaplain once described to me how he has to actively make sure that he is not preferentially visiting patients who were “YAVIS”—young, attractive, verbal, intelligent, or social. This is straight out of Us versus Them—recall Susan Fiske’s work showing how extreme out-group members, such as the homeless or addicts, are processed differently in the frontal cortex than other people. And it is also straight out of Josh Greene’s tragedy of the commons versus tragedy of the commonsense morality, where acting morally toward an Us is automatic, while doing so for a Them takes work.

  The ease of empathizing with people like us starts at the level of autonomic building blocks of empathy—in one study of ritual fire walkers in Spain, heart rate changes in the walkers synchronized with spectators—but only those who were relatives. In line with that distinction, taking the perspective of a loved one in pain activates the ACC; doing the same for a stranger activates the TPJ, that region central to Theory of Mind.22

  This extends to broader versions of Us versus Them. As introduced in chapter 3, we have a stronger sensorimotor response in our hands when the hand we see being poked with a needle is of our race; the stronger one’s implicit in-group bias, the stronger this effect. Meanwhile, other studies show that the stronger the discrepancy in patterns of neural activation when observing an in-group versus an out-group person in pain, the lower the chances of helping the latter.23 Thus it’s no surprise that feeling the same degree of empathy or achieving the same level of perspective taking for a Them as for an Us requires greater frontocortical activation. This is the domain where you must suppress the automatic and implicit urges to be indifferent, if not repulsed, and do the creative, motivated work of finding the affective commonalities.*24

  Categorical boundaries to the extension of empathy also run along socioeconomic lines, but in an asymmetrical manner. What does that mean? That when it comes to empathy and compassion, rich people tend to suck. This has been explored at length in a series of studies by Dacher Keltner of UC Berkeley. Across the socioeconomic spectrum, on the average, the wealthier people are, the less empathy they report for people in distress and the less compassionately they act. Moreover, wealthier people are less adept at recognizing other people’s emotions and in experimental settings are greedier and more likely to cheat or steal. Two of the findings were picked up by the media as irresistible: (a) wealthier people (as assessed by the cost of the car they were driving) are less likely than poor people to stop for pedestrians at crosswalks; (b) suppose there’s a bowl of candy in the lab; invite test subjects, after they finish doing some task, to grab some candy on the way out, telling them that whatever’s left over will be given to some kids—the wealthier take more candy.25

  So do miserable, greedy, unempathic people become wealthy, or does being wealthy increase the odds of a person’s becoming that way? As a cool manipulation, Keltner primed subjects to focus either on their socioeconomic success (by asking them to compare themselves with people less well off than them) or on the opposite. Make people feel wealthy, and they take more candy from children.

  What explains this pattern? A number of interrelated factors, built around the system justification described in chapter 12—wealthier people are more likely to endorse greed as being good, to view the class system as fair and meritocratic, and to view their success as an act of independence—all great ways to decide that someone else’s distress is beneath your notice or concern.

  It is a particularly uphill battle when we are asked to empathize with the pain of people we dislike, whom we morally disapprove of—remember how their misfortune doesn’t simply fail to activate the ACC but instead it activates mesolimbic dopamine reward pathways. Thus the process of taking their perspective and feeling their pain (as other than grounds for gloating) is a dramatic cognitive challenge rather than something remotely automatic.26

  The cognitive “costs” of empathizing with someone distant are shown by increasing people’s cognitive load (i.e., making their frontal cortex work harder by forcing it to override a habitual behavior)—they become less helpful to strangers but not to family members. “Empathy fatigue” can thus be viewed as the state when the cognitive load of repeated exposure to the pain of Thems whose perspective is challenging to take has exhausted the frontal cortex. The notions of cognitive work and load also help explain why people are more charitable when contemplating one person in need than a group. To quote Mother Teresa, “If I look at the mass, I will never act. If I look at the one, I will.” Or to cite a quote attributed to someone who never seems to have achieved enough empathy to be vulnerable to empathy fatigue, Joseph Stalin: “The death of one man is a tragedy; the death of millions is a statistic.”27

  And probably most reliably, those mentalizing pathways are activated when we switch from focusing on what it would feel like if this were happening to us to focusing on what it must feel like for them. Thus when subjects are instructed to switch from first- to third-person perspective, there’s not just activation of the TPJ but also frontal activation with the top-down regulatory task “Stop thinking about yourself.”28

  Thus we have themes that closely resemble those from the last few chapters. When it comes to empathic states, “emotion” and “cognition” are totally false dichotomies; you need both, but with the balance between the two shifting on a continuum, and the cognition end of it has to do the heavy lifting when the differences between you and the person in pain initially swap the similarities.

  Time now for one of the great sideshows in empathy science.

  A MYTHIC LEAP FORWARD

  In the early 1990s scientists at the University of Parma in Italy, led by Giacomo Rizzolatti and Vittorio Gallese, reported something that, depending on your tastes, ranged from really interesting to revolutionary. They had been studying an area of the brain called the premotor cortex (PMC) in rhesus monkeys, examining what sorts of stimuli would cause individual neurons there to activate. Back to the PMC from chapter 2. “Executive” neurons in the PFC decide something, passing the news to the rest of the frontal cortex just behind it. Which sends projections to the PMC just behind it. Which sends projections one step further back, to the motor cortex, which then sends commands to muscles. Thus the PMC straddles the divide between thinking about and carrying out a movement.29

  The group had discovered some mighty quirky PMC neurons. Suppose a monkey carried out a behavior—grasping some food and bringing it to her mouth. Naturally, some neurons in the PMC would have activated. If she did a different movement—grasping an object and placing it in a container—a different (partially overlapping) array of PMC neurons were involved. What the group reported was that some of the bring-food-to-mouth neurons would also activate if the monkey observed someone else (monkey or human) making that movement. Same for some of the place-object-in-container neurons. Same for subtler movem
ents like facial expressions. Consistently, about 10 percent of the PMC neurons devoted to doing movement X also activated when observing someone else doing movement X—very odd for neurons a few steps away from commanding muscles to move. The neurons were concerned with the mirroring of movements. And thus were “mirror neurons” announced to the world.

  Naturally, everyone looked for mirror neurons in humans, and their existence in roughly the same part of the brain*30 was soon inferred with brain imaging studies (“inferred” because that approach tells you about the activity of large numbers of neurons at a time, rather than single ones). Then individual neurons were shown to be mirroresque in humans (in patients undergoing neurosurgery to control a rare type of epilepsy).31

  The mirroring can be quite abstract. It can be cross-modal—see someone doing movement A, and some mirror neuron activates; hear the sound of someone doing movement A, and the same occurs. And the neurons can gestalt a scene, firing even if only part of the observed movement is obscured.32

  Most interesting, mirror neurons didn’t simply track movement. Find a mirror neuron that responds to the sight of someone picking up a cup of tea to drink. The sight of someone picking up the tea to clear the table doesn’t activate it. In other words, mirror neurons can incorporate intentionality into their response.

  Thus mirror neuron activity correlates with circumstances of imitation, either conscious or otherwise, including imitating the idea of an action, as well as the intent behind it. Nevertheless, no one has actually shown a causal relationship, that automatic or conscious mimicry requires mirror neuron activation. Moreover, the mirror neuron/imitation link is complicated by the cells having been first identified in rhesus monkeys—a species that does not show imitation of behavior.

  But assuming that mirror neurons are indeed involved, the question becomes what purpose mimicry serves. Various possibilities have been raised and debated.