Behave: The Biology of Humans at Our Best and Worst

by Robert M. Sapolsky

  Soon other species were heard from. Oxytocin lets sheep learn the smell of their offspring and facilitates female monkeys grooming their offspring. Spray oxytocin up a woman’s nose (a way to get the neuropeptide past the blood-brain barrier and into the brain), and she’ll find babies to look more appealing. Moreover, women with variants of genes that produce higher levels of oxytocin or oxytocin receptors average higher levels of touching their infants and more synchronized gazing with them.

  So oxytocin is central to female mammals nursing, wanting to nurse their child, and remembering which one is their child. Males then got into the act, as vasopressin plays a role in paternal behavior. A female rodent giving birth increases vasopressin and vasopressin receptor levels throughout the body, including the brain, of the nearby father. Among monkeys, experienced fathers have more dendrites in frontal cortical neurons containing vasopressin receptors. Moreover, administering vasopressin enhances paternal behaviors. However, an ethological caveat: this occurs only in species where males are paternal (e.g., prairie voles and marmoset monkeys).24*

  Then, dozens of millions of years ago, some rodent and primate species independently evolved monogamous pair-bonding, along with the neuropeptides central to the process.25 Among marmoset and titi monkeys, which both pair-bond, oxytocin strengthens the bond, increasing a monkey’s preference for huddling with her partner over huddling with a stranger. Then there was a study that is embarrassingly similar to stereotypical human couples. Among pair-bonding tamarin monkeys, lots of grooming and physical contact predicted high oxytocin levels in female members of a pair. What predicted high levels of oxytocin in males? Lots of sex.

  Beautiful, pioneering work by Thomas Insel of the National Institute of Mental Health, Larry Young of Emory University, and Sue Carter of the University of Illinois has made a species of vole arguably the most celebrated rodent on earth.26 Most voles (e.g., montane voles) are polygamous. In contrast, prairie voles, in a salute to Garrison Keillor, form monogamous mating pairs for life. Naturally, this isn’t quite the case—while they are “social pair-bonders” with their permanent relationships, they’re not quite perfect “sexual pair-bonders,” as males might mess around on the side. Nonetheless, prairie voles pair-bond more than other voles, prompting Insel, Young, and Carter to figure out why.

  First finding: sex releases oxytocin and vasopressin in the nucleus accumbens of female and male voles, respectively. Obvious theory: prairie voles release more of the stuff during sex than do polygamous voles, causing a more rewarding buzz, encouraging the individuals to stick with their partner. But prairie voles don’t release more neuropeptides than montane voles. Instead, prairie voles have more of the pertinent receptors in the nucleus accumbens than do polygamous voles.* Moreover, male prairie voles with a variant of the vasopressin receptor gene that produced more receptors in the nucleus accumbens were stronger pair-bonders. Then the scientists conducted two tour de force studies. First they engineered the brains of male mice to express the prairie vole version of the vasopressin receptor in their brains, and they groomed and huddled more with familiar females (but not with strangers). Then the scientists engineered the brains of male montane voles to have more vasopressin receptors in the nucleus accumbens; the males became more socially affiliative with individual females.*

  What about versions of vasopressin receptor genes in other species? When compared with chimps, bonobos have a variant associated with more receptor expression and far more social bonding between females and males (although, in contrast to prairie voles, bonobos are anything but monogamous).27

  How about humans? This is tough to study, because you can’t measure these neuropeptides in tiny brain regions in humans and instead have to examine levels in the circulation, a fairly indirect measure.

  Nevertheless, these neuropeptides appear to play a role in human pair-bonding.28 For starters, circulating oxytocin levels are elevated in couples when they’ve first hooked up. Furthermore, the higher the levels, the more physical affection, the more behaviors are synchronized, the more long-lasting the relationship, and the happier interviewers rate couples to be.

  Even more interesting were studies where oxytocin (or a control spray) was administered intranasally. In one fun study, couples had to discuss one of their conflicts; oxytocin up their noses, and they’d be rated as communicating more positively and would secrete less stress hormones. Another study suggests that oxytocin unconsciously strengthens the pair-bond. Heterosexual male volunteers, with or without an oxytocin spritz, interacted with an attractive female researcher, doing some nonsense task. Among men in stable relationships, oxytocin increased their distance from the woman an average of four to six inches. Single guys, no effect. (Why didn’t oxytocin make them stand closer? The researchers indicated that they were already about as close as one could get away with.) If the experimenter was male, no effect. Moreover, oxytocin caused males in relationships to spend less time looking at pictures of attractive women. Importantly, oxytocin didn’t make men rate these women as less attractive; they were simply less interested.29

  Thus, oxytocin and vasopressin facilitate bonding between parent and child and between couples.* Now for something truly charming that evolution has cooked up recently. Sometime in the last fifty thousand years (i.e., less than 0.1 percent of the time that oxytocin has existed), the brains of humans and domesticated wolves evolved a new response to oxytocin: when a dog and its owner (but not a stranger) interact, they secrete oxytocin.30 The more of that time is spent gazing at each other, the bigger the rise. Give dogs oxytocin, and they gaze longer at their humans . . . which raises the humans’ oxytocin levels. So a hormone that evolved for mother-infant bonding plays a role in this bizarre, unprecedented form of bonding between species.

  In line with its effects on bonding, oxytocin inhibits the central amygdala, suppresses fear and anxiety, and activates the “calm, vegetative” parasympathetic nervous system. Moreover, people with an oxytocin receptor gene variant associated with more sensitive parenting also have less of a cardiovascular startle response. In the words of Sue Carter, exposure to oxytocin is “a physiological metaphor for safety.” Furthermore, oxytocin reduces aggression in rodents, and mice whose oxytocin system was silenced (by deleting the gene for oxytocin or its receptor) were abnormally aggressive.31

  Other studies showed that people rate faces as more trustworthy, and are more trusting in economic games, when given oxytocin (oxytocin had no effect when someone thought they were playing with a computer, showing that this was about social behavior).32 This increased trust was interesting. Normally, if the other player does something duplicitous in the game, subjects are less trusting in subsequent rounds; in contrast, oxytocin-treated investors didn’t modify their behavior in this way. Stated scientifically, “oxytocin inoculated betrayal aversion among investors”; stated caustically, oxytocin makes people irrational dupes; stated more angelically, oxytocin makes people turn the other cheek.

  More prosocial effects of oxytocin emerged. It made people better at detecting happy (versus angry, fearful, or neutral) faces or words with positive (versus negative) social connotations, when these were displayed briefly. Moreover, oxytocin made people more charitable. People with the version of the oxytocin receptor gene associated with more sensitive parenting were rated by observers as more prosocial (when discussing a time of personal suffering), as well as more sensitive to social approval. And the neuropeptide made people more responsive to social reinforcement, enhancing performance in a task where correct or wrong answers elicited a smile or frown, respectively (while having no effect when right and wrong answers elicited different-colored lights).33

  So oxytocin elicits prosocial behavior, and oxytocin is released when we experience prosocial behavior (being trusted in a game, receiving a warm touch, and so on). In other words, a warm and fuzzy positive feedback loop.34

  Obviously, oxytocin and vasopressin are the grooviest hormones in the uni
verse.* Pour them into the water supply, and people will be more charitable, trusting, and empathic. We’d be better parents and would make love, not war (mostly platonic love, though, since people in relationships would give wide berths to everyone else). Best of all, we’d buy all sorts of useless crap, trusting the promotional banners in stores once oxytocin starts spraying out of the ventilation system.

  Okay, time to settle down a bit.

  Prosociality Versus Sociality

  Are oxytocin and vasopressin about prosociality or social competence? Do these hormones make us see happy faces everywhere or become more interested in gathering accurate social information about faces? The latter isn’t necessarily prosocial; after all, accurate information about someone’s emotions makes them easier to manipulate.

  The Groovy Neuropeptide School supports the idea of ubiquitous prosociality.35 But the neuropeptides also foster social interest and competence. They make people look at eyes longer, increasing accuracy in reading emotions. Moreover, oxytocin enhances activity in the temporoparietal juncture (that region involved in Theory of Mind) when people do a social-recognition task. The hormone increases the accuracy of assessments of other people’s thoughts, with a gender twist—women improve at detecting kinship relations, while men improve at detecting dominance relations. In addition, oxytocin increases accuracy in remembering faces and their emotional expressions, and people with the “sensitive parenting” oxytocin receptor gene variant are particularly adept at assessing emotions. Similarly, the hormones facilitate rodents’ learning of an individual’s smell, but not nonsocial odors.

  Neuroimaging research shows that these neuropeptides are about social competence, as well as prosociality.36 For example, variants of a gene related to oxytocin signaling* are associated with differing degrees of activation of the fusiform face area when looking at faces.

  Findings like these suggest that abnormalities in these neuropeptides increase the risk of disorders of impaired sociality, namely autism spectrum disorders (ASD) (strikingly, people with ASD show blunted fusiform responses to faces).37 Remarkably, ASD has been linked to gene variants related to oxytocin and vasopressin, to nongenetic mechanisms for silencing the oxytocin receptor gene, and to lower levels of the receptor itself. Moreover, the neuropeptides improve social skills in some individuals with ASD—e.g., enhancing eye contact.

  Thus, sometimes oxytocin and vasopressin make us more prosocial, but sometimes they make us more avid and accurate social information gatherers. Nonetheless, there is a happy-face bias, since accuracy is most enhanced for positive emotions.38

  Time for more complications.

  Contingent Effects of Oxytocin and Vasopressin

  Recall testosterone’s contingent effects (e.g., making a monkey more aggressive, but only toward individuals he already dominates). Naturally, these neuropeptides’ effects are also contingent.39

  One factor already mentioned is gender: oxytocin enhances different aspects of social competence in women and men. Moreover, oxytocin’s calming effects on the amygdala are more consistent in men than in women. Predictably, neurons that make these neuropeptides are regulated by both estrogen and testosterone.40

  As a really interesting contingent effect, oxytocin enhances charitability—but only in people who are already so. This mirrors testosterone’s only raising aggression in aggression-prone people. Hormones rarely act outside the context of the individual and his or her environment.41

  Finally, a fascinating study shows cultural contingencies in oxytocin’s actions.42 During stress, Americans seek emotional support (e.g., telling a friend about their problem) more readily than do East Asians. In one study oxytocin receptor gene variants were identified in American and Korean subjects. Under unstressful circumstances, neither cultural background nor receptor variant affected support-seeking behavior. During stressful periods, support seeking rose among subjects with the receptor variant associated with enhanced sensitivity to social feedback and approval—but only among the Americans (including Korean Americans). What does oxytocin do to support-seeking behavior? It depends on whether you’re stressed. And on the genetic variant of your oxytocin receptor. And on your culture. More to come in chapters 8 and 9.

  And the Dark Side of These Neuropeptides

  As we saw, oxytocin (and vasopressin) decreases aggression in rodent females. Except for aggression in defense of one’s pups, which the neuropeptide increases via effects in the central amygdala (with its involvement in instinctual fear).43

  This readily fits with these neuropeptides enhancing maternalism, including snarling don’t-get-one-step-closer maternalism. Similarly, vasopressin enhances aggression in paternal prairie vole males. This finding comes with a familiar additional contingency. The more aggressive the male prairie vole, the less that aggression decreases after blocking of his vasopressin system—just as in the case of testosterone, with increased experience, aggression is maintained by social learning rather than by a hormone/neuropeptide. Moreover, vasopressin increases aggression most in male rodents who are already aggressive—yet another biological effect depending on individual and social context.44

  And now to really upend our view of these feel-good neuropeptides. For starters, back to oxytocin enhancing trust and cooperation in an economic game—but not if the other player is anonymous and in a different room. When playing against strangers, oxytocin decreases cooperation, enhances envy when luck is bad, and enhances gloating when it’s good.45

  Finally, beautiful studies by Carsten de Dreu of the University of Amsterdam showed just how unwarm and unfuzzy oxytocin can be.46 In the first, male subjects formed two teams; each subject chose how much of his money to put into a pot shared with teammates. As usual, oxytocin increased such generosity. Then participants played the Prisoner’s Dilemma with someone from the other team.* When financial stakes were high, making subjects more motivated, oxytocin made them more likely to preemptively stab the other player in the back. Thus, oxytocin makes you more prosocial to people like you (i.e., your teammates) but spontaneously lousy to Others who are a threat. As emphasized by De Dreu, perhaps oxytocin evolved to enhance social competence to make us better at identifying who is an Us.

  In De Dreu’s second study, Dutch student subjects took the Implicit Association Test of unconscious bias.* And oxytocin exaggerated biases against two out-groups, namely Middle Easterners and Germans.47

  Then came the study’s truly revealing second part. Subjects had to decide whether it was okay to kill one person in order to save five. In the scenario the potential sacrificial lamb’s name was either stereotypically Dutch (Dirk or Peter), German (Markus or Helmut), or Middle Eastern (Ahmed or Youssef); the five people in danger were unnamed. Remarkably, oxytocin made subjects less likely to sacrifice good ol’ Dirk or Peter, rather than Helmut or Ahmed.

  Oxytocin, the luv hormone, makes us more prosocial to Us and worse to everyone else. That’s not generic prosociality. That’s ethnocentrism and xenophobia. In other words, the actions of these neuropeptides depend dramatically on context—who you are, your environment, and who that person is. As we will see in chapter 8, the same applies to the regulation of genes relevant to these neuropeptides.

  THE ENDOCRINOLOGY OF AGGRESSION IN FEMALES

  Help!

  This topic confuses me. Here’s why:

  This is a domain where the ratios of two hormones can matter more than their absolute levels, where the brain responds the same way to (a) two units of estrogen plus one unit of progesterone and (b) two gazillion units of estrogen plus one gazillion units of progesterone. This requires some complex neurobiology.

  Hormone levels are extremely dynamic, with hundredfold changes in some within hours—no male’s testes ever had to navigate the endocrinology of ovulation or childbirth. Among other things, re-creating such endocrine fluctuations in lab animals is tough.

  There’s dizzying variability across species. Some breed yea
r-round, others only in particular seasons; nursing inhibits ovulation in some, stimulates it in others.

  Progesterone rarely works in the brain as itself. Instead it’s usually converted into various “neurosteroids” with differing actions in different brain regions. And “estrogen” describes a soup of related hormones, none of which work identically.

  Finally, one must debunk the myth that females are always nice and affiliative (unless, of course, they’re aggressively protecting their babies, which is cool and inspirational).

  Maternal Aggression

  Levels of aggression rise in rodents during pregnancy, peaking around parturition.*48 Appropriately, the highest levels occur in species and breeds with the greatest threat of infanticide.49

  During late pregnancy, estrogen and progesterone increase maternal aggression by increasing oxytocin release in certain brain regions, bringing us back to oxytocin promoting maternal aggression.50

  Two complications illustrate some endocrine principles.* Estrogen contributes to maternal aggression. But estrogen can also reduce aggression and enhance empathy and emotional recognition. It turns out there are two different types of receptors for estrogen in the brain, mediating these opposing effects and with their levels independently regulated. Thus, same hormone, same levels, different outcome if the brain is set up to respond differently.51

  The other complication: As noted, progesterone, working with estrogen, promotes maternal aggression. However, on its own it decreases aggression and anxiety. Same hormone, same levels, diametrically opposite outcomes depending on the presence of a second hormone.52

  Progesterone decreases anxiety through a thoroughly cool route. When it enters neurons, it is converted to another steroid;* this binds to GABA receptors, making them more sensitive to the inhibitory effects of GABA, thereby calming the brain. Thus, direct cross-talk between hormones and neurotransmitters.