Aggression III. 105m Robert Sapolsky video.

Part 3 of 4 in a broad biological survey on empathy and aggression. The discussion drills deeper into the timeline of biological causes starting with the neurology and reviewing how the brain mixes reality and metaphor in its moral wiring. Then a fascinating and deep discussion of the role of serotonin in aggression and impulsive behavior shows how difficult it is to find cause and effect in brain science. Then there is a discussion of the releasing stimuli for aggression followed by the hormonal effects. Perhaps the most interesting part of the lecture covers the environmental influences for moral behavior: bake it into the cerebellum or you'll have to think it through! Aha, this whole series of 4 videos is about "the biology of morality": sweet!

19. Aggression III

Metaphor in the Brain (continued)

Sapolsky continues exploring how the brain mixes reality and metaphor. These effects strongly support Jonathan Haight's findings that moral reasoning is mostly after the fact rationalization for moral affect (the affective decisions of the limbic system). The dictionary (1913 Webster) says "affective" means "pertaining to or exciting emotion" and "affect" means "to influence or move, as the feelings or passions".

Anterior cingulate is a brain region (nucleus) that both senses pain and empathically senses the pain of another. Despite tremendous interest in mirror neurons which get excited if you and someone else are doing the same motor action (but not when doing it yourself or watching others), there is no proof that mirror neurons play a role in the anterior cingulate.

Serotonin is a neurotransmitter which seems to affect aggression and impulsive behavior. In animal studies less serotonin yields more impulsive and aggressive behavior. In humans, one studies the serotonin breakdown products and finds that lower levels of such products correlate with higher levels of aggression and antisocial behavior (caveats: correlation is not causation: cause and effect may be reversed; breakdown products may come from serotonin use in other parts of the brain). By driving down serotonin levels in animal studies, there is some evidence that serotonin causes impulsive and aggressive behavior. In humans giving a drug that mimics serotonin in the short term ("buzzing serotonin pathways") leads to increased metabolism in the frontal cortex except in individuals with a history of antisocial violence. This suggests serotonin supports frontal cortex inhibitions of inappropriate limbic impulses. Sociopaths have lower than normal metabolic rates in the frontal cortex.

Serotonin breakdown biochemistry: tryptophan (TH: tryptophan hydroxylase) -> 5-HTP (5-Hydroxy-L-tryptophan; enzyme TBH is not listed at -> Serotonin (MAO: Monoamine oxidase, COMT) -> 5-HIAA (which can be measured in bloodstream, cerebospinal fluid, or urine). Measuring serotonin levels in this indirect fashion can be problematic because there could be a problem in the breakdown process itself instead of with serotonin. Sapolsky reports that some studies failed to control for these effects and reported results which are completely uninterpretable.

The rate-limiting step is the biochemical reaction that is slower or harder to accomplish than the other steps in a biochemical pathway. TH is the rate-liming step in serotonin metabolism. There are studies showing that variations in TH in humans and primates correlate with aggression (but they are very small statistical effects). There are studies on correlations with the two variations on MAO showing a gene-environment interaction: the "bad" gene doesn't lead to more aggression unless there is also abuse.

Alcohol has lots of messy effects on the brain (no specific effect). There is no significant relationship between alcohol and aggression! Alcohol causes aggressive people to be more aggressive; it causes non-aggressive people to be more inhibited. Alcohol merely magnifies the preexisting social tendency. People who believe their blood alcohol levels have risen tend to become more aggressive: alcohol socially "allows" you to become more aggressive. It has a modulating function. Anthropological studies of Polynesian cultures who were taught drinking after WWII adopted the behavior and attitude toward alcohol of the colonial power teaching them to drink (US, UK: aggression; FR: sexual promiscuity).

Releasing stimuli for aggression

Pseudomyrmex ants respond aggressively to vibrations in their Acacia tree homes as a symbiotic protection service against herbivores. Humans have no auditory, olfactory or other direct cues for aggressive behavior. Although there is amygdal activation with exposure to the sweat of frightened individuals, it is subliminal and does not necessarily lead to aggressive behavior.

Pain is the most reliable stimulus for triggering aggressive behavior. Frustration is another trigger: a rat trained to press a lever will get frustrated when the lever stops working and bite its cagemate: displacement aggression. The stressed and frustrated rat will have high glucocorticoid levels until it bites its cagemate. Displacing aggression on somebody else in species after species is stress reducing. Male baboons occasionally have rape (forced sex with a female who actively attempts to get away and resist): Sapolsky has seen it three times: when an alpha male is displaced from its #1 position. Although such displaced males will usually just mope quietly or beat up on other lower-ranking guys occasionally they displace their frustration on females by rape. Frustration displacement may account for the increase in violence typically associated with poverty and recessions. Frustration, pain, and stress are reliable predictors of aggressive behavior. But it is a modulatory factor: only those predisposed to aggression will vent it when facing pain, frustration and stress.

John B. Calhoun studied overpopulation in rats in the 50s and reported that rats in a smaller enclosure dramatically increased aggression to the point where they started killing and cannibalizing each other. Urban sociologists expected the next generation of large cities to feature cannibalism. Everyone feared the menace of overpopulated inner cities. Eventually more careful studies were conducted which determined that aggression does not increase with overcrowding. Overcrowding modulates aggressive behavior. Calhoun and other early researchers failed to quantitatively measure the aggression and simply reported the dramatic behaviors of the few hyper-aggressive rats dealing with the stress of overcrowding and having more rats around them to vent their frustrations upon.

Hormonal effects

Testosterone is required for the normal range of aggressive behavior in every species looked at. Castration reduces aggressive behavior but not to zero. More prior experience with aggression leads to higher rates of aggression in the castrated males. Add back in small levels of testosterone (10% of normal) and the aggressive behavior returns. If testosterone levels reach 10 times normal levels (anabolic steroid abusers), then aggressive behavior increases. Testosterone is necessary but not sufficient for aggressive behavior, the brain is not sensitive to small differences in levels: testosterone has a modulatory effect: it exaggerates preexisting tendencies. High testosterone shortens the lag time between action potentials (neuron firings) in the amygdala if and only if it is already excited.

The world's only research colony of hyenas living in the Berkeley Hills originated as pups brought in from East Africa without adult models. Without the social learning from seeing mom terrorizing all the males, it took longer for the females to establish their dominance.

Testosterone and other androgens are synthesized in the adrenal gland in females at about 5% male levels. Aggression in females is also modulated by adrenal androgens.

Perimenstral (a little before and after menses) hormones are associated with increased aggression in females. Since female baboons also display perimenstral aggressive behavior, the cause is evidently biological and so Sapolsky dismisses the cultural and psychodynamic explanations. Some anthropologists have argued that perimenstral effects don't occur in societies that are freer about their bodies and more sexually uninhibited. Sapolsky reports that perimenstral mood shifts are pretty universal in human cultures. But there is some psychology because women told that their period is approaching, will become more irritable (self-fulfilling prophesy). The significant others of perimenstral women also become more irritable. Perimenstral behavior also correlates with depression and social withdrawal. In baboons, low-ranking females withdraw whereas high-ranking ones get irritable. Ethology (interviewing an animal in its own language) plays its profound role again!

The ratios of estrogen and progesterone drop significantly in the perimenstral period. 10% of women have severe perimenstral syndrome correlated with a large drop in progesterone around the time of menses. Progesterone binds to GABA, the main inhibitory neurotransmitter in mammals, at the binding site for minor tranquilizers (which also binds benzodiazepines such as Valium and Librium). An interpretation is that progesterone acts as a mild sedative decreasing anxiety except for that time of the month when its levels are low. In another subset of women with perimenstral syndrome, endogenous opioid levels (specifically beta-endorphin, a neurotransmitter and hormone with "nice effects on mood") decrease significantly around menses.

Glucocorticoids, epinephrine, & norepinephrine rise with fight (aggression) or flight (not aggression): activation of the sympathetic nervous system. The opposite of love is indifference. Give someone pain or physical discomfort and their sympathetic nervous system typically activates as part of stress response. In sociopaths, the response is much reduced (elevated pain thresholds with less sympathetic responsiveness to pain). Since empathy is feeling someone else's pain, this effect may explain some sociopathic behavior.

Environmental triggers of aggression: broad theories: 3 schools

1. Environment is irrelevant: Konrad Lorenz 1966 book "On Aggression" was very influential: aggression is universal (present in all individuals) and has no environmental requirements. Famous quote: "there is no love without hate". Hydraulic model of aggression: aggressive drive builds up until smaller and smaller environmental triggers will release it. If aggression is not released, it will eventually lead to spontaneous aggression. Once aggression is released, the aggressive drive is depleted starting a refractory period. Critiques: aggression is not universal nor inevitable in humans (well, at least not after seventh grade). Aggression is not self-depleting, instead it is self-reinforcing (witness crowd violence at sporting events). Aggression stimulates more aggression: it legitimizes and habituates one to it.

2. Aggression is ultimately all about frustration, pain, stress, fear, and anxiety. Extreme version of the frustration displacement model: since the amygdala is about fear and aggression, all aggression is based on some form of fear or stress. Unemployment, spousal abuse, and child abuse are correlated. In baboons, about 50% of aggression occurs after a fight or someone loses access to a resource. The poor are more subject to stress and so are more violent. Moreover, the aggression is directed at other low-ranking organisms. Negative result: during periods of famine, aggression tends to go down despite the high stress that everyone suffers. So at least in this realm, the extreme frustration displacement theory is off. "Behavioral fat": aggression occurs when there are fewer stresses (lions during zebra migration have the surplus resources to fight for dominance). Clan or retributive violence can be interpreted as competing for the one who does the last retribution.

3. Behaviorists argue that aggressive behavior can be shaped by positive and negative reinforcements. Studies looking at laws and crime rates can test the theory. When the death penalty becomes law, the number of premeditated murders does in fact decrease. However, crimes of passion and impulsive murders are not affected by the death penalty. Generally changing the severity of punishment does not change the rate of murder.

Early environmental (developmental & upbringing) influences on aggression (and empathy, etc.) is mainly about learning the context for appropriate aggressive, empathic, compassionate, and cooperative behavior. Most of the research on this question is framed by the question: how do moral standards develop in kids? During the first few days or weeks of life babies start distinguishing between animate and inanimate things. The fusiform cortex which responds to faces does not work as well in autistic children who cannot distinguish animate from inanimate. Before the development of a self (distinguishing self from other: an ego boundary) at about one year of age, kids identify themselves with mom: they report an owwwy on their finger if mom has cut her finger.

Kids develop theory of mind (that is, the recognition that others may have different information, thoughts, and feelings than they have) between ages 3 and 5. Kids may have a good faculty for theory of mind when reading a story but when in an emotionally charged situation, they may not be effective in thinking through the differential information, thoughts, and feelings of others.

Is theory of mind a prerequisite for empathy? Is it possible to feel someone else's pain without understanding that there is a someone else who has different thoughts and feelings? The thinking is that yes theory of mind is a prerequisite for empathy. Sociopaths have an effective theory of mind faculty yet little or no empathy. So theory of mind is not a sufficient condition for empathy. But if an adult is crying, a 15 month old will often offer them their pacifier to help them feel better. Do 15 month olds have empathy without theory of mind? Perhaps, but it might also be the child is irritated by the crying and is just trying to stop it (or mimicking the parental behavior of sticking pacifiers in the mouths of upset babies). One experiment tested kids aged 3-6 months with three skits. One skit exhibited a prosocial interaction (someone helping another), another neutral behavior, and the third antisocial behavior (taking something away from someone). The kids prefer to look at the individual doing the prosocial helpful behavior. So some parts of empathy manifest before theory of mind.

Lawrence Kohlberg at Harvard built upon Swiss psychologist Jean Piaget's work of stages of cognitive development (object permanence: if you cover an object is it surprising when it gets uncovered?) to develop Kohlberg's famed stages of moral development in kids: six sequential stages falling in three categories of moral reasoning: 1) pre-conventional (you behave morally because you might get caught and either get punished or lose a reward: ego centered motivations for moral behavior), 2) conventional (social norms: shared group conventions: following the rules or laws), and post-conventional (transcendent: behaving in a way that breaks the rules or laws: there are things more important than what everyone else thinks: civil disobedience, bad laws may deserve to be broken). Studies show that not all adults advance through the stages: many stay in the pre-conventional or conventional stage into adulthood. One study looked at a girl wanting to play ball with some boys. In the pre-conventional stage they argue it is fair since she has not played. In the conventional stage, they argue it will be bad for the team, so no (instead of individual consequences using standards of norms to reason instead).

Criticisms of Kohlberg's model. Like most studies in the 60s, most of them used male subjects. Kohlberg's model fits boys better than girls. Boys are more about justice and girls are more about social affiliation (a gender critique). Moral development in other cultures doesn't seem to fit Kohlberg's model well either. People who are trained in childhood to respect ability more than effort are far more likely to use social norm moral reasoning than social justice reasoning and so are less likely to reach a post-conventional Kohlberg stage. How do kids learn that it is sometimes OK to lie? ("No, grandma, I do not like this sweater." "I already have one of those.") How do kids learn the difference between rules (less breakable: must be followed) and principles (involve judicious weighing). When do we learn about bad laws? How do we learn the difference (consistently see distinction) between intended harm and successful harm? At earlier stages kids discount harm that doesn't actually happen.

Peer groups & community

Felton Earls at Harvard made very careful studies about how many violent acts kids were exposed to growing up, He found that those growing up in a violent community (witness or victim to violence) led to adults who are more prone to violence (they rationalize and are desensitized to it). The effect is even stronger if the violence happened in the family. Many studies now show that the effects of watching a lot of violent TV or movies or video games modulates behavior: it magnifies the effect only in those who have a pre-existing tendency to violence.

In a study comparing male homicide rates by age in three cities (one Canadian: Toronto: 50; one American: Chicago: 600; and one British: London: 30), Martin Daly and Margo Wilson found that the frequency of murder by age increases rapidly until the early 20s and then starts declining significantly by age 25. It is a maxim in criminology that there is a big decrease in criminal behavior after age 25. So although age is an important predictor, the community you grow up in is the most significant predictor of a subject's likelihood to commit murder.

Harry Harlow raising captive rhesus monkeys in isolation showed that (as with sexual behavior) their fixed action patterns are normal, but they never learned the appropriate social context for the behavior. So they show the normal behaviors of dominance and subordination, but they do not know when and to whom those displays should be given. In baboons, you inherit your rank from your mother. The social training for appropriate social behavior based on hereditary rank begins as early as the first week after birth. Similarly the Berkeley Hills hyenas took longer to establish their dominance hierarchy since they were raised without social context.

How important are parents and how important is peer group? Judith Rich Harris argues in her book "The Nurture Assumption" that peer influences are vastly more important than had been recognized. She observed that in immigrant families by age 4 or 5 the kids are developing the accent of the local community not that of their parents; they are becoming embarrassed by their parents' accent; they answer questions given in the home language using the outside language. We pick up the accent of our peer group, not our parents. "At the end of the day, parents are mostly good for determining which peer groups their kids have access to." Humans are very subject to conformity and conventionalized behavior in groups: once separated into groups we start to have disparaging beliefs about the other group; we overidentify with whatever category or group we are assigned. Within a week of working on the set of Planet of the Apes, actors sat at lunch with people from the same species that they were acting in the movie. All the actor chimps ate together even though their best friends were orangutans. Harris reinterpreted the result that kids growing up in a family without a father had a higher chance of antisocial violence as an adult: such families tend to be poorer and it is the peer socialization which leads to the effect.

Steven Levitt (economist) and John Donohue (lawyer) determined that a significant part of the decrease in the crime rate since the 1970s (now thought to be 50%) is due to the Roe v. Wade decision that legalized abortion. As each state legalized abortion, there would be a 12-15 year lag before the crime rate would drop. At the beginning of the decrease in crime, it would be due to fewer teenagers entering the criminal justice system. As time goes on there are fewer criminals in the age range that grew up with safe, legal abortion. Evidently a huge predictor of growing up to be a violent antisocial individual is being an unwanted child. Abortion reduces the number of unwanted children which significantly reduces the crime rate.

Explicit declarative learning: you learn a fact, you know that you know it, and you can consciously use it, strategize with it, and use it in an executive way (requires hippocampus to store it)

Implicit procedural learning: your body knows it better than your head (stored in the cerebellum)

Is there a correlation between moral reasoning and moral behavior? The literature does not find a particularly good connection. Those who reach Kohlberg's post-conventional stage of moral reasoning tend not to be more likely to act heroically. The Carnegie Foundation studies those who do brave unexpected moral acts and finds they grow up in an environment with a very strong, consistent, repeated stated imperative to act morally, to act bravely, and to not care what other people think. When interviewed, such people invariably report that they didn't think (for example before running into the burning building or jumping into the freezing river), so it is implicit procedural learning that spurs them to action. Something overlearned in childhood that does not require reasoning skills and effort.

Josh Green determined that players who cheated at least once (in a game designed to make it easy to cheat and easy for the researches to know when they were cheating) had increased activity in their frontal cortex. Those who didn't cheat never had frontal cortex activation: it was an implicit pathway: they were not even tempted. Perhaps their baked in moral rules made the question verboten.

Sapolsky concludes that he thinks a lot of what moral development is about is developing implicit rules by overlearning in childhood.