Stereotyping is a form of reputation, one that's applied to groups rather than individuals. It's something people think they know about each other in advance of meeting them, and something that affects how they treat them when they do.
Much of this assessment goes on unconsciously, and even against our better judgement. Last year, psychologist Elizabeth Phelps and her colleagues showed that implicit race bias, measured by associations between words and faces, was the best predictor of how much people trusted those of different races - better than more explicit traits, such as political affiliation. (You can try tests of implicit association for yourself. They'll mess with your mind.)
Now, in a study published today, Phelps' team reveals that prejudging people based on what you know about their past and prejudging them based on the colour of their skin activate the same areas of the brain.
Games of trust
The researchers got 40 subjects with an average age of just over 20 to play a 'trust game' via a computer screen. Player 1 is given some money and has to decide how much of it to share with player 2. The experimenters quadrupled this, so that player 2 received $4 for every dollar that player 1 parted with. Then player 2 had to decide how much, if any, of the windfall to return to player 1.
The 'rational' thing to do is to never share and never give anything back, but most people do both - because they know that if two people can trust one another, they will be able to work together to their mutual benefit. The advantages of trust are encoded in our norms and emotions, our sense of what's fair and ethical.
In the new study, 16 of the subjects were white, three were black, 13 were Asian, and the rest belonged to other races. Each saw a still picture of player 2, and had to decide how much of $10 to split with him (it was always a 'him').
Subjects played 150 rapid, one-shot rounds with men of all races (although the paper analyses only how players responded to black and white faces). Meanwhile, his or her brain was scanned using fMRI.
There was no average bias in who subjects trusted. But 22 of the 40 showed some bias: 14 were pro-white, and 8 pro-black.
Bias in the brain
Two brain areas turned out to be especially active in responding to race and using that information in decision-making: the amygdala and the striatum.
The subjects' amygdalas were more active when they interacted with a black player. The amygdala is part of the brain's emotional circuitry. It's particularly (although not exclusively) associated with negative emotions such as fear. Amygdala activity, though, did not correlate with individual bias - that is, people who gave relatively less money to black people did not show more amygdala activity than those who were unbiased.
The striatum, on the other hand, became more active when a player decided to trust someone of a race he or she was biased against - when the decision was perceived as risky. This brain region has been shown to get involved in executive function, responding to and predicting rewarding and aversive experiences, and making decisions in novel environments.
The finding builds a bridge to another (pdf) by the same team from a few years ago, which showed that the striatum responds similarly to individual reputation. When subjects in that study met a trust-game partner about whom they had heard bad things, their striatum lit up more than it did when they encountered someone they had heard was good - just as it does when people meet someone of a race they are biased against.
There's probably no 'reputation' bit of the brain, in other words, because to the brain, reputation doesn't exist independent of other forms of social information. And there's probably no 'race' bit, either, for the same reason.
Instead, direct experience, hearsay, prejudice, emotion and rationality all go into the neural pot, and all inform our behavior.
All that information and activity are integrated to produce an estimate of how risky it would be to trust a particular person, what the potential rewards of doing so might be - and, ultimately, a decision. This is a gnarly paper, and, to me, one of the main lessons is that it's all very complicated.
Phelps and her colleagues suggest that the amygdala might be involved in providing an "initial, automatic evalutation", which then feeds, along with a bunch of other things, conscious and otherwise, into the striatum's decision-making apparatus. But, they say, this is "solely a hypothetical framework for future investigation".
