The Neural Substrate of Trust and Reputation Management

Chris and Uta Frith
University College London
When Leo Kanner first diagnosed a group of 11 children as autistic in 1943, he described the syndrome as one of "extreme aloneness." ("Aut" is greek for "self," and autism translates as "the state of being unto one's self.") The syndrome afflicts 1 in every 160 individuals, and it leaves them emotionally isolated, incapable of engaging in many of the social interactions that most of us take for granted. An autistic individual can be highly intelligent (as measured with IQ tests), but will still have profound difficulties dealing with other people. We know that autism is a biological disorder with a strong genetic component and a basis in the brain. It's much harder, however, to figure out what happens in the brain during social interactions. After all, it's not easy to be social inside a brain scanner. You're stuck inside a small tube, unable to do anything but press a few buttons. Scientists have solved this problem by playing simple games in which two players interact while their brains are being scanned. Chiu et al. employed this technique in a recent study, which was the first to examine brain activity in autistic individuals while they were engaged in social interactions. Just Trust Me The subjects were playing a "trust" game involving money. At the start of each round, an "investor" receives a stake of $20. He or she can keep it all or "invest" some of it with a "trustee." Any money that gets invested is tripled, and the trustee then has the option of returning some portion of that amount back to the investor. For example, if the investor keeps $10 and invests $10, then the trustee has $30 to divide ($10 x 3). If the trustee decides to keep all the money, however, the investor cannot do anything about it. In principle, investors can greatly increase how much money they have by investing a large proportion of their money. Such transactions, however, require trust. There is always the risk that your partner will take the money and run, leaving you with nothing. To effectively play the game, investors must engage in a very specific aspect of social interaction: mind reading. Before you can trust somebody, it's important to understand what he or she is thinking. Read Montague's lab has studied the neural basis of this game in great detail. He has shown that the game, when played by non-autistic individuals, triggers a characteristic pattern of activity in the cingulate cortex. This strip of cortex in the middle of the brain runs all the way from the front to the back. When players are deciding how much to invest (known as the "self phase" of the game), the middle of this strip becomes highly active. No such activity appears when players learn how much their partner is going to pay in return, however (known as the "other phase"). Instead, there is activity in the front and the back of the cingulate. What's interesting about these different patterns of activity is that they only occur when people play the game with another human being. When subjects play against a computer, the brain reacts quite differently. This difference suggests that the neural activity in the cingulate has a specific role in social interactions, and is not simply computing financial gains and losses. What can the trust game experiment reveal about the autistic brain? One possibility is that autistic individuals would treat people like computers. They would fail to exhibit the characteristic patterns of brain activity that define social interactions. The results turned out to be more subtle. The autistic subjects showed normal activity during the "other phase" (when they were learning how much their partner was going to repay them), but they did not show the activation in the mid-cingulate during the "self phase" (when they were deciding how much to invest). Why was there only a difference during one phase of the game? Imagine yourself at the moment you make an investment. This instance is your only chance to influence the other player. You are trying to read and manipulate his or her mind. At the same time, you are trying to build a reputation as a person who can be trusted. Worrying about our Reputation Our speculation is that this process of reputation management is impaired in autistic individuals, because it depends on the ability to read the minds of others. This hypothesis can be tested experimentally. If we are concerned with our reputation then our behavior will be strongly affected by whether or not an audience is present to observe our actions. Consider, for instance, another sharing game known as the dictator game. One player is given $100 and is allowed to share any amount he or she chooses with the other player. In this situation, the rational thing to do would be to give the other player no money at all, because the second player is powerless to respond. Even "dictators" will typically dole out a small proportion of the money, however. When there is an audience for the transaction, dictators give away even more money. Presumably, they do not want to have a reputation for meanness or for acting unfairly. If autistic people are not concerned with their own reputation, then their behaviour should not be affected by the presence of an audience. The study by Chiu et al. is an exciting advance because it shows that brain imaging can be used to reveal the neural basis of subtle and hitherto invisible social processes that we engage in every day. It also increases our understanding of the specific social problems in autism. -- Edited by Mind Matters at 03/11/2008 7:03 AM