On the surprising finding that expected utility is literally computed in the brain

Advocates of neuroeconomics sometimes argue that one of the most surprising findings in neuroeconomic studies is that expected utilities are literally computed in the brain. This claim is scrutinized closely in the paper. Not surprisingly, the tenability of the claim is shown to depend critically on what is meant by 'literal computation' and 'surprising'. It is argued that the findings do not show that expected utilities are literally computed, if by 'literal computation' we mean a particular kind of mental activity (namely, deliberate choice). It is also argued that the findings do not lend empirical support to expected utility theory, if that theory is taken to be a theory about individual and aggregate behavior. Findings that in some particular tasks expected utility theory predicts neural activity fairly accurately do not imply that expected utility theory also yields accurate predictions of behavior at 'higher levels', individual behavior and aggregate behavior. What the findings do suggest, it is argued, is that expected utility theory is useful for understanding decision-making at the neural level. Once one accepts that expected utility theory provides a potentially fruitful tool for predicting neural activity in neurobiology, the findings cease to be surprising.