The ability to anticipate future outcomes is crucial to decision making, and yet people exhibit a variety of striking biases in their predictions (for example, we tend to be overconfident, and prone to believe we will have more control over events than we actually do). On the one hand, these biases can lead to suboptimal decisions, be they in financial investment or mate selection. On the other hand, they appear to serve an adaptive function, improving subjective well-being and motivating action.
One pervasive bias in prediction is that of unrealistic optimism. Independent of culture, age, or gender, people tend to overestimate the likelihood of positive future outcomes and underestimate the likelihood of negative future outcomes. Over the last five years we have begun to investigate its neurobiological basis (Sharot, Riccardi, Raio & Phelps, 2007; Sharot, Korn & Dolan, 2011). Our most recent work addresses a fundamental puzzle posed by the existence of optimism: how can unrealistically positive expectations be maintained when information challenging those predictions is readily available? According to standard learning theories, an agent should correct its expectations when confronted with disconfirming information, eventually leading to unbiased predictions. We have found, however, that people update their beliefs selectively: they incorporate information that enforces optimism, but dismiss information that contradicts it (Sharot, Korn & Dolan 2011). This selectivity appears to be mediated by a relative failure of frontal regions to code for errors that call for a negative update of predictions (i.e. that would otherwise reduce optimism). We are currently exploring cognitive and biological ways to enhance or reduce the optimism bias.