It has been suggested that many aspects of reward-guided behaviour can be understood within the framework of a computational account of decision making. The account emphasises representation of expectations about decision outcomes and the revision of future expectations in the light of the prediction error-the discrepancy between the actual outcome and prior expectation. Frontal cortex and striatum are implicated in such processes in humans, monkeys, and rats suggesting they are ubiquitous and found in many species. Disagreement remains over the exact contribution made by each brain region. A growing body of work even suggests analogous processes may account for behaviour outside the domain of reward-guided decision making, for example, when people and animals learn about visual and social environments.