Indifference probabilities. Subjects' behaviour was analysed using a linear regression technique (see text for method of calculation) to estimate the large-reinforcer probability at which they were indifferent between a 4-pellet uncertain large reinforcer and a 1-pellet certain small reinforcer. Rational choice, and optimal choice in this task, would be an indifference probability of 0.25 (that is, it is rational to be indifferent between a certain 1-pellet reinforcer and a 4-pellet reinforcer delivered with a probability of 0.25), shown by the dotted line. Lower indifference probabilities imply risk-prone behaviour; higher indifference probabilities imply risk-averse behaviour. Preoperative and successive postoperative indifference probabilities are shown for AcbC-lesioned and sham-operated control rats (** p < .01, difference from controls).