Participants repeatedly choose between options (arms of a bandit) with different and unknown probabilities of reward. This task models exploration versus exploitation trade-offs.