The standard studies focus on the magnitude question, but I’m wondering if there are studies on the question of the type of reward. Suppose a subject is indifferent between an infusion of glucose of a certain size, and finding out that she has won a certain amount of money. The specific cues predict the glucose, but instead she finds that she is to be paid the money. Indifference means that there’s no need for a reward prediction error concerning magnitude. But it’s surely surprising, relevant to learning, etc., that she got cash rather than glucose…