AbstractRecent interest has focused on a class of decision problems in which subjects encounter options serially and must decide when to leave an option in search of a better one, rather than directly comparing simultaneously presented options. Although such problems have a rich history in animal foraging and economics, relatively little is known about their neural substrates. Suggestively, however, a separate literature has argued that the key decision variable in these tasks – the opportunity cost of time, given by the average reward rate – may also govern behavioral vigor and may be reported by tonic dopamine (DA).In this study, we test whether this putative dopaminergic opportunity cost signal plays an analogous role in serial decisions by examining the behavior of patients with Parkinson’s disease (PD), on and off their DA replacement medication, in a patch-foraging task. In these tasks, subjects’ decisions about when to leave a depleting resource implicitly reflect their beliefs about the opportunity cost of time spent harvesting that resource. Consistent with the opportunity cost hypothesis, umedicated patients harvested longer than matched controls, and medication remediated this deficit. These effects were not explained by motor perseveration. Our results suggest a functional role for DA, and an associated cognitive deficit in PD, in a type of decision process that may be distinct from (but related to) the neuromodulator’s well studied roles in behavioral invigoration and learning from rewards.Significance StatementThis study addresses two important questions whose answers are, unexpectedly, linked. First, what is the scope of cognitive functions of the neuromodulator dopamine, whose contributions – for instance, as assessed by both the motoric and more subtle cognitive deficits of patients with PD, which depletes dopamine – range from movement to reward and decision-making? Second, what are the neural mechanisms supporting an important but understudied class of problems, in which, rather than choose among a set of alternatives (like apples and oranges), one makes serial decisions about whether to stick with an option (like a job, or a mate) or seek another? We demonstrate a novel cognitive deficit in PD that integrates this function into the web of DA’s contributions.
Why Don't Present-Biased Agents Make Commitments?
