Planning in view of future needs: a bayesian model of anticipated motivation

Traditional neuroeconomic theories of decision-making assume that utilities are based on intrinsic values of outcomes and that those values depend on how salient are outcomes in relation to the current motivational state. The fact that humans, and possibly also other animals, are able to plan in view of future motivations is not accounted by this view. So far, it is not clear which are the structures and the computational mechanisms employed by the brain during these processes. In this article, we present a Bayesian computational model that describes how the brain considers future motivations and assigns value to outcomes in relation to this information. We compare our model of anticipated motivation with a model that implements the standard perspective in decision-making and assigns value only based on the animal's current motivations. The results of our simulations indicate an advantage of the model of anticipated motivation in volatile environments. Finally we connect our computational proposal to animal and human studies on prospection and foresight abilities and to neurophysiological investigations on their neural underpinnings.