Friday, November 15, 2013
Behavioral Social Neuroscience Seminar
A Neuro-Computational Account of Economic Choices
Philippe Domenech, MD. PhD., Laboratoire de Neurosciences Cognitives, Ecole Normale Supérieure, Paris, France
When ordering a dessert from a set menu, choosing between the apple pie and the strawberry tart solely depends on the subjective values attributed to each sweet: there is no correct answer, only options that are more or less desirable according to one's subjective preferences. An important assumption in economy is that value-based decision-making entails two separable stages: (1) the aggregation of all options' meaningful features into subjective values ordered on a common scale (valuation) and (2) the selection of the option associated with the most valuable outcome (selection). However, most models of economic choices implicitly assume that the selection stage trivially reduces to an instantaneous maximization. As a result, most neurobiological studies on economic choices have focused on characterizing the brain valuation system. Although this scientific endeavor yielded strong evidence linking orbito-frontal cortex and valuation, we still know very little on the algorithm at play at the selection stage and its implementation in the human brain. I will present new behavioral and fMRI data showing that valuation and selection involve distinct brain networks, and show how drift-diffusion key computations can be assigned to separate functional brain regions by crossing model-based fMRI, multivoxel pattern analysis, and DCM.