Fri, Sept 02
Speaker: Iris Vilares
Title: Prior and likelihood uncertainty are deferentially represented in the human brain.
Abstract: Uncertainty shapes our perceptions of the world and the decisions we make. Two aspects of uncertainty are commonly distinguished: uncertainty in previously acquired knowledge (prior) and uncertainty in current sensory information (likelihood). Humans take these types of uncertainty into account for successful behavior. Although recent theories make predictions about the neural representations of uncertainty, direct experimental support is lacking. Here we varied prior and likelihood uncertainty in a perceptual decision-making task while measuring human brain activity using functional magnetic resonance imaging (fMRI). Representations of prior uncertainty were identified in reward-related areas including putamen, amygdala, insula, and orbitofrontal cortex. In contrast, likelihood uncertainty was associated with brain activity in visual occipital areas. Our results indicate that the human brain uses different networks to encode uncertainty about prior and likelihood, offering insights into the neural mechanisms that allow humans to make decisions close to the optimal defined by a Bayesian statistical framework.
Advisor: Dr. Kording