A Computational Role for Arousal in Optimal Inference
Degree type
Graduate group
Discipline
Subject
Computational modeling
Decision making
Inference
Learning
Locus coeruleus
Neuroscience and Neurobiology
Funder
Grant number
License
Copyright date
Distributor
Related resources
Author
Contributor
Abstract
Making accurate predictions is one of the most critical functions of the brain. Whether made by a monkey deciding where to forage, a deer deciding which way to run, or a wall-street broker deciding how to invest, decisions are informed by expectations about possible future outcomes. These expectations are learned over time through experience and are rapidly adjusted when they fail to match observations. Here I propose and support the thesis that learning systems in the brain optimize the accuracy of predictions in a changing world, even though this necessitates becoming insensitive to incoming sensory information under some conditions. Furthermore I propose a biologically inspired model for achieving accurate predictions and suggest a novel role for the arousal system in optimally adjusting the influence of incoming sensory information. I support these theses with a series of experiments that utilize computational modeling, as well as behavioral and pupillometric measurements in humans.