Patrick Simen - Oberlin College
Francois Rivest - Royal Military College of Canada
Wednesday, July 8, 2015
1:00 PM, 215 Tolentine
"Keeping Track of Time With A Drift-Diffusion Model"
Drift-diffusion models (DDMs) are widely used to model the neural and psychological basis of two-choice decision making. Until recently, they have not been used to model how humans and other animals keep track of time, raising the question of why a simple diffusion process is good enough for one capability but not for the other. We show that it is theoretically suitable for both. We outline how a time-adaptive DDM (TDDM) can account for the statistical regularities observed in timing behavior, as well as how it can generate new predictions about the speed of learning new durations. We further show how a simplified, Poisson model of neural spiking activity underlying the diffusion process can connect the abstract TDDM to physiology. In this "opponent Poisson" version of the TDDM (TopDDM), the balance of excitation and inhibition determines the precision of timing, and offers the possibility of studying the brain's balance of these quantities with behavioral measures. Current work focuses on extending the TopDDM modeling approach to the simultaneous timing of multiple intervals, an effort which will be informed by data collected at Villanova by our collaborator, Matt Matell.