"Interval Timing and Decision Making with an Opponent Poisson Diffusion Model" is the title of the talk to be presented by Oberlin College Assistant Professor of Neuroscience Patrick Simen.

Abstract:  Diffusion models explain patterns of human choice probabilities and response times in perceptual decision tasks (e.g., tasks in which a person decides which of two objects is heavier, or which of two sounds is louder).  In diffusion models of human behavior, a random quantity representing "evidence" builds up over time in a person's brain until a boundary is reached, yielding a decision. 

Recently, my colleagues and I have applied diffusion models to a different task:  interval timing.  Despite major differences between these task types, human response time distributions in both are typically time scale invariant—their standard deviations divided by their means are constant, even as the means differ across conditions.  To understand how diffusion could explain time scale invariance, we analyzed diffusion processes that emerge when excitatory and inhibitory Poisson "spikes" (representing action potentials in the brain) are accumulated over time.  

Assuming that spike rates are proportional to evidence strength (in decision making) or to clock speed (in timing) leads to the prediction of time scale invariant response times.  Most predictions held up, suggesting that human decision making and timing may derive from a common process of diffusion in the brain.

A reception for Professor Simen begins at 4:10 p.m. in the Anderson Lounge, Wright Lab, second floor.