-----------------------------------------------------------

Special Applied Math Seminar

Monday, Nov 16, 2009, 4:00 PM
Cullimore Lecture Hall, Room 611
New Jersey Institute of Technology

Spike-Time Reliability of Layered Neural Oscillator Networks

Kevin Lin

Department of Mathematics, University of Arizona

Abstract

This talk concerns some recent results on the reliability of large networks of coupled neural oscillators driven by fluctuating stimuli. Reliability means that upon repeated presentations of a given stimulus, the network gives essentially the same response each time; whether a network is reliable can impact its ability to encode information via the precise timing of spikes. I will explain how questions about reliability can be formulated and studied within the framework of random dynamical systems (RDS) theory. Focusing on certain layered network models, I will explain -- via a combination of qualitative theory and numerical simulations -- how factors like network architecture affect reliability. I will also discuss the effects of noise, and show that some types of noise affect reliability more seriously than others.

This is joint work with Eric Shea-Brown and Lai-Sang Young. No prior knowledge of RDS theory will be assumed.