May 1, `02: Simon Schultz, CNS
<br>
<br>
Title: Studying the information content of spike trains
<br>
<br>
In recent years I and a number of colleagues have pushed an approach for 
studying spike train structure which involves breaking the Shannon 
information content of the spike train down into components according to 
the order of correlation present. This has made it feasible to examine 
temporal, correlational and rate coding mechanisms within a single 
mathematical framework. In this talk I will give an overview of our 
approach, and discuss how it relates to other similar attempts in the 
literature, dating back to the reconstruction filters used by Bialek and 
colleagues over 10 years ago. After explaining some of the practical 
issues relating to implementation of our algorithm, I will describe the 
results we have obtained by applying this approach to experimental data 
from the rat barrel cortex, from primate visual cortex, and to a model 
of the thalamocortical relay system.
<br>
<!-- <br>
see <a href="http://www.cns.nyu.edu/~schultz/papers.html">simon's
page</a> for details.
<br> -->
<br>
<a href="course.html">back</a> to CNS seminar page