Title: Statistical Analysis of Neural Spike Train Data

Abstract: There are two basic problems in the analysis of neural
data. The "encoding" problem concerns how information is processed in
neural spike trains; On the other hand , the "decoding" problem
concerns how much information is in a spike train: in particular, how
well can we estimate the stimulus that give rise to the spike train?
Statistical models which builds a relationship between the stimulus
and the instantiation of spike trains can provide a unified solution
to these two coding problems in some cases. In this talk I will review
some single neuron models, in particular, the integrate-and-fire model
which is very simple but biophisically plausible. I will also describe
recent work on various techniques for detecting functional connections
between neurons and the nature of those connections.  A multistate
special event model will be proposed and the properties of the model
will be discussed.