|
Activity Number:
|
206
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #306760 |
|
Title:
|
Intensity Estimates for Spike Train Data Observed under Multiple Behavioral States
|
|
Author(s):
|
Matt Gregas*+
|
|
Companies:
|
Harvard School of Public Health
|
|
Address:
|
Harvard School of Public Health, Boston, MA, 02115,
|
|
Keywords:
|
intensity function ; spike trains ; Poisson process
|
|
Abstract:
|
Spike trains are the fundamental experimental unit of many neurophysiological experiments. Intensity function estimates are usually obtained from a collection of spike trains that are modeled as independent realizations of a Poisson process with common intensity function. In practice the intensity function often depends upon the behavioral state of the subject. Thus the intensity function that generates the spike trains may be a concatinization of several separate intensity functions that are joined at behavioral markers. That is, the intensity function is dependent upon the behavioral epoch of the experiment. The lengths of these epochs vary across the experimental replications. We propose a local likelihood methodology that allow us to estimate the intensity function using all available spike train information while correctly accounting for the different lengths.
|
