Generalized Linear Models for Point Process Analyses of Neural Spiking Activity

Abstract

Most systems neuroscience experiments entail applying a stimulus and measuring the response which is commonly the spiking activity of one or more neurons (Brown et al. 2004; Paninski 2004; Truccolo et al. 2005; see entries “Spike Train,” “Spike Train Analysis: Overview”). The typical objective of the experiment is to define how spiking activity in a given brain area changes in response to the stimulus under a range of different conditions. From a statistical standpoint, this type of experimental design requires a regression model in which the observations can be spiking activity and the regressors can be the stimulus and any other relevant covariates. The generalized... Keywords Generalize Linear Model Point Process Spike Train Exponential Family Interspike Interval