Notes on Generalized Linear Models of Neurons
It offers a tractable framework for analyzing neural behavior, but is incremental as it builds on existing GLM literature without introducing new methods.
The paper summarizes the application of generalized linear models (GLMs) to neural spike train data, providing equations and extensions for modeling spatio-temporal receptive fields and network activity in multiple neurons.
Experimental neuroscience increasingly requires tractable models for analyzing and predicting the behavior of neurons and networks. The generalized linear model (GLM) is an increasingly popular statistical framework for analyzing neural data that is flexible, exhibits rich dynamic behavior and is computationally tractable (Paninski, 2004; Pillow et al., 2008; Truccolo et al., 2005). What follows is a brief summary of the primary equations governing the application of GLM's to spike trains with a few sentences linking this work to the larger statistical literature. Latter sections include extensions of a basic GLM to model spatio-temporal receptive fields as well as network activity in an arbitrary numbers of neurons.