A generalized affine model for analysis of neural time series data.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


The Generalized Affine Model (GAM)

The GAM is a latent variable model developed to study large populations of simultaneously recorded neurons. It incorporates latent variables that can either add to or multiply the stimulus response of individual neurons. This model is a generalization of several other models that have appeared in the computational neuroscience literature, including:

A more thorough explanation of the modeling framework can be found in our preprint on biorxiv: Characterizing the nonlinear structure of shared variability in cortical neuron populations using latent variable models

The doc directory contains scripts that show how to use the model on several simulated datasets (coming soon).

The GAM optimizes model parameters using Mark Schmidt's minFunc package, which is located in the lib directory and should work out of the box. If not you may need to run the mexAll.m file from the lib/minFunc_2012 directory.