Gaussian-process factor analysis (GPFA)

GPFA extracts low-d latent trajectories from noisy, high-d time series data. It combines linear dimensionality reduction (factor analysis) with Gaussian-process temporal smoothing in a unified probabilistic framework. GPFA is particularly useful for exploratory analysis of spike trains recorded simultaneously from multiple neurons on individual experimental trials.

Matlab codepack (version 2.02, 108 KB)
References: Yu et al., J Neurophysiol, 2009; Churchland et al., Nat Neurosci, 2010