Detecting correlation changes in electrophysiological data

A correlation multi-variate analysis of variance (MANOVA) test to statistically analyze changing patterns of multi-electrode array (MEA) electrophysiology data is developed. The approach enables us not only to detect significant mean changes, but also significant correlation changes in response to external stimuli. Furthermore, a method to single out hot-spot variables in the MEA data both for the mean and correlation is provided. Our methods have been validated using both simulated spike data and recordings from sheep inferotemporal cortex.