We have developed an adaptation of multi-variate analysis of variance (MANOVA) to analyze statistically both local and global patterns of multi-electrode array (MEA) electrophysiology data where the activities of many (typically >100) neurons have been recorded simultaneously. Whereas simple application of standard MANOVA techniques prohibits extraction of useful information in this kind of data, our new approach, MEANOVA (=MEA + MANOVA), allows a more useful and powerful approach to analyze such complex neurophysiological data. The MEANOVA test enables the detection of the "hot-spots" in the MEA data and has been validated using recordings from the rat olfactory bulb. To further validate the power of this approach, we have also applied the MEANOVA test to data obtained from a simple computational network model. This MEANOVA software and other useful statistical methods for MEA data can be downloaded from http://www.sussex.ac.uk/Users/pmh20.