Separation of Phase-Locked and Non-Phase-Locked Activities Using Empirical Null-Distributions Obtained from Randomized Plus/Minus Averagings of Event-Related Epochs

This paper proposes a novel non- parametric statistics-based method for extracting phase-locked target source activities from non-phase- locked background activities in MEG event-related measurements. The method derives, at each pixel location and at each time point, an empirical proba- bility distribution of non-phase-locked activities from plus/minus averagings of raw epochs. The statisti- cal threshold that extracts the target source activ- ities is derived based on the empirical distributions obtained at all spatio-temporal locations. Here, the multiple comparison problems are taken into account by using the maximum statistics, i.e., by using two step procedure: standardizing these empirical distri- butions and deriving an empirical distribution of the maximum pseudo T values. The results of numer- ical experiments are presented to demonstrate the method's effectiveness.