The above results indicate that statistically correct delimination of significant time-frequency regions of ERD/ERS is possible over an arbitrary time and frequency ranges (e.g. no high-pass filtering or a priori choice of the frequency band of interest are needed). Careful choice of the statistics and correction for multiplicity retains the power of the procedure also in cases of a low number of repetitions (57) or an artificially increased size of the problem (calculated for an area where over 80% of the time was expected to exhibit no ERD/ERS).
In estimating the time-frequency energy density, MP offers better sensitivity and higher resolution than STFT. However, both these methods give similar and consistent results in the context of delimination of time-frequency regions of significant changes of energy. Taking into account the high computational cost of the MP procedure, we may consider the STFT estimator as an alternative for cases when speed is more important than sensitivity.
Presented methodology and software gives the missing criterion for the significance of event-related changes in the time-frequency energy density of signals. It introduces objective criteria to related neurophysiological research and allows to investigate quantitatively e.g. the following issues: