The spectral dynamics and its applications in EEG

The method of spectral dynamics is introduced to process slow changes in EEG. This method is based on evaluation of distance between EEG spectra. The typical application field is pharmaco-EEG after single dose drug administration, where distances between pre and post drug EEG spectra are computed. The distance between two spectra may be, however, defined in several different ways. The paper deals with the well known Lp metrics and with the metrics alpha that plays an important role in discriminating stationary processes with different spectra. The Lp and alpha metrics are non-equivalent and their properties, consistency and robustness, are studied in a simple statistical model.