A real-time system for the spectral analysis of the EEG

We have developed an inexpensive and portable system for processing multiple channels of electroencephalograms (EEG) in real-time to assist the electroencephalographer in identifying subtle changes in these data, particularly in the interpretation of long records. Up to 16 channels of data are spectrally decomposed with a selection of bandwidth, windows, scaling methods, epoch averaging, and smoothing options available. The resulting power spectral estimate may be displayed or printed in a variety of formats, including color encoding of selectable spectral bands. Both the sampled EEG and its spectra may be stored for off-line reprocessing, for archiving, or for statistical analysis in the time or frequency domains.