Two channel segmentation of EEG
Isak Gath, Ariel Michaeli · 2003
Dual-channel adaptive segmentation of the electroencephalogram (EEG) signal is proposed. The purpose is to divide the signals into quasi-stationary segments, on the basis of information common for the two channels. The segment boundary is determined by changes in the cross spectrum of the two signals. It has been shown theoretically, as well as by simulation studies, that this method is sensitive to changes common to both channels, whereas segmentation does not occur as a result of changes in each channel separately. >