Spectral estimation of short segments of EEG signals using least squares waveshaping filters

Electroencephalogram (EEG), the manifestations of brain's electrical activity as recorded on the scalp, has become an indispensable tool in clinical neurophysiology and related fields. The main objective of EEG signal analysis is to extract valid in formation from EEG-records and for this purpose many analysis techniques have been used. As far as background activity is concerned, spectral analysis is important to determine the rhythms present. Often short segments of data is to be analysed in order to adequately characterise the temporally rapid changes that occur in EEG. This paper is concerned with this aspect of EEG analysis viz., spectral estimation of short segment EEG data. Autore gressive (AR) spectral estimation techniques are known to provide better resolution than classical periodogram methods when short segments of data are selected for analysis. We suggest in this paper a method which works better than the normal AR spectral estimation for very short segments of data. It has been observed that the energy in the EEG data segment is concentrated not in the beginning but somewhere in between the initial and the final positions thus confirming the fact that EEG is a mixed delay signal. This position where the energy is concentrated has been obtained with the help of least squares waveshaping filter. It is also shown that the knowledge of the position where the energy in the signal is concentrated can be used in making a better spectral estirna tion of short segments of EEG data. The study has been made first with the simulated data and then with the actual EEG data. It is also observed that the performance of the present method and AR method become comparable as the length of data segment increases.