Analysis of EEG transients by means of matching pursuit
- 225 Downloads
Matching pursuit (MP), a new technique of time-frequency signal analysis, was applied to simulated signals and the awake and sleep EEG. With the MP algorithm, waveforms from a very large class of functions were fitted to the local signal structures in a recursive procedure. By means of this technique, sleep spindles were localized in the time-frequency plane with high precision, and their intensities and time spans were found. The MP technique makes following the temporal evolution of transients and their propagation in brains possible. It opens up new possibilities in EEG research providing a means of investigation of dynamic processes in brains in a much finer time-frequency scale than any other method available at present.
KeywordsMatching Pursuit Wavelet analysis EEG Sleep spindles
Unable to display preview. Download preview PDF.
- 1.Davis, G., S. Mallat, and Z. Zhang. Adaptive time-frequency decomposition with matching pursuit. In: Wavelets Theory, Algorithms and Applications, edited by C. Chui, L. Montefusco, and L. Pucio. Boston: Academic Press, 1994, in press.Google Scholar
- 2.Delapierre, G., E. Dreano, D. Samson-Doffus, J. Senant, J. F. Menard, and D. De Brucq. Mise au point chez le nourisson d'un critere de detection automatique des fuseaux de sommeil.Rev. EEG Neurophysiol. Clin. 16:311–316, 1986.Google Scholar