Possibilities of classification of topographically distributed neurophysiological multi-channel data

Author(s): Wolter, S., Klee, B., Schontube, E.

Journal/Book: Int J Clin Monit Comput. 1996; 13: Spuiboulevard 50, PO Box 17, 3300 AA Dordrecht, Netherlands. Kluwer Academic Publ. 27-34.

Abstract: Progress in quantifying states of cerebral function and in the further development of automated EEG processing demands the application of suitable methods for the reduction of neurophysiological multi-channel data as well as their automatic classification. The method used here for reducing multi-channel data was to gain distributions of parametric descriptors from EEG data from computer-aided topographic electroencephalometry (CATEEM(R)), for example the relative and absolute band power in the frequency bands delta, theta, alpha 1, alpha 2, beta 1, beta 2, total power, median and mode frequency, and other parameters. These values were subjected to cluster analysis. The classification of EEG parameters was carried out by means of discrimination analysis and neural networks. The practicability of both procedures was demonstrated in the reduction and classification of EEG data in the context of a normed study involving 104 healthy adults. These data have been used as the basis for a new evaluation study of 60 additional intraoperative EEG recordings obtained with CATEEM(R). In that newly started study, the effects of sedative and anaesthetic drugs on EEG behavior and psychophysiologic behavior remain to be investigated.

Note: Article L Rolz, Humboldt Univ Berlin Charite, Dept Med, Anesthesiol & Intens Care Clin, Schumannstr 20-21, D-10117 Berlin, Germany

Keyword(s): classification; computer; electroencephalography; EEG norm variations; multi-variate approach; discrimination analysis; neural networks

Zurück | Weiter