EEG Background Activity in Childhood Epilepsy: Valid Parameterisation by the Partial Autocorrelation Function¹

Abstract

The relationship between EEG background activity and the clinical course is studied in a group of 42 patients with a primary myoclonic-astatic epilepsy. According to the occurrence of seizures the probands are divided into two groups. One group is comprised of patients with an unfavourable course of the disease and the other group is comprised of probands with a favourable one. It is shown by a combined approach of variance analysis, linear discriminant analysis and factor analysis that the EEG background activity parameterized by the partial autocorrelation function is significantly different in the two clinically defined groups. The error rate for assigning a proband, only from the proband's parameterized EEG, to the correct clinical class is estimated to be 0.24.
A factor analysis of EEG parameters yields a discriminant factor interpreted as rhythm-power-factor. The group with an unfavourable course is characterized by rhythms in the theta-band and a greater total power than in the favourable group which shows rhythms in the alpha-band. Hypersynchronous potentials in the EEG of this study are not related to the clinical course of the epilepsy. This study confirms that the EEG background activity contains important information about the convulsibility in primary myoclonic-astatic epilepsy.

Key words

Myoclonic-astatic epilepsy - EEG background activity - Time series analysis - Partial autocorrelation function - Pattern classification