A New Method of Electroencephalographic Analysis for the Determination of Stable and Unstable Processes

Abstract

Long-term EEG analysis is a highly specialized method in the investigation of physiological and disturbed sleep. At present the tendency toward the computer-assisted evaluation of EEGs is gaining ground at an international level. The problem with EEG analysis seems to be the diversity of the clinical and scientific demands on the EEG. So, in order to be able to assess the numerous and different specific graphic elements and EEG patterns, continued visual evaluation of the daytime and sleep EEG is necessary. For the objective assessment of sleeping behavior, however, automatic EEG evaluation is imperative. This consists primarily of quantitative EEG analysis using spectral-analytical evaluation methods [1, 3. 6, 10, 13] or modified methods based on image pattern analysis [5, 7, 8, 12, 14]. An example of the former is EEG mapping [9, 11]. The latter method is currently mainly used for the analysis of specific sleep patterns and of activity in the delta band of the sleep EEG [4, 5, 8]. Despite advanced computer-asisted EEG evaluation, too little importance is still attached to the cyclic structure of sleep which is reflected in the EEG. Brain function, just as all the other bodily functions (immunological, endocrine, exocrine, etc.) is subjected to a great number of rhythmic and stochastic variations in the 10^-3–107-s range, caused by internal and external factors [2]. The internal factors include all the genetically determined regulatory processes within the body, and the external factors are the organism-to-environment relationships.