APPLICATION OF NEURAL NETWORKS FOR EEG-ANALYSIS - CONSIDERATIONS AND FIRST RESULTS

This paper presents the results of the practical use of artificial neu ral networks in the field of EEG analysis. It describes the general me thodology of application as well as a case study of a discrimination o f depressive and psychotic patients using 16-channel long-term EEG dat a prepared by classical pre-processing (spectral decomposition). This study shows advantages and current limits concerning different levels of generalisation capabilities using a representative application exam ple.