Statistical handling of EEG-data.

Abstract

This contribution focusses on the treatment of EEG data as time series. More emphasis is placed on spontaneous activity than on event-related potentials. Smoothing techniques are discussed, both for the nonparametric estimation of regression functions and of densities. Some new promising methods are presented. The fundamentals of time series analysis, and in particular of spectrum analysis, are reviewed. For the difficult area of parameterization of EEG time series, we compare desirable characteristica of parameters with the methods already in use.