New methods of time series analysis of non-stationary EEG data: eigenstructure decompositions of time varying autoregressions

Objective: Those who analyze EEC data require quantitative techniques that can be validly applied to time series exhibiting ranges of nonstationary be havior. Our objective is to introduce a new analysis technique based on for mal non-stationary time series models. This novel method provides a decompo sition of the time series into a set of 'latent' components with time-varyi ng frequency content. The identification of these components can lead to pr actical insights and quantitative comparisons of changes in frequency struc ture over time in EEG time series. Methods: The technique begins with the development of time-varying autoregr essive models of the EEG time series. Such models have been previously used in EEG analysis but we extend their utility by the introduction of eigenst ructure decomposition methods. We review the basis and implementation of th is method and report on the analysis of two channel EEG data recorded durin g 3 generalized tonic-clonic seizures induced in an individual as part of a course of electroconvulsive therapy for major depression. Results: This technique identified EEG patterns consistent with prior repor ts. In addition, it quantified a decrease in dominant frequency content ove r the seizures and suggested for the first time that this decrease is conti nuous across the end of the seizures. The analysis also suggested that the seizure EEG may be best modeled by the combination of multiple processes, w hereas post-ictally there appears to be one dominant process. There was als o preliminary evidence that these features may differ as a function of ECT therapeutic effectiveness. Conclusions: Eigenanalysis of time-varying autoregressive models has promis e for improving the analysis of EEG time series. (C) 1999 Elsevier Science Ireland Ltd. All rights reserved.