Adaptive predictive modelling for the analysis of the epileptic EEG

Abstract

A signal processing model for the epileptic EEG is proposed. This is used to formulate an analysis model, based on linear prediction. This formulation is implemented as a number of adaptive filters and applied for the detection of epileptic spikes. The theory behind the method is explained and the implementation is described. Results are presented and compared for two adaptive filter realizations. The algorithm is compntationally efficient and can be implemented in real-time on a small microcomputer system for on-line analysis. Intermediate results, produced by this method may be used for further analysis. Generalization for the detection of other EEG transients and the removal of artifacts can be easily achieved.

Linear Prediction Modelling for the Analysis of the Epileptic EEG

“…Parametric modelling is a technique used for time series analysis in which a mathematical model is fitted to a sampled signal. If the model forms a good approximation to the signal's observed behavior it can then be used in a wide range of applications, such as spectral estimation, Linear Prediction Coding (LPC) for data compression, speech synthesis and feature extraction for pattern classification problems [6][7][8].…”