Comparison of FFT and AR techniques for scalp EEG analysis

Abstract

Scalp electroencephalogram (EEG) with bipolar montage is used in most infirmaries for monitoring epilepsy. However, scalp EEG is unpopular as compared to IEEG (intra-cranial EEG) in the research field. Most researchers used IEEG and scalp EEG with unipolar montage. Bipolar montage is also rarely used in the research in contrast to unipolar montage. The main aim of this paper is to investigate and determine a suitable method for processing EEG data using bipolar montage directly from the hospital archive. Two well-known methods namely, the Fast Fourier Transform (FFT) and the Autoregressive (AR) will be analyzed and compared based on their power spectrums. Results obtained based on monitored frequencies showed that the AR method is better than FFT in delineating the epilepsy region which can be visually observed and recognizable. © 2008 Springer-Verlag.