Use of K-means clustering in migraine detection by using EEG records under flash stimulation

Abstract

In this study, a new migraine analysis method was proposed by using EEG (electroencephalography) signals under flash stimulation in time domain. EEG signal is one of the most complex biomedical signals due to its complex nature. Therefore, these types of signals are commonly pre-processed before the analysis procedure. Since, pre-processing techniques affect the analysis results positively or negatively, the achievements of these pre-processing techniques are very important. Flash-stimulated and non-stimulated EEG signals obtained from healthy subjects and migraine patients. First, a digital band pass FIR (Finite Impulse Response) filter is used to select beta band of the T5-T3 channel EEG signals. Then a time domain based pre-processing technique, histogram, was employed to obtain the features. Histogram differences in the case of flash stimulation (or non-stimulation) calculated and used as features for the healthy subjects and migraine patients. These features are applied to a k-means clustering algorithm to see clustering results of the proposed technique. Silhoutte clustering results show that, a good clustering performance is evaluated as 86.6% correct clustering rate (CCR) with the used techniques in migraine analysis. Results provide a good insight in this area for future studies. Key words: Migraine, EEG, K-means cluster, histogram.