The analysis of EEG texture content for seizure prediction

Texture features have been among major tools in image analysis field for decades. Their applications in various computerized biomedical image processing teak advantage of the fact that computers are better than human observers at analyzing second order statistics. The attempt to make use of a similar approach to provide additional insight into EEG pattern recognition was presented by A. Petrosian et al. (Annual Meeting of the American EEG Society, New Orleans, LA, USA, p.88, Oct. 10-15, 1993). This study represents further development of suggested methods in recognition of different interictal, preictal, ictal, and postictal stages from EEG recordings. Practical application that is relevant to this study includes prediction of epileptic activity in patients before an actual seizure occurs. Although this study is preliminary and was carried out on the data obtained from one patient, the results showed feasibility of using signal texture information for distinguishing different abnormal patterns.<>