A Platform for Electroencephalogram Analysis and Seizure Prediction in Epilepsy

When a focal seizure is generated, synchronized brain activity spreads out to other brain areas. Therefore, the neural information hidden in electroencephalogram (EEG) before the onset of seizure was recently investigated to find useful information for seizure prediction. Up to now several parameters such as phase coupling, nonlinear characteristics, etc. were proposed and applied to prediction algorithms. Although the predictability of seizure was shown in previous oversea researches, there are still several issues that need further investigations, such as the role of other physiological systems like the autonomic nervous system, the influence of interference and patient vigilance on the accuracy of seizure prediction, the elimination of the above influence by advanced signal processing methods, and so on. Furthermore, a seizure prediction study needs lots of clinical data and examination, and complex, heavy computation and algorithm development, which rely on intense cooperation between neurologists and engineers. Upon a full support by Department of Neurology, Chang Gung Memorial Hospital, this project is aimed to build a platform for EEG analysis, data display and database establishment, on which the seizure predictability and related algorithm can be studied. This study is scheduled in three years. In the first year, a platform for basic EEG analyses such as coherence, nonlinear analyses, etc. will be developed; the EEG characteristics and heart rate va riability during interictal, preictal, and ictal will be also investigated. In the second year, the predictability of available and newly developed EEG parameters and prediction algorithms will be assessed. In the third year, the influence of interference and patient vigilance on seizure prediction will be tested, and the advanced signal processing methods will be employed to reduce the effect of such interferences. Not only for studying seizure prediction, a new research team between engineering and medicine for epilepsy investigations will also be formed. Our scheduled platform and studies can be further integrated with brain imaging such as magnetic resonance imaging and single photon emission computed tomography for the combined investigation of neurophysiology and anatomical information.

Project ID:PB9706-1245
External Project ID:NSC96-2221-E182-001-MY3