PROJECT TITLE :

Low-Complexity Seizure Prediction From iEEG/sEEG Using Spectral Power and Ratios of Spectral Power

ABSTRACT:

Prediction of seizures may be a troublesome problem as the EEG patterns aren't wide-sense stationary and change from seizure to seizure, electrode to electrode, and from patient to patient. This paper presents a novel patient-specific algorithm for prediction of seizures in epileptic patients from either one or 2 single-channel or bipolar channel intra-cranial or scalp electroencephalogram (EEG) recordings with low hardware complexity. Spectral power features are extracted and their ratios are computed. For each channel, a complete of forty four features together with eight absolute spectral powers, eight relative spectral powers and 28 spectral power ratios are extracted each two seconds using a 4-second window with a 50% overlap. These options are then ranked and selected in a very patient-specific manner employing a two-step feature selection. Selected features are any processed by a second-order Kalman filter and then input to a linear support vector machine (SVM) classifier. The algorithm is tested on the intra-cranial EEG (iEEG) from the Freiburg database and scalp EEG (sEEG) from the MIT Physionet database. The Freiburg database contains eighty seizures among 18 patients in 427 hours of recordings. The MIT EEG database contains seventy eight seizures from seventeen youngsters in 647 hours of recordings. It is shown that the proposed algorithm can achieve a sensitivity of 100p.c and a median false positive rate (FPR) of zero.0324 per hour for the iEEG (Freiburg) database and a sensitivity of 98.68percent and an average FPR of 0.0465 per hour for the sEEG (MIT) database. These results are obtained with leave-one-out cross-validation where the seizure being tested is often unseen from the training set. The proposed algorithm also incorporates a low complexity because the spectral powers will be computed using FFT. The area and power consumption of the proposed linear SVM are two to 3 orders of magnitude but a radial basis operate kernel SVM (RBF-SVM) classifier. Furthermore, the entire energy consumption of a system using linear - VM is reduced by 8p.c to twenty three% compared to system using RBF-SVM.