A Novel Epileptic Seizure Detection Algorithm Based on Analysis of EEG and ECG Signals Using Probabilistic Neural Network

In this paper, we present a novel epileptic seizure detection algorithm based on analysis of electroencephalogram (EEG) and electrocardiogram (ECG) signals to detect seizure onsets that are not associated with rhythmic EEG activity. In this algorithm, spectral and spatial features are extracted from seizure and non-seizure EEG signals by Gabor functions and combined with four extracted features of ECG signals to form feature vector. Then a probabilistic neural network (PNN) classifier is used to determine an optimal nonlinear decision boundary. This proposed algorithm can automatically detect the presence of seizures which can be important advance facilitating timely medical intervention. Our algorithm is evaluated on 12 records of physionet database. The obtained results indicate that the proposed algorithm can recognize 98.25% of seizures with a false detection rate of 12.47%.