Epileptic seizure prediction based on local mean decomposition and deep convolutional neural network

A reliable seizure prediction system has important implications for improving the quality of epileptic patients’ life and opening new therapeutic possibilities for human health. In this paper, a new method combining local mean decomposition (LMD) and convolutional neural network (CNN) is proposed for seizure prediction. Firstly, the LMD is employed to decompose the raw EEG signals into a string of product functions (PFs). Subsequently, three PFs (PF₂–PF₄) are selected to learn the EEG features automatically using the deep CNN. In order to obtain the most important information from the features extracted by the CNN, the principal components analysis is applied to remove the redundant features. After that, these features are fed into the Bayesian linear discriminant analysis for classifying the cerebral state as interictal or preictal. The proposed method achieves a sensitivity of 87.7% with the false prediction rate of 0.25/h using intracranial EEG signals of 21 patients from a publicly available EEG dataset. The experimental results suggest that the proposed method can become a potential approach for predicting the impending seizures in clinical application.