Abstract:Seizures can cause great physical and mental trauma to patients, and accurate prediction of seizures can assist physicians to take treatment measures for patients in a timely manner. In order to accurately predict seizures, this paper proposes a seizure prediction method that integrates EEG features and multichannel EEG interaction features. Firstly, the multi-scale symbolic alignment transfer entropy is proposed to analyze the multi-channel EEG signal interaction information, generate the synchronization matrix, and screen the important EEG channels related to seizures by significance analysis to reduce the interference of unnecessary features to the classification. Next, the power spectral density energy diagram (PSDED), which characterizes the EEG signal, and the synchronization matrix diagram (SMD), which describes the interaction characteristics of brain channels, were generated for the EEG signals after screening the channels; the two feature maps were fused, and deep convolutional neural network (DCNN) was used to classify and identify the EEG signals of epilepsy patients, which improved the learning ability and generalization ability, and the classification accuracy could reach 96.825%; finally, a prediction evaluation system was used on the basis of classification to evaluate the seizure prediction performance, and the prediction sensitivity reached 96.66% and the false detection rate could reach 0.03/h when the seizure prediction range (SPH) was 10 minutes and the seizure onset period (SOP) was 10 minutes.When the SPH is 30 min and SOP is 10 min, the prediction sensitivity reaches 93.17% and the false detection rate can reach 0.05/h. Compared with the results of existing studies, the method in this paper has better prediction sensitivity and lower false detection rate.