Automated detection of abnormalities from an EEG recording of epilepsy patients with a compact convolutional neural network

Electroencephalography (EEG) is essential for the diagnosis of epilepsy, but it requires expertise and experience to identify abnormalities. It thus crucial develop automated models detection abnormalities in EEGs related epilepsy. This paper describes development a novel class compact convolutional neural networks (CNNs) detecting abnormal patterns electrodes The designed model inspired by CNN developed brain-computer interfacing called multichannel EEGNet (mEEGNet). Unlike EEGNet, proposed model, mEEGNet, has same number electrode inputs outputs detect patterns. mEEGNet was evaluated with clinical dataset consisting 29 cases juvenile childhood absence epilepsy labeled expert. labels were given paroxysmal discharges visually observed both ictal (seizure) interictal (nonseizure) durations. Results showed that detected area under curve, F1-values, sensitivity equivalent or higher than those existing CNNs. Moreover, parameters much smaller other models. To our knowledge, validated machine learning through this research largest literature.