Convolutional neural network with autoencoder-assisted multiclass labelling for seizure detection based on scalp electroencephalography

Hirokazu Takahashi, Ali Emami, Takashi Shinozaki, Naoto Kunii, Takeshi Matsuo, Kensuke Kawai

Abstract: Highlights•A convolutional neural network (CNN) is useful to process images of EEG plots.•CNN and patient-specific autoencoders (AE) were combined to detect seizure.•AE was used to make a “non-seizure-but-abnormal” label.•The classification performance of AE-CNN was comparable to that of the original CNN.•The false seizure alarm rate in AE-CNN was one-fifth of that in the original CNN.