ResSleepNet: Automatic sleep stage classification on raw single-channel EEG

Abstract

Automatic sleep stage classification is an important paradigm in intelligence and promises considerable advantages to the health. Electroencephalography (EEG) is a reflection of the electrophysiological activities of brain neurons. Most current automated methods require the multiple electroencephalogram channels and rely on hand-engineered features which require prior knowledge about sleep stage scoring. Consequently, research efforts to purse for an automatic stage scoring based on machine learning techniques have been carried out over the last years. In this paper, a deep learning model, named ResSleepNet, is proposed for automatic sleep stage scoring based on raw single-channel EEG, it can automatically learn features from raw single channel EEG signal, and build an automatic sleep staging model for assisted sleep staging. The model is applied to an open-access database named Sleep-EDF, and the results demonstrated that the model scored the EEG epochs with the accuracy of 87.9%.