Automatic sleep stage classification for daytime nap based on hopfield neural network

In this study, automatic method of sleep stage classification for daytime nap is investigated. The ultimate objective is to identify the changing of sleep level during one´s nap. The sleep data is recorded according to the polysomnographic (PSG) measurement. The Electroencephalograph (EEG) is analyzed for sleep stage classification. Totally, 4 parameters are selected and calculated for each 20-second segment of EEG data. The main method is based on Hopfield Neural Network (HNN). The neural network is trained by using standard mode. The sleep stages are classified based on HNN for each consecutive segment. The obtained result showed about 80.6% consistence comparing with the visual inspection. The automatic classification results indicated the changing of sleep level during nap, which can be useful for daytime nap sleep evaluation.