A Dual-Discriminator GAN for Sleep EEG Signal Synthesis

Open Access

Abstract: The interpretation of one’s overnight sleep process based on EEG signal is of importance for the inspection and treatment of various sleep-related disorders. The automatic sleep staging models have benefits as the assistant computerized tools to release the clinicians from the laborious task of manual sleep stage scoring. However, the issues of data insufficient and class imbalance are common for clinical data. The data problem is crucial to be solved when applying the automatic sleep staging models for real clinics. In this research, the network architecture of GAN (Generative Adversarial Network) is investigated by using one generator and two discriminators for the synthesis task of sleep EEG signals. The data augmentation performance by the dual-discriminator GAN and the sleep stage classification performance by combining with a typical machine learning classifier are evaluated on the sleep recording of subjects. The obtained results showed that the constructed dual-discriminator GAN is effective to generate samples which are closer to the time and frequency characteristics of real sleep EEG signal. It would be a useful method to solve the data problems for the training and optimization of automatic sleep stage classifiers in the field of sleep staging.
Keywords: Generative adversarial network; Electroencephalograph; Data augmentation; Sleep staging; Dual-discriminator