SleepSatelightFTC: A Lightweight and Interpretable Deep Learning Model for Single-Channel EEG-Based Sleep Stage Classification

Sleep scoring by experts is necessary for diagnosing sleep disorders. To this end, electroencephalography (EEG) is an essential physiological examination. As manual sleep scoring based on EEG signals is time-consuming and labor-intensive, an automated method is highly desired. One promising automation technology is deep learning, which has performed well or better than experts in sleep scoring. However, deep learning lacks adequate interpretability, which is crucial for ensuring safety and accountability, especially for complex inference processes. We propose SleepSatelightFTC, a lightweight model that achieves comparable performance to state-of-the-art models with only one-third of their parameters. Based on the rules for sleep scoring, self-attention is applied to each of the time- and frequency-domain inputs, a raw EEG signal and its amplitude spectrum. The simple method of continuously connecting the intermediate outputs of the epoch-wise model has resulted in a highly lightweight architecture. On the Sleep-EDF-78 dataset, our model achieves an accuracy of 84.8% and a kappa coefficient of 0.787 while requiring significantly fewer parameters ( $0.47\times 10^{6}$ ) compared to existing models (1.3– $4.54\times 10^{6}$ ). The visualization of feature importance obtained from self-attention confirms that the proposed model learns representative waveform features, including K-complexes and sleep spindles.