Real-Time Classification of Deep and Non-Deep Sleep With Comparative Intervention Experiments

Sleep is an essential part of human life, and sleep quality is a critical indicator of overall health. This paper presents a system that utilizes a Brain-Computer Interface and a Deep Learning Network for the real-time classification of non-deep sleep and deep sleep. By collecting, uploading, preprocessing, and classifying sleep EEG data in real-time, the system quantitatively evaluates the subjects' sleep quality. Additionally, we conducted sleep experiments on the same subjects under varying environments and interventions and compared the results. Using our deep learning model, we calculated the proportion of deep sleep in each experimental group and validated the effects of fatigue state and white noise environment on improving the proportion. Based on this system, we examined the specific effects of different interventions (fatigue state and white noise intervention) on sleep states, providing concrete data to determine whether these factors enhance sleep quality.