Sleep Stage Identification of Split Wave EEG Signals Using CNN-RNN

Sleep disturbance can indeed have a profound impact on various aspects of health and well-being. Monitoring sleep quality and detecting sleep disorders are crucial to maintaining overall health. EEG data plays a significant role in understanding sleep patterns and identifying different sleep stages. Sleep is divided into five levels: Awake, Rapid eye movement (REM), Non-REM 1, Non-REM 2, and Non-REM 3. This research aims to improve the accuracy of identifying sleep levels from EEG signals by dividing the EEG signal into specific frequency bands (Delta, Theta, Alpha, Beta waves) using a Wavelet filter and then applying a Convolutional Neural Network (CNN) and a Recurrent Neural Network (RNN) in parallel for sleep level identification. The experimental results showed that the split signal waves approach achieved an accuracy of 98.34%, indicating a promising advancement in sleep stage identification. This high accuracy is particularly impressive, especially compared to the accuracy achieved using simultaneous signal processing.