Enhanced Central Sleep Apnea Detection Using a Hybrid Deep Learning Model with EEG Signals

Sleep apnea is a serious sleep disorder that disrupts normal breathing and is classified into obstructive, central, and complex types. Central sleep apnea (CSA), caused by impaired neurological control of breathing, can lead to serious health problems such as cardiovascular disease and neurocognitive decline. Traditional CSA diagnosis methods are based on manual polysomnography analysis, which is time consuming and error prone. This study uses a hybrid deep learning model that combines convolutional neural networks (CNN), bidirectional long short-term memory (BiLSTM), and recurrent neural networks (RNN). CNN extracts spatial features, BiLSTM captures bidirectional temporal dependencies, and RNN models sequential patterns to accurately classify apnea events. When evaluated using the MIT-BIH Polysomnographic database, the proposed model achieved an accuracy of 87.8%, highlighting its potential to enhance CSA diagnosis and improve patient outcomes through automation.