Multimodal apnea detection: advancements through convolutional neural networks and STFT analysis of EEG, ECG, and nasal signals to tackle key challenges in innovation.

Sleep apnea is a prevalent chronic disorder posing significant health risks, requiring prompt diagnosis for effective treatment. This study introduces a reliable and efficient method using simultaneous electroencephalography (EEG), electrocardiography (ECG), and nasal airflow signals to distinguish apnea subtypes. The approach involves signal preprocessing with bandpass filtering and applying the short-time Fourier transform (STFT) to create spectrograms for 30-second segments. A convolutional neural network (CNN) classifies normal events, obstructive sleep apnea (OSA), central sleep apnea (CSA), and hypopnea. Results show our method achieved 98.01% accuracy, highlighting its potential to enhance personalized care for sleep apnea patients.