A spatio-temporal learning-based model for sleep apnea detection using single-lead ECG signals.

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Sleep Medicine

Background:

• Sleep apnea (SA) is a prevalent chronic disorder linked to severe health issues like stroke and cardiovascular disease.
• Many individuals with SA remain undiagnosed due to asymptomatic or unrecognized sleep events.
• Existing ECG-based SA detection methods lack the accuracy required for clinical application.

Purpose of the Study:

• To develop an advanced, end-to-end spatio-temporal learning model for accurate sleep apnea detection using single-lead ECG signals.
• To improve the performance of home-based sleep monitoring systems for sleep apnea screening.
• To provide a reliable tool for identifying individuals at high risk of sleep apnea for timely intervention.

Main Methods:

• An end-to-end deep learning architecture was designed, featuring multiple spatio-temporal blocks.
• Each block integrates Convolutional Neural Network (CNN) for spatial feature extraction and Bi-Gated Recurrent Unit (BiGRU) for temporal dynamics.
• The model was trained and validated on the PhysioNet Apnea-ECG and University College Dublin Sleep Apnea Database (UCDDB).

Main Results:

• The proposed spatio-temporal learning model achieved state-of-the-art performance on both the Apnea-ECG and UCDDB datasets.
• Results demonstrated significant superiority over existing ECG-based sleep apnea detection techniques.
• The model effectively captured both morphological (spatial) and temporal features from ECG signals.

Conclusions:

• The developed ECG-based method shows high potential for accurate sleep apnea detection.
• This approach can be integrated into portable devices for effective home-based sleep monitoring.
• The method offers a promising solution for early screening of sleep apnea, facilitating timely medical interventions.