A Novel Algorithm for the Automatic Detection of Sleep Apnea From Single-Lead ECG

Area of Science:

• Biomedical Engineering
• Cardiology
• Sleep Medicine

Background:

• Sleep apnea is a common disorder with significant health implications.
• Current diagnostic methods can be invasive or require specialized equipment.
• There is a need for non-invasive, automated sleep apnea detection methods.

Purpose of the Study:

• To develop and validate a novel methodology for automatic sleep apnea detection using single-lead electrocardiogram (ECG) signals.
• To introduce new ECG-derived features for improved apnea detection accuracy.
• To assess the performance of the proposed method against established techniques.

Main Methods:

• Utilized two novel ECG-derived features (QRS complex morphology, heart rate-respiration coupling) and two standard heart rate variability features.
• Employed orthogonal subspace projections to extract respiratory information from ECG.
• Applied a least-squares support vector machine classifier with an RBF kernel to 80 ECG recordings.

Main Results:

• Achieved approximately 85% accuracy on a minute-by-minute basis for detecting both hypopneas and apneas across two independent datasets.
• Demonstrated 100% accuracy in distinguishing between apnea and normal recordings.
• The methodology successfully determined the contamination level of each ECG minute, aiding artifact detection.

Conclusions:

• The proposed ECG-based methodology offers high accuracy for automated sleep apnea detection, comparable to existing algorithms.
• Solely relying on ECG sensors is sufficient for achieving good diagnostic performance.
• The ability to assess ECG segment contamination enhances the reliability and interpretability of automated sleep apnea detection.