Single channel photoplethysmography-based obstructive sleep apnea detection and arrhythmia classification

Area of Science:

• Biomedical Engineering
• Cardiology
• Sleep Medicine

Background:

• Early diagnosis and prevention of obstructive sleep apnea (OSA) and its complications require simplified monitoring.
• Developing user-friendly approaches is essential for widespread clinical adoption.

Purpose of the Study:

• To propose algorithms for OSA detection and arrhythmia classification using single-channel photoplethysmography (PPG).
• To enable early screening of OSA and associated cardiac irregularities.

Main Methods:

• Extracted fourteen features from PPG signals of 30 clinically diagnosed OSA patients.
• Evaluated the performance of a support vector machine (SVM) for apnea event detection based on varying feature inputs.
• Developed a multi-classification algorithm using modified Hausdorff distance for recognizing sinus rhythm and four common arrhythmias linked to sleep apnea.

Main Results:

• A feature set including meanPP, SDPP, RMSSD, meanAm, and meank1 offered an optimal balance between performance and complexity for OSA detection.
• The arrhythmia classification algorithm achieved an average sensitivity of 83.79%, specificity of 95.91%, and accuracy of 93.47% for classifying regular rhythm and four types of arrhythmia.

Conclusions:

• Single-channel PPG-based OSA detection and arrhythmia classification present a viable and promising strategy.
• This method facilitates early screening and diagnosis of OSA and its related arrhythmias.