PPG-Based Sleep Apnea Detection
PPG enables wearable sleep apnea screening by detecting the characteristic oxygen desaturation events (SpO2 dips below 90%), heart rate surges following apnea termination, and sympathetic nervous system activation signatures encoded in IBI and pulse amplitude. Wearable PPG achieves AHI estimation correlation r = 0.7–0.85 with polysomnography-derived AHI.
Obstructive sleep apnea (OSA) produces a stereotyped PPG signature during apneic events: progressive SpO2 decrease (typically 3–5% below baseline), loss of respiratory modulation in the PPG amplitude envelope, followed by abrupt SpO2 recovery and heart rate surge upon apnea termination. The autonomic arousal at apnea termination produces a distinctive sympathetic burst visible in the PPG as a transient vasoconstriction (reduced AC amplitude) followed by vasodilation. These signatures repeat cyclically at the AHI-dependent rate.
The ODI (Oxygen Desaturation Index) — number of SpO2 dips below 3% or 4% per hour — is the primary PPG-derived metric for OSA severity screening. ODI correlates with polysomnography AHI (r = 0.82–0.92 in clinical studies), and ODI4 > 15 events/hour has sensitivity 93% and specificity 80% for moderate-to-severe OSA (AHI > 15). Consumer SpO2 wearables including Withings ScanWatch, Apple Watch Ultra, and Samsung Galaxy Watch are used clinically for OSA screening, with CE marking in Europe and FDA Breakthrough Device designation for some devices.
Beyond oxygen saturation, machine learning models fusing PPG-derived features including IBI variability, pulse amplitude envelope (PAE), PTT, and photoplethysmography-derived respiratory rate (PDRR) achieve AUC 0.88–0.93 for OSA/non-OSA classification and RMSE of 4–6 events/hour for AHI regression — approaching the inter-rater reliability of manual polysomnography scoring (kappa ~0.7).
Frequently Asked Questions
Can a fitness tracker diagnose sleep apnea?
No. Consumer PPG devices can flag high-risk nights warranting formal polysomnography evaluation but cannot provide diagnostic-grade AHI scores. FDA-cleared sleep apnea diagnostic devices require prescription and clinical validation to ISO 14155 standards.
How does PPG compare to polysomnography for sleep apnea?
Polysomnography (PSG) with electrodes, respiratory belts, and airflow sensors remains the gold standard. Wearable PPG achieves 70–85% sensitivity for moderate-severe OSA at 75–85% specificity in validation studies — suitable for screening but not diagnosis.
Does REM sleep affect PPG sleep apnea detection?
Yes. REM-related OSA often involves longer, more complex apneas with greater SpO2 desaturation and larger autonomic surges. Algorithms trained without REM stage labeling underestimate REM AHI by 20–40% due to different morphological signatures.