Predicting Heart Rate Estimation Accuracy: Photoplethysmography Sensor Measurements on Eight Anatomical Sites.

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Wearable Technology

Background:

• Photoplethysmography (PPG) wearables enable continuous cardiac monitoring outside clinical settings.
• Accurate heart rate (HR) estimation is crucial for effective remote patient monitoring.
• Evaluating PPG signal quality across diverse anatomical sites is essential for reliable HR data.

Purpose of the Study:

• To assess the accuracy of infrared PPG sensors for HR estimation across eight anatomical locations.
• To characterize the relationship between PPG waveform features and HR estimation accuracy.
• To compare the efficacy of perfusion index (PI) and cardiac power ratio (PR) in predicting HR errors.

Main Methods:

• Collected PPG data from eight anatomical sites (fingertip, finger base, wrist, forearm, earlobe, mastoid bone, temple, forehead).
• Used Electrocardiogram (ECG) as the ground truth for HR measurement.
• Quantified associations between PPG waveform characteristics (PI, PR) and HR error.

Main Results:

• Perfusion index (PI) effectively predicted HR errors at sites with strong pulsatile components but was less reliable at sites with weaker pulses.
• Cardiac power ratio (PR) provided more accurate HR error predictions, particularly at sites with weaker pulsatile characteristics.
• PR enables the use of all recordings by allowing reliability assessment, unlike PI which often leads to discarding data.

Conclusions:

• Cardiac power ratio (PR) is a more reliable metric for predicting HR estimation errors from PPG signals compared to perfusion index (PI).
• PR facilitates a more comprehensive use of PPG data by enabling reliability assessment rather than data exclusion.
• These findings enhance the clinical relevance of PPG wearables by improving the trustworthiness of HR measurements across various anatomical sites.