A Framework for Extracting Heart Rate Variability Features from Earbud-PPG for Stress Detection

Area of Science:

• Biomedical Engineering
• Health Informatics
• Signal Processing

Background:

• Unmanaged stress poses significant physiological and mental health risks.
• Continuous stress monitoring is essential for proactive health management.
• Wearable sensors offer a promising avenue for unobtrusive stress tracking.

Purpose of the Study:

• To investigate the efficacy of consumer-grade earbud photoplethysmography (PPG) sensors for stress classification.
• To develop a deep learning framework for extracting heart rate variability (HRV) features from noisy earbud PPG signals.
• To improve the accuracy and reliability of stress detection using wearable technology.

Main Methods:

• A deep learning model was developed to predict HRV features from earbud PPG data.
• A knowledge transfer approach was employed to utilize predicted HRV features for stress classification.
• The proposed framework was compared against a state-of-the-art HRV feature extraction library.

Main Results:

• The proposed framework demonstrated superior performance in HRV feature extraction compared to existing methods.
• Stress detection accuracy, sensitivity, and specificity showed at least a 5% improvement.
• The study successfully classified stress periods using data from consumer-grade earbuds.

Conclusions:

• Consumer-grade earbud PPG sensors, coupled with deep learning, can effectively monitor stress.
• The developed framework offers a significant improvement over current state-of-the-art methods for stress detection.
• This approach paves the way for accessible and continuous stress management tools.