Breathing Rate Estimation from Head-Worn Photoplethysmography Sensor Data Using Machine Learning.

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Wearable Technology

Background:

• Breathing rate is a critical vital sign, but its monitoring is complex.
• Estimating respiration from heart activity (photoplethysmography) is challenging, especially outside labs and with new wearable sensors.
• Existing algorithms struggle with motion artifacts and novel sensor placements.

Purpose of the Study:

• To develop and validate a novel algorithm for accurate breathing rate estimation.
• To utilize photoplethysmography (PPG) data from a head-worn virtual reality mask.
• To address challenges of motion artifacts and sensor placement in wearable respiratory monitoring.

Main Methods:

• Developed a new algorithm using advanced signal processing and machine learning techniques.
• Integrated a novel quality assessment and motion artifact removal procedure.
• Evaluated the algorithm on two datasets comprising 37 subjects, comparing it to existing methods.

Main Results:

• The proposed algorithm achieved a mean absolute error of 1.38 breaths per minute.
• A Pearson's correlation coefficient of 0.86 was obtained, demonstrating strong agreement.
• The novel algorithm outperformed existing approaches in breathing rate estimation accuracy.

Conclusions:

• Reliable breathing rate estimation is feasible using PPG data from head-worn devices.
• The developed algorithm offers a promising solution for non-invasive respiratory monitoring in real-world conditions.
• Advanced signal processing and machine learning can overcome limitations of current wearable respiratory monitoring techniques.