Estimation of Respiratory Rate during Biking with a Single Sensor Functional Near-Infrared Spectroscopy (fNIRS)...

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Wearable Technology

Background:

• Wearable systems are increasingly used for continuous vital sign monitoring.
• Conventional bio-signals are susceptible to motion artifacts, challenging respiratory rate (RR) estimation during physical activity.
• Functional Near-Infrared Spectroscopy (fNIRS) offers a less movement-vulnerable alternative for physiological monitoring.

Purpose of the Study:

• To propose and validate a novel fusion-based method for estimating respiratory rate (RR) during bicycling using wearable fNIRS signals.
• To address the limitations of traditional bio-signal-based RR estimation during dynamic physical activities.

Main Methods:

• Extracted five respiratory modulations from oxygenated hemoglobin concentration (O2Hb) signals based on amplitude, frequency, and intensity.
• Computed dominant frequencies of each modulation using Fast Fourier Transform (FFT).
• Fused dominant frequencies via averaging to estimate RR, validated on 22 subjects during bicycling.

Main Results:

• The proposed fusion method demonstrated superior performance compared to traditional band-pass filtering.
• Achieved a significantly lower mean absolute error in RR estimation (3.66 vs. 11.06 breaths per minute, p<0.05).
• Outperformed RR estimations derived from individual modulation analyses.

Conclusions:

• The developed fusion method effectively estimates RR from fNIRS signals during bicycling.
• Highlights the practical limitations of conventional bio-signals for RR monitoring during physical exertion.
• Suggests fNIRS as a viable technology for robust respiratory monitoring in wearable systems.