Toward a Robust Estimation of Respiratory Rate using Cardiovascular Biomarkers: Robustness Analysis under Pain Stimulation

Abstract

Respiration can modulate the cardiovascular system through the autonomic nervous system (ANS), deriving numerous methods for monitoring respiration based on cardiovascular biomarkers. However, the sensitivity of the ANS to environmental changes can negatively affect these methods, which suggests the necessity to evaluate their performance in estimating respiratory rate (RR). This paper aims to propose a method for robust estimation of RR using a biodegradable piezoelectric sensor by analyzing the robustness differences of these biomarkers under pain stimulation. In an electrocutaneous stimulus experiment conducted with 15 participants, arterial pulse waves near the elbow and wrist were measured, as well as the electrocardiogram and fingertip photoplethysmogram. The robustness of six biomarkers was quantified using respiratory quality index (RQI) and mean absolute percentage error (MAPE). Heart rate derived from the arterial pulse wave near the elbow achieves the best robustness (RQI = 85.67±12.84%, MAPE = 2.22±1.81%) of all biomarkers, whereas pulse wave velocity (PWV) from the elbow to the wrist performs best (RQI = 70.39±12.15%, MAPE = 3.47±1.69%) of the three biomarkers of PWV. Therefore, the robustness of biomarkers varies, as does the same biomarker measured at different sites. Our results reveal the heterogeneity of respiratory modulation on the cardiovascular system and demonstrate the robustness of the biomarkers of the arterial pulse wave near the elbow in estimating RR. This study can help smart wearables perfect respiratory monitoring and contribute a robust method for respiratory monitoring using a biodegradable piezoelectric sensor.

Ziqiang Xu

Doctor of Engineering

My research interests include non-invasive neurosensing techniques and wearables.