Elucidating the Reactivity of Reflectance-Based Core-Body Photoplethysmogram to Posture and Respiratory Changes

At-home cardiovascular monitoring has become widespread with the commercialization of the photoplethysmogram (PPG) in wearable devices. A tremendous opportunity for comprehensive monitoring of chronotropic, inotropic, and vascular characteristics of the cardiovascular system can be unlocked if PPG data are combined with signals of other modalities at the core body. However, little is known about the quality of core-body PPG or its reactivity to environmental conditions. To elucidate core-body PPG signals, we compared the reactivity and quality of sternum and back PPG signals to finger-based PPG in 24 healthy participants during a protocol with posture and breathing changes. We found that the signal quality of core-body PPG was affected by body location, wavelength, LED/photodiode configuration, posture, and breathing pattern. Notably, reflectance-mode red-wavelength core-body quality was less than that of transmittance-mode finger red-wavelength PPG ( ${p}~\lt 0.05$ ). However, green-wavelength core-body PPG quality was comparable ( ${p}~\gt 0.05$ , sternum) or significantly greater ( ${p}~\lt 0.001$ , back). We also found that PPG amplitude reactivity for core-body signals was in accordance with that of the finger ( ${p}~\gt 0.05$ or ${p}~\lt 0.05$ , sign( $\mu _{\text {core}}$ ) = sign( $\mu _{\text {periphery}}$ )) for posture changes but opposite ( ${p}~\lt 0.05$ , sign( $\mu _{\text {core}}$ ) $\neq $ sign( $\mu _{\text {periphery}}$ )) or significantly blunted ( ${p}~\lt 0.05$ , sign( $\mu _{\text {core}}$ ) = sign( $\mu _{\text {periphery}}$ ), abs( $\mu _{\text {core}}$ ) < abs( $\mu _{\text {periphery}}$ )) during deep and resistive breathing. We, thus, demonstrate the importance of studying the reactivity of vasculature at different sensor locations, as their reactivity to various provocations is heterogeneous. Our findings can be used to develop improved hardware for sensing core-body PPG and algorithms for interpreting the signals acquired, enabling clinicians to more comprehensively understand patient cardiovascular health.