Impact of sensor configuration and melanin concentration on reflective pulse oximetry using Monte Carlo simulations

Abstract

Abstract We investigate the impact of melanin concentration C Mel and photoplethysmography (PPG) sensor configuration on signal quality and estimation accuracy of oxygen saturation  $$\mathrm {SpO_2}$$ . We deploy Monte Carlo (MC) simulations of photon-skin interactions to estimate arterial oxygen saturations  $$\mathrm {SaO_2}$$ ranging from 70 to 100% and melanin concentration C Mel ranging from 2.55 to 30.5%. We analysed the effects of red and infrared wavelengths (624 nm, 660 nm, 850 nm, and 940 nm), beam profiles (LED and VCSEL), as well as beam incidence angles of light sources (0°, 45°, and − 45°) on Perfusion Index (PI), reflective $$\mathrm {SpO_2}$$  estimation, and signal-to-noise ratio (SNR) for source-detector distances ranging from 2  mm to 9 mm. Maximum PI was observed for in the red spectral range at 624 nm and in the infrared spectral range at 940 nm. In contrast, reflective pulse oximetry provided more accurate results at 660 nm and 850 nm, independent of wavelength combination. We conclude that the VCSEL beam profile at 0° is the optimal light source for a wide range of applications, as it offers a balance between PI and absolute $$\mathrm {SpO_2}$$  estimations error.