Spectrophotometric Study of Turbid Media Samples and Artery Phantoms for Modeling of Photoplethysmography Process

Abstract

Photoplethysmography (PPG) has long been explored and applied as a useful modality to measure numerous physiologically relevant parameters within cardiologic and home-health applications. Despite these successes, the fundamental process of light-tissue interaction underlying PPG signals remains poorly understood. This dissertation research develops the tools needed for experimental and numerical investigations of the underlying light-matter interaction of PPG signals within the framework of radiative transfer theory (RT) to improve the understanding of physiological factors attributing to their formation. We have developed and validated a multiparameter spectrophotometry (MPS) system and a novel collection of algorithms for recovery of bulk optical parameters of turbid media samples from measured signals of diffuse reflectance, diffuse transmittance, and forward transmittance from 460 to 1000 nm without integrating spheres. A particle swarm optimization (PSO) based inverse solver has been combined with GPU-executed Monte Carlo (MC) simulations to rapidly retrieve optical parameters defined by the RT theory. Extensive validation results of our MPS system are presented that include determination of RT parameters of Intralipid and bovine milk samples as functions of wavelength. Additionally, we demonstrate the ability of our novel MPS approach to measure a competitive threshold concentration of detectable bacteria in milk. The MPS system and the inverse solver were then applied in a PPG modeling study. Multiple polydimethylsiloxane (PDMS)-human blood phantoms were constructed to simulate up to eight blood vessels embedded in a tissue phantom. Scattered light was measured as PPG signals under various values of blood pressure at multiple wavelengths and quantitatively compared to those obtained by MC simulations. Our results demonstrate the potential of our phantom model to advance the understanding of factors that give rise to the photoplethysmogram. Future improvements of our phantom study may provide additional insight on extraction of useful biomarkers from PPG measurement to monitor cardiovascular health in clinics.