Photoplethysmographic pulse decomposition analysis for stiffness differentiation in microfluidic phantoms

Microvascular stiffness is a critical physiological parameter associated with vascular health, yet its non-invasive assessment remains challenging. The stiffness of the vascular affects the photoplethysmography (PPG) signal, which represents pulsatile blood volume changes and holds promise for vascular stiffness estimation. The most common noninvasive method is using carotid-femoral pulse wave velocity. However, this method requires two measurement sides and carotid artery plaque may affect the results. Thus, using a single measurement side is preferrable. Aging affects blood vessel stiffness and PPG pulses from different ages also have different shapes. Hence, study of the pulse shape may reveal the stiffness of the blood vessels. In this study, we used a pulse decomposition analysis method to extract five standard log-normal pulses from each PPG pulse. Each standard log-normal pulse bears amplitude (A), center (x), and width (s) parameters; each parameter is numbered based on the chronological appearance of the standard log-normal pulses. Features were extracted based on these fifteen parameters, including the peak locations. To validate this approach, we employed microfluidic phantoms at different stiffness levels with either low or high pulse pressure, enabling systematic evaluation of the decomposition features in relation to vascular compliance. We found that A3 to A4 ratio worked well in low pressure to separate different stiffness. Both s1 and distance between peak 1 and 2 gave moderate results in both pressure levels.