Modeling-Based Radial Pressure Waveform Reconstruction Using Photoplethysmography Signals

This study introduces a model linking photoplethysmography (PPG) dynamics to radial pressure waveform (RPW), which could be integrated into digital twins, that enables the reconstruction of RPW from PPG measurements using pulse pressure extrema. Built upon existing literature and supervised symbolic regression on anesthesia data, the model was validated on 581 continuous 10 seconds subsequences from 24 patients. Calibration through an unscented Kalman filter ensured patient-specific accuracy, yielding an averaged Pearson correlation coefficient of 0.955 for the reconstructed signal. The model’s ordinary differential equation (ODE) with three parameters showed consistency with existing models. The stable, identifiable parameters underscore the model’s robustness. The proposed model gives some insights into the physiology hidden behind the PPG and paves the way for RPW reconstruction using non-invasive measurements.