Comparison of pulse rate variability and morphological features of photoplethysmograms in estimation of blood pressure

Photoplethysmography is an optical technique that produces a wealth of information about cardiovascular health. Therefore, the technology has become integral part personal health monitoring devices. Given importance blood pressure measurement and control in physical mental health, recent years, estimation from photoplethysmography been active area research with promising results. Most studies on subject rely morphological features photoplethysmogram. These are highly prone to noise, changes sensor placement, skin properties; including colour. To address these limitations, we investigated feasibility using pulse rate variability which known be less aforementioned limitations. this end, collected high quality photoplethysmograms bespoke, research-grade device 18 healthy subjects. Approximately 15 min continuous waveforms were each subject. We trained machine learning models based different feature sets compared their performances. The model alone outperformed features, root mean squared error (RMSE) 6.32 vs 7.23 mmHg. However, best performance was obtained combined set (RMSE: 5.71 mmHg). Combined, evidence shows BP PRV, or conjunction feasible. In light limitations pressure, our findings lend support further use features.