Advancing remote photoplethysmography (rPPG) to facilitate cardiac monitoring in naturalistic settings using webcam technology.

Remote photoplethysmography (rPPG) is a technique for the monitoring of cardiac signals using simple consumer-grade cameras, such as webcams. Its cost-effectiveness and accessibility make it a promising tool for widespread application, particularly in remote and resource-limited research settings. However, it remains unclear whether rPPG can accurately capture the moment-to-moment fluctuations in heart rate (HR) required for robust HR variability (HRV) estimation. In this study, we collected a novel dataset of 77 participants who completed three behavioral tasks under controlled laboratory conditions, designed to mirror the recording setup typically encountered in home-based online experiments. We validated four rPPG pipeline configurations against the electrocardiogram (ECG) ground truth, using a comprehensive set of measures capturing both signal quality and cardiac variability. Moreover, using data from an emotion anticipation task, we assessed whether, in a typical experimental setting, analyses based on rPPG data yielded statistical results comparable to those obtained from ECG. Our results indicate that rPPG is a promising tool for detecting HR and HRV at the group level, particularly in high-powered studies. At the individual level, rPPG reliably captures average HR, but the accuracy of individual-level HRV estimates remains limited. The findings underscore rPPG's potential as a low-cost, noninvasive research tool, while highlighting the importance of managing its limitations such as sensitivity to excessive movement and poor illumination through thoughtful study planning and execution.