Consistency of determinism detection in sparse photoplethysmogram recordings

Abstract:
A human photoplethysmogram is a biological signal widely used for heart rate estimation, and it has a great potential for a variety of physiological and mental health monitoring applications. Photoplethysmogram is a highly complex, chaotic signal, and health monitoring applications beyond the heart rate estimation, such as mental health monitoring, may possibly require the extraction of photoplethysmogram dynamical properties that can be achieved by the application of non-linear time series analysis. However, the applicability of nonlinear time series analysis imposes various requirements on the time series such as a sufficiently high sampling rate. This study aimed to investigate the effect of the sampling rate on the determinism of photoplethysmogram signals taken in reference conditions. The determinism, which is one of the fundamental properties of the dynamical system, was quantified by the recurrence quantification analysis, and the robustness of the determinism identification was verified by the surrogate data method. Results demonstrated that for the sampling frequency range 409.6~24.09 Hz only a minor decrease of determinism quantity is observed, thus confirming the consistency of the determinism detection in the photoplethysmogram time series of investigated sampling rates.