Non-Contact Short-Term Heart Rate Variability Analysis Under Paced Respiration Based on a Robust Fiber Optic Sensor System

In this article, a low-cost fiber optic sensor (FOS)-based system is proposed for the long-term stable measurement of ballistocardiography (BCG), which can be used to analyze short-term heart rate variability (HRV). The core part of the system is a single-mode fiber-based Mach–Zehnder interferometer (MZI), which consists of a
$1\times2$
coupler and a
$3\times3$
coupler. Due to the asymmetric structure of the two MZI arms, the interference spectrum will be observed, which will move linearly with the change of the transverse stress. The incident light is emitted from a low-cost distributed feedback (DFB) laser and coupled into the MZI sensor. The vibrations induced by human respiration and heartbeat will result in the variation of light phase in MZI, and further give rise to the change of interfering light intensity, which is detected by three photodetectors (PDs). The FOS-based BCG interbeat interval (IBI) shows high correlations with electrocardiography (ECG)-derived IBI, with correlation coefficients of 0.9971 and 0.9581 for healthy and premature beats subjects, respectively. During the one-month stability test, the FOS-based BCG monitoring system realized high accuracy. An inverse relationship between BCG-derived HRV amplitude and respiratory rate is revealed in this study. In this work, we also discover the phenomenon that breathing at 0.1 Hz causes larger respiratory sinus arrhythmia (RSA) amplitude, and resonant frequency breathing is observed in BCG-derived IBI signals. In addition, the time-domain, frequency-domain, and nonlinear analysis results of BCG HRV and ECG HRV at 0.1 and 0.25 Hz respiratory rates are highly correlated. In addition to high accuracy, the FOS-based BCG monitoring system has the unique advantage of flexible configuration to obtain continuous and accurate HR information without contact. This system has the potential to become a preferred alternative for ECG in daily monitoring.