Multi-site cardiac rhythm monitoring via multi-channel SCG system and exercise-induced physiological analysis.

Purpose: This study aims to achieve fine-grained multi-chamber monitoring and further investigate exercise-induced cardiac vibration patterns across thoracic sites through spatiotemporally resolved mechanical analysis and quantitative characterization, based on a custom-built multi-channel seismocardiography (MSCG) system integrated with electrocardiography (ECG). This approach holds clinical importance for early screening and dynamic management of cardiac structural abnormalities. Methods: 1) For the multi-channel MSCG signals from continuous cardiac vibration monitoring across multiple thoracic regions (from our collected dataset and public dataset), a signal processing pipeline was established to extract temporal intervals and amplitude parameters from cardiac cycles. 2) These temporal disparities of corresponding feature points among channels, which could reflect the sequential mechanical activities of different cardiac chambers, were physiologically interpreted based on cardiac dynamic, thereby enabling a detailed analysis of chamber-specific timing relationships. 3) Furthermore, a pre- and post-exercise comparative protocol, analogous to stress ECG testing, was implemented to analyze multi-channel SCG feature changes and establish correlations between exercise-induced cardiac mechanical alterations and SCG parameters. Results: 1) Synchronized multi-channel SCG and ECG signals were successfully captured by the custom-built high-precision dual-modal acquisition system, enabling precise identification of characteristic feature points across all channels. 2) Statistical analysis of cardiac cycles from individual subjects revealed that following exercise, humans exhibited significant forward shifts in aortic valve closure (AC) and mitral valve opening (MO) points across all five channels, consistent with exercise-induced heart rate elevation and shortened cardiac phases. 3) Regarding amplitude, the mitral valve site showed the earliest MC point, though no consistent spatial sequence emerged for other feature points during exercise. Conclusion: 1) The spatiotemporal disparities across MSCG channels indeed correspond to the distinct physiological activities of the underlying cardiac chambers. 2) Exercise-induced cardiac mechanical dynamics were spatiotemporally resolved monitored by the system, revealing quantifiable timing shifts in valvular events undetectable by single-channel approaches. 3) The finding of consistent timing shifts in the AC and MO feature points under post-exercise scenario obtained via multi-channel monitoring, which supports the non-invasive assessment of cardiac function under exertion and in pathological conditions involving altered ventricular dynamics.