Wearable monitoring of central blood pressure using seismocardiography

This thesis investigates how seismocardiography (SCG)—measurement of chest vibrations generated by cardiac activity may offer a viable basis for beat-to-beat blood pressure (BP) estimation. The findings suggest that timing features derived from heart-proximal mechanical events have the potential to improve the robustness of cuffless BP monitoring, particularly by reducing sensitivity to variations in peripheral circulation that commonly affect wearable sensors.

The research addresses a long-standing clinical and engineering goal: continuous, unobtrusive, and reliable BP monitoring suitable for healthcare, research, and daily life. Two in vivo datasets were analyzed: (i) measurements from healthy participants performing physiological tasks that induce rapid changes in cardiovascular state, and (ii) recordings from a highly invasive clinical procedure under deep anesthesia, providing arterial blood pressure as a reference. These datasets enabled assessment of accuracy during fast hemodynamic fluctuations under very different conditions.

To better understand mechanisms that confound pulse-wave–based approaches, a controllable circulatory phantom was developed. This system allows independent manipulation of pulsation rate and pressure while adjusting vessel elasticity, diameter, and fluid viscosity. Experiments show that such vascular properties can alter pulse-transit timing, highlighting why fixed calibration models may fail outside their original conditions.

Drawing on these insights, the thesis evaluates SCG-based electro-mechanical timing markers for beat-to-beat BP tracking and discusses potential clinical and wearable applications, including cerebral autoregulation assessment and sleep monitoring. Overall, the work provides a comprehensive examination of SCG for continuous BP monitoring, outlines its opportunities and limitations, and proposes a practical path toward more reliable non-invasive cardiovascular assessment in the future.