Simplified single neuron model for robust local pulse wave velocity sensing using a tetherless bioimpedance device.

ABSTRACT

Pulse arrival time (PAT), Pulse transit time (PTT), and Pulse Wave Velocity (PWV) have all been used as metrics for assessing a number of cardiovascular applications, including arterial stiffness and cuffless blood pressure monitoring. These have been measured using various sensing methods, including electrocardiogram (ECG) with photoplethysmogram (PPG), two PPG sensors, or two Bioimpedance (BioZ) sensors. Our study addresses the mathematical inaccuracies of previous bioimpedance approaches and incorporates PTT weights for the peak-peak (PTTpp), middle-middle (PTTmm), and foot-foot (PTTff) segments of the sensing signal into a single neuron model to determine a more accurate and stable PWV. In addition, we developed a tetherless bioimpedance device and compared our PTT estimation approaches, which yielded PWV across six subjects and two different arteries. Specifically, using our model, we found that the most reliable combination of weights corresponding to PTTpp, PTTmm, and PTTff was (0.260, 0.704, 0.036) for the brachial artery and (0.104, 0.858, 0.038) for radial artery. This model consistently yielded stable values across repetitions, with PWV values of 5.2 m/s, 5.3 m/s, and 5.9 m/s for the brachial artery and values of 5.8 m/s, 6.6 m/s, and 6.5 m/s for the radial artery. This system and model offer the possibility of obtaining higher reliability PTT and PWV values yielding better monitoring of cardiovascular health measures such as blood pressure and arterial stiffness.