LAUSR

Objective. Methods for separating the forward–backward components from blood pulse waves rely on simultaneously measured pressure and flow velocity from a target artery site. Modelling approaches for flow velocity simplify the wave separation analysis (WSA), providing a methodological and instrumentational advantage over the former; however, current methods are limited to the aortic site. In this work, a multi-Gaussian decomposition (MGD) modelled WSA (MGDWSA) is developed for a non-aortic site such as the carotid artery. While the model is an adaptation of the existing wave separation theory, it does not rely on the information of measured or modelled flow velocity. Approach. The proposed model decomposes the arterial pressure waveform using weighted and shifted multi-Gaussians, which are then uniquely combined to yield the forward (P F(t)) and backward (P B(t)) pressure wave. A study using the database of healthy (virtual) subjects was used to evaluate the performance of MGDWSA at the carotid artery and was compared against reference flow-based WSA methods. Main results. The MGD modelled pressure waveform yielded a root-mean-square error (RMSE) < 0.35 mmHg. Reliable forward–backward components with a group average RMSE <2.5 mmHg for P F(t) and P B(t) were obtained. When compared with the reference counterparts, the pulse pressures (ΔP F and ΔP B), as well as reflection quantification indices, showed a statistically significant strong correlation (r > 0.96, p < 0.0001) and (r > 0.83, p < 0.0001) respectively, with an insignificant (p > 0.05) bias. Significance. This study reports WSA for carotid pressure waveforms without assumptions on flow conditions. The proposed method has the potential to adapt and widen the vascular health assessment techniques incorporating pulse wave dynamics.