Fig. 1.
Our study consists of 2 parts: the development of a new methodology (i.e., the creation of a virtual population) and its clinical application [i.e., the assessment of foot-to-foot pulse wave velocity (PWV)]. By varying the cardiac and arterial parameters of the 1-dimensional (1D) model within healthy ranges, we create a set of 7,776 simulations. Rejection criteria (filter #1) are applied to eliminate nonphysiological data. Using the remaining 3,325 cases, we compute the physiological index of interest (i.e., foot-to-foot PWV) and reject 5 cases for which the PWV algorithm fails (filter #2). Peripheral and central PWV indexes are computed for each of the 3,320 cases using pressure waves measured at the dots in the Numerical model box. We also compute an index of local sensitivity analysis Īi,k that describes the effect of parameters variation on PWV values.