Background: We described a model allowing in vivo non-linear mechanical characterization of human common carotid arteries (CCAs) (Masson 2011). It relied on 14 different parameters corresponding to geometric properties (number 1 to 3), fibrillar constituents (n. 4 to 9), perivascular parameters (n. 10 and 11) and contractile properties (n. 12 to 14). Because of the non-linearity and the high number of parameters of the objective function, convergence towards an optimal solution is difficult. We propose here to quantify the contribution of each parameter for optimizing the model.

Methods: We studied 58 subjects. The 14 parameters were first determined using Masson’s method, then each parameter was independently changed by +/− 1, 5 and 15%. Changes in objective function were computed and compared. Parameters were ranked.

Results: Geometric parameters contributed the most (rank 1, 2 and 3), followed by contractile components. The elastic component parameters contributed the least to the energy function. The ranking depended on the amplitude of imposed variation, especially for geometrical parameters. Some fibrillar parameters (7 and 9) had marked contribution for negative imposed change (rank 2 and 4) but not for positive one (rank 13 and 14). Convergence of modelling was more reliably obtained when using stepwise procedures based on the ranking than with standard least-square procedures.

Conclusion: The contribution of parameters in large artery energy function is unequal and stepwise introduction of parameters improves the optimization procedure. Reduction in the number of parameters might be made possible by a smart selection of the parameters.