Generalized sensitivity analysis applied to vascular refilling models

In the process of estimating parameters on inverse problems, one needs to use an ensemble of techniques to obtain sufficient information on the model parameters. One way of doing so is by using traditional sensitivity functions (TSF) to analyze the behavior of the model parameters for a given model output. However, this is of limited utility since one usually takes measurements on a noisy environment and high correlation among parameters could be present. Generalized sensitivity functions (GSF), introduced by Thomaseth and Cobelli, overcome the aforementioned limitations on parameter estimation. In this paper, we discuss the positive features and utility, as well as the shortcomings of GSF in the model analysis, and in the parameter identification and estimation of vascular refilling models.