Sensitivity analysis in parameter estimation of physiological systems

In the identification of physiological systems it is often of practical interest knowing which time intervals during experiments provide most information on specific model parameters. This problem is usually approached using parameter sensitivities of model outputs. However, this kind of analysis does not provide always a correct answer, especially if strong correlation among the parameters is present. For this reason, the authors consider in this paper alternative approaches based on information-theoretic criteria. They use as a case study the widely used minimal model of glucose disappearance. Their results demonstrate the usefulness of one particular approach for defining the relevant time intervals for the identification of specific parameters.