Qualitative experiment design in physiological system identification

A qualitative experiment design problem for identifying compartmental models of physiological systems from multi-input-multi-output tracer experiments is examined. Qualitative experiment design consists of selecting the sites of inputs and outputs, among those experimentally feasible, which guarantee a priori unique identifiability of model parameters. Using this method, a minimal input-output configuration that ensures a priori unique identifiability is defined. The problem is of practical relevance when dealing with physiological systems in which severe constraints exist on experiment design for both ethical and practical reasons. Fundamentals of physiological systems, tracer experiments, compartmental models, and a priori identifiability are discussed. As a case study, a ten-compartment model of leucine metabolism and the multi-input-output tracer experiment designed for its identification are described.<>