How accurately can cardiac conductivity values be determined from heart potential measurements?

Although realistic cardiac electrophysiological simulations require accurate model parameters, no fully experimentally determined sets of six cardiac conductivity values exist. The present authors have recently proposed a method to determine the six bidomain conductivity values, for the extra- and intracellular domains in the longitudinal, transverse and normal directions, from measurements of potential made in cardiac tissue in vivo. The method uses a 3D mathematical model, a microelectrode measuring array and a novel inversion technique, which retrieves the conductivities and the fibre rotation angle from the potential measurements. In this work, a number of different data analysis methods are compared for realistically large sets of these simulated potential measurements and the best method is identified. Using synthetic data it is found that the three extracellular conductivities can be retrieved extremely accurately, with relative errors of less than 5%, even with noise of up to 40% added to the potential measurements. In addition, the intracellular longitudinal conductivity and the fibre rotation can be retrieved with relative errors at worst around the added noise. The intracellular transverse and normal conductivities are often more difficult to retrieve, with relative errors of around four times the added noise.