Quantitative comparison of two cardiac electrophysiology models using personalisation to optical and MR data

In order to translate the important modelling work into clinical tools, the selection of the best model for a given application is crucial. In this paper, we quantitatively compare personalization of two different cardiac electrophysiology models on the same dataset, in order to help such a selection. One is a phenomenological model, the Aliev-Panfilov model (1996), and the other one is a simplified ionic model, the Mitchell-Schaeffer model (2003). In the preliminary steps of model personalization, we optimise the forward problem with the determination of an optimum time integration scheme for each model, which could result in stable and accurate simulations without the use of unnecessary expensive high temporal and spatial resolutions. Next, we personalize the two models by optimising their respective parameters, to match the depolarisation and repolarisation maps obtained ex-vivo from optical imaging of large porcine healthy heart. Last, we compare the personalization results of the two different models.