Comparison of monodomain and bidomain models for whole-heart propagation studies
Mark Potse, Bruno Dubé, and Alain Vinet
In The 33rd International Congress on Electrocardiology,
Cologne, June 2006.


The work presented in this paper was originally directed by Dr Ramesh M. Gulrajani, who died unexpectedly before this paper was drafted. We gratefully dedicate this paper to the memory of Dr Gulrajani.


poster in PNG format (2693×3969)


A more comprehensive version of this paper was published in IEEE Trans Biomed Eng, December 2006.


question  The bidomain model is the most realistic mathematical expression for macroscopic simulation of cardiac muscle. However, it is computationally much more demanding than the less realistic monodomain model. A bidomain model of the human heart depends on expensive supercomputers to run, while a monodomain model can work on a standard PC. We have investigated if a monodomain model suffices for propagation studies.

methods  We developed a bidomain reaction-diffusion model of the electrical activity of the human heart, incorporating a realistic cardiac anatomy, anisotropic ventricles with transmural fiber rotation, ventricular blood, and a recent ionic model for the human ventricular myocyte. A 0.2-mm grid was used to discretize the equations, leading to 50 million nodes. The model can also operate in monodomain mode, and as a separate forward model to compute extracellular potentials (Ve) from given membrane potentials (Vm). This allows a fair comparison of monodomain and bidomain results.

results  The differences between the monodomain and bidomain models in propagating Vm and Ve were very small. Propagation was 5% faster in a bidomain model. Simulated Vm were not significantly different. Differences in Ve were very small compared to differences due to the inclusion of intracavitary blood.

conclusion  Monodomain models can be used to compute propagating action potentials in simulations that do not involve applied currents. A separate forward model can be used to compute highly realistic Ve from the simulated Vm. Although the Ve distribution implicit to the monodomain formulation differs much from that produced by a bidomain model, the difference is a smooth function that contributes little to the dynamics of Vm. Consequently, the monodomain model produces an appropriate distribution of membrane currents to calculate Ve afterward.


This work was supported by the Natural Sciences and Engineering Research Council of Canada. The work of M. Potse was supported in part by FRSQ, Québec, and in part by The Netherlands Organization for Scientific Research (NWO). Computational resources for this work were provided by the Réseau québécois de calcul de haute performance (RQCHP).