Robust calculations of the inverse electrocardiographic problem may require accurate specification of boundary conditions at the torso and cardiac surfaces. In particular, the numerical specification of the no-flux condition on the torso is difficult because surface normals must be computed, and because the torso may alternatively be considered infinitely far away from the heart. Using the method of fundamental solutions (MFS) proposed in [1], this boundary conditions can be taken into account in different manners. Specifically, the no-flux condition on the torso can be ignored, or weighted with respect to the Dirichlet boundary condition associated to the torso data, or can be strongly enforced through a saddle-point problem. In this article, we compare these approaches. In this work we provide a preliminary comparison of these different strategies.

This study received financial support from the French Government as part of the “Investissements d’avenir” program managed by the National Research Agency (ANR), Grant reference ANR-10-IAHU-04 and from the Conseil R´egional Aquitaine as part of the project “Assimilation de donn´ees en canc´erologie et cardiologie”.