To address the goal of our simulation study, we carried out two types of simulations. First, we wished to determine the effect of selectively adding inhomogeneous regions to an otherwise homogeneous model. To this end, we computed potentials for each of the same 30 time instants, first with a completely homogeneous model and again after including a single inhomogeneity. Then we compared the solutions statistically with the homogeneous solution as the reference. The inhomogeneities studied did not necessarily belong to a single closed region. For example, all of the bones were tested as a single inhomogeneity. Some larger structures, such as the lungs, were considered both individually and in combination in order to evaluate the effect of similar structures in different proximities to the heart. We considered the following individual inhomogeneities: the bones, the left lung, the right lung, both lungs, subcutaneous fat, isotropic skeletal muscle, anisotropic skeletal muscle at 5 different anisotropy ratios (3:1, 5:1, 7:1, 10:1 and 15:1), the major blood vessels in the vicinity of the heart, and three epicardial fatpads, two located in the atrioventricular groove and the other one on the apex of the heart. ``Including'' an inhomogeneity actually meant assigning an appropriate conductivity value to all elements with a group number corresponding to that inhomogenetity. The rest of the tetrahedra were assigned the average isotropic conductivity, listed in Figure 3 as homogeneous torso, of .239 S/m.
The second type of simulation was a complement of the first - its goal was to evaluate the effect of removing a single inhomogeneity from an otherwise completely inhomogeneous model. In this case, the reference model, or full model, included all available inhomogeneities. The skeletal muscle in this model was assigned a 7:1 anisotropy ratio. Simulations were then run at the same time steps but with a specific region assigned the average isotropic conductivity tensor. The potentials from these simulations were compared to those from the reference model. The list of inhomogeneities removed from the full model was similar to that from the first study: subcutaneous fat, 7:1 skeletal muscle, both lungs, the bones, the major blood vessels, all the fatpads, and each individual fatpad.