Numerical model of electrical potential within the human head

A realistic subject-speci c human head model was constructed based on structural magnetic resonance imaging (sMRI) data. Electrical conductivities were assigned inhomogeneously according to tissue type and variability within each head segment. A three-dimensional (3D) nite-di erence method (FDM) was used to compute the evolution of the electrical potential from a single electrical dipole within the brain. The Douglas–Rachford FDM and three versions of iterative FDM were tested on a threelayer concentric sphere model. The successive over-relaxation (SOR) iterative method showed the best convergence properties and hence was used to compute the electrical potential within a realistic head model. The e ect of using inhomogeneous rather than homogeneous conductivities within head segments of this model was shown to be important