High resolution boundary element forward models for sensor arrays: an application to electroencephalography

The forward problem of electroencephalography (EEG) involves mapping current sources in the brain to voltages at sensors (electrodes) on the scalp. Boundary element modeling is a numerical method which has been widely applied to this problem. Here, the boundary element approach is validated for spherical models for which analytic solutions can be calculated. Three aspects of the basic approach were varied: the type of potential approximation (de Munck, J.C., IEEE Trans. Biomed. Eng., vol.39, no.9, p.986-90, 1992); the use of a solid angle correction (Heller, L., Digital Image Synthesis and Inverse Optics, Proc. SPIE, vol.1351, p.376-90, 1990); the isolated skull correction (Hamalainen, M.S. and Sarvas, J., IEEE Trans. Biomed. Eng., vol.36, no.2, p.165-71, 1989). The effects of locally non-uniform meshes were also considered. The calculations were performed using six different mesh resolutions, including mesh resolutions significantly higher than those previously used for forward model evaluation.