Validation of the Two-Level Approach for the Solution of the EEG Inverse Problem in an Anisotropic Realistic Head Model

The electrical activity of a neurological epileptic disorder can be determined in a noninvasive way by analyzing the electroencephalogram (EEG). The EEG source localization procedure requires the solution of an inverse problem. We implemented the two-level inverse algorithm, which combines the computational efficiency of the semi-analytical three-shell spherical model with the accuracy of the anisotropic realistic head model. This paper mainly deals with the implementation and the validation of the numerical two-level inverse algorithm using an anisotropic head model. Results obtained from the EEG source localization are compared with measurements of depth electrodes. The results show the fast localization and noise robustness of the numerical procedure.