Purpose: Source localization based on EEG/MEG data is a widely used technique to investigate neuronal activity. It tries to localize focal sources (dipoles) in order to represent the external measured signal (EEG/MEG) solving an inverse problem. Thereby the accuracy of the results depends on the given volume conductor model. Volume conductor modeling using the Finite Element Method (FEM) opens the possibility of taking into account the anisotropic conductivity of, e.g., the white matter tracts. In our study we investigated the influence of this anisotropy on solving the forward and inverse problem using an animal model. Material: Using a T1-weighted MR image, we segmented the head of a rabbit into four different tissue layers (skin, skull, grey and white matter). Additionally, we performed diffusion-tensors imaging to obtain the anisotropy of the white matter tissue. The orientation of the diffusion tensors was used to model anisotropic conductivity tensors in the white matter of the rabbit brain employing an adopted anisotropic ratio of 1:10. 650 dipoles in the cortical region in intervals of 1mm and radial orientation served as sources. Using the anisotropic model we performed EEG simulation to asses EEG potentials at 100 electrodes placed on the rabbit head. With the computed potentials we performed source localization using the same model but with isotropic conductivity tensors. This corresponds to the assumption of a source localization with an isotropic model and realistic measurement data, which include effects of anisotropic conductivity. Results: All dipoles were shifted in their location and changed in their orientation due to the isotropic model with data derived from the anisotropic model. The shift was up to 2mm with a mean of 0.69mm. The averaged orientation deviation was 23.7 degrees and the mean magnitude change of the dipole was determined with a value of 24.2 percent. Conclusion: In this study we have shown the influence of conductivity anisotropy on EEG source simulation and localization with the help of an FEM model of a rabbit head. Volume conductor modeling in EEG source localization procedures including anisotropy will improve accuracy of the localization results.
