Electroencephalographic source localization (ESL) is a highly valuable tool in the presurgical evaluation of epilepsy. For the accurate estimation of sources from the electroencephalogram (EEG), the use of realistic and individual head models has become increasingly important1. In realistic volume conductor modeling, the skull plays an important role due to its complex geometry and low conductivity compared to the other tissues inside the head2. In this study, we analyzed the influence of different skull modeling approaches on ESL, using data of six operated epileptic patients. For these patients, high-resolution CT images, pre- and postoperative MRI as well as 27-channel EEG with marked interictal epileptiform discharges, were available. These matched data sets allowed us to analyze the influence of using the skull geometry segmented from CT compared to the geometry based on MRI or on a CT_template3 on the source localization procedure. Four realistic head models with different skull compartments, based on finite difference methods, were constructed for each patient: (i) Three models contained skulls with compact and spongy bone compartments as well as air-filled cavities, and (ii) one model included a single compartment MRI-based skull. We investigated dipole estimation using single versus averaged spike analysis. Furthermore, to analyze at which time point the dipole estimations laid closer to the resection, ESL was performed at the half-rising phase and peak of the spike. The estimated sources for each model were validated against the resection, as indicated by the postoperative MRI. We obtained a mean distance to the resection over all patients of ~10 mm, using the averaged spike. Moreover, single spike analysis was highly influenced by SNR, yielding better estimations at the spike peak. Although averaging reduced the SNR effects, it did not always result in dipole estimations lying closer to the resection. Despite the use of low SNR 27-channel EEG, our skull modeling approaches did not lead to significant differences in the localization of the irritative zone. The proposed head models improved the accuracy of dipole estimation in the temporal lobe by considering the complex skull base geometry. Therefore, all models can be used in the presurgical evaluation of epilepsy. References: 1. Vorwerk, J. et al. (2014) A guideline for head volume conductor modeling in EEG and MEG. NeuroImage 100:590–607. 2. Montes-Restrepo, V. et al. (2014) Influence of skull modeling approaches on EEG source localization. Brain Topography 27:95–111. 3. Rorden, C. et al. (2012) Age-specific CT and MRI templates for spatial normalization. NeuroImage 61:957–65.