Objective. To measure the impact of skull-to-brain conductivity ratios on interictal spikes source localizations, using high resolution EEG (HR EEG). In previous studies, two ratios were mainly employed: 1/80 and 1/40. Consequences of the employed ratios on source localization results are poorly studied. Methods. Twenty patients with drug-resistant epilepsy were studied using HR EEG (sixty-four scalp electrodes). For each patient, three-layers realistic head models based on individual MRI were elaborated using boundary element model. For each interictal spike, source localization was performed six times, using six skull-to-brain conductivity ratios (1/80, 1/50, 1/40, 1/30, 1/20 and 1/10), exploring all the spectrum of values reported in the literature. We then measured distances between the different sources obtained and between the sources and the anterior commissure (in order to estimate sources depth). Results. We measured a mean distance of 5.3 mm (sd: 3 mm) between the sources obtained with 1/40 versus 1/80 ratio. This distance increased when the discrepancy between the two evaluated ratios increased. We measured a mean distance of 14.2 mm (sd: 4.9 mm) between sources obtained with 1/10 ratio versus 1/80 ratio. Sources localized using 1/40 ratio were 4.3 mm closer to the anterior commissure than sources localized using 1/80 ratio. Significance. Skull-to-brain conductivity ratio is an often-neglected parameter in source localization studies. The different ratios mainly used in the litterature (1/80 and 1/40) lead to significant differences in source localizations. These variations mainly occur in source depth. A more accurate estimation of skull-to-brain conductivity is needed to increase source localization accuracy. Abbreviations. ECD: equivalent current dipole; EIT: electric impedance tomography, HR EEG: High resolution Electroencephalography, IIS: Inter ictal spikes, MEG: Magnetoencephalography, MRI: Magnetic resonance imaging, mS/m: milli-Siemens/m, S/m: Siemens/m, SD: Standard deviation.

客观。使用高分辨率脑电图 (HR EEG) 测量颅脑电导率比对发作间期尖峰源定位的影响。在以往的研究中，主要采用两个比率：1/80 和 1/40。使用比率对源定位结果的影响研究很少。方法. 使用 HR EEG（64 个头皮电极）研究了 20 名耐药性癫痫患者。对于每位患者，使用边界元模型详细阐述了基于个体 MRI 的三层真实头部模型。对于每个发作间期尖峰，源定位进行了六次，使用六个颅脑电导率比（1/80、1/50、1/40、1/30、1/20 和 1/10），探索所有文献报道的数值范围。然后，我们测量了获得的不同源之间以及源与前连合之间的距离（以估计源深度）。结果. 我们测量了以 1/40 与 1/80 比率获得的源之间的平均距离为 5.3 毫米（标准差：3 毫米）。当两个评估比率之间的差异增加时，这个距离就会增加。我们测量了以 1/10 比率与 1/80 比率获得的源之间的平均距离为 14.2 毫米（标准差：4.9 毫米）。使用 1/40 比率定位的源比使用 1/80 比率定位的源更接近前连合 4.3 毫米。意义. 颅脑电导率是源定位研究中经常被忽视的参数。文献中主要使用的不同比率（1/80 和 1/40）导致源定位的显着差异。这些变化主要发生在源深度。需要更准确地估计颅脑电导率以提高源定位精度。缩写。ECD：等效电流偶极子；EIT：电阻抗断层扫描，HR EEG：高分辨率脑电图，IIS：发作间期尖峰，MEG：脑磁图，MRI：磁共振成像，mS/m：milli-Siemens/m，S/m：Siemens/m，SD：标准偏差。