Abstracts

Rationale: Sphenoidal electrodes (SpE) have been shown to be helpful in detecting mesial temporal spikes. Unfortunately, including SpE in realistic EEG source imaging (ESI) is difficult, because the exact positions of these electrodes are unknown, and also because the existing ESI software usually calculate the potential only on the surface of the head and not at the locations of SpE. We now report, for the first time, the feasibility to determine the locations of SpE and to incorporate this information into ESI. In addition, the impact of including SpE on source localization in mesial temporal lobe epilepsy (MTLE) is assessed. Methods: Nine patients with refractory MTLE, based on their clinical semiology and electrophysiologic data, were included for this study. The potentials recorded at 10-20 electrodes, a pair of subtemporal electrodes (FT9/10 or F9/10) and SpE were used for ESI. The locations of the scalp electrodes were determined using a digitizer. To find the positions of SpE two orthogonal X-rays (sagittal and coronal) were taken while the scalp electrodes were attached to the head and the sphenoidal needle was inserted. An in-house computer program was then used to find the locations of the tip of the needle relative to the surface electrodes on the X-rays. This information was used to determine the 3D coordinates of the tip of the needle. A finite element method (FEM) based on the patient's MRI was used for forward solution calculations. This method allows computation of the electrical potential at the true locations of SpE. For each patient, at least 10 similar spikes were selected for averaging. Then a single fixed dipole was fitted to the measured data over the time interval starting at the midway between the baseline and the peak, and ending at the peak of each averaged spike. The SpE were then turned off and the calculations were repeated. For each patient the distance between the dipoles in presence and in absence of SpE was calculated. T-tests (at the 5% significance level) were then performed to determine if the observed differences between locations of the dipoles in two cases were significant. Results: The maximum/minimum observed distance between the two dipoles (with and without SpE) was 36.5/0.3 mm and the average was 17.9 mm. The t-tests, however, showed that only the difference in the vertical direction was significant. In fact in 78% (7) of the cases the dipole location in presence of SpE was determined to be lower than the corresponding dipole when these electrodes were turned off. In the patient with the biggest dipole shift, inclusion of SpE moved the location of the source from the right inferior insula to the right anterior temporal pole. Conclusions: We report on a technique to determine the actual locations of SpE for realistic ESI. Our results show that inclusion of SpE could significantly change ESI results in patients with MTLE, even when surface subtemporal electrodes are present.