Abstracts

Rationale: The most important aspect of successful epilepsy surgery is the accurate and complete identification of the ictal onset region. Research has shown that the placement of subdural electrodes combined with video EEG monitoring (Phase 2 monitoring) is one of the best ways to define the extents of the ictal onset region. The subdural EEG recordings are interpreted by epileptologists to identify which contacts are involved in the onset and spread of the seizures. The present study investigates the utility of dipolar source estimation as a tool to aid epileptologists in the identification of the ictal region.Methods: Three patients who underwent Phase 2 monitoring in the EMU at Dell Children’s Medical center were selected to participate. The age range was from four years to 15 years of age. Following implantation of the intracranial EEG (icEEG) electrodes, a CT scan was done to identify the locations of each electrode contact. The CT was co-registered with the patient’s structural MRI and the three dimension coordinates of the electrode placements were identified. A three dimensional model was generated with an automated algorithm to identify grey matter. A high resolution boundary element model (BEM) was generated for dipole localization of electrocortical fields. The number of subdural contacts ranged from 48 to 72. One patient had bilateral subdural grid placements (72 contacts), one patient had left hemisphere grid placement, and the other patient had right hemisphere grid placement. The two patients with unilateral subdural contacts also had depth electrodes placed. The icEEG was digitized at 512 Hz and filtered from 1 to 70 Hz. Seizure onsets were identified by an epileptologist and an equivalent current dipole source estimate was computed and projected onto the patient’s MRI and the 3D model generated from the MRI.Results: Seizures were captured in all three patients. The number of seizures per patient ranged from four to 53 events. Seizure onset was determined by an epileptologist and the icEEG contacts involved were identified. For all three patients, the dipole source estimates for seizure onset correlated with the ictal onset region identified using the icEEG. Additionally, the dipolar source estimates were correlated with interictal discharge localization based on the icEEC. Post-surgically, the dipolar source estimates were correlated with the region of the resection in all cases.Conclusions: This study demonstrates that it is possible to generate accurate 3D source estimates from intracranial EEG waveforms. The correlation between the icEEG, the dipolar source estimates, and the surgical region was very high for all patients and suggests that dipolar source estimates projected onto a 3D model of the patient’s brain can be used in surgical planning during Phase 2 evaluations in the EMU.