Abstracts

Rationale: Stereo-electroencephalography (SEEG) can assist clinicians in delineating brain areas generating epileptic activity. However, the placement of electrodes based on non-invasive imaging methods and clinical assessment may miss epileptogenic zones. When implanted SEEG electrodes cannot provide confident localization, the risk of seizure recurrence increases. One approach to address this challenge is implanting additional electrodes at the risk of increased surgical complications. Here, we address this challenge by localizing the ictal activity in epilepsy patients with distributed source modeling of SEEG. Specifically, we evaluated whether the distributed source modeling can still disclose the same ictal onset zones when measurements from the electrode’s vicinity were synthetically removed.

Methods: This study was approved by the Institute Review Board of Taipei Veteran General Hospital. Two medically refractory epilepsy patients gave written informed consent before participating in this study. Patient 1 had nine electrodes implanted at bilateral temporal lobes. Patient 2 had 13 electrodes distributed at the right frontal and parietal cortices. Each electrode (5-mm spacing between contact centers; Ad-Tech, Racine, WI, USA) had 6 or 8 contacts. SEEG data were sampled at 2,048 Hz. T1-weighted MRI was collected before and after electrode implantation. Episodes of ictal onsets were annotated by clinicians. Noise-normalized distributed source modeling of SEEG data based on individual patients’ brain anatomy was calculated at each time point using the minimum-norm estimates.

Results: The estimated neural current distribution for Patient 1 using all SEEG electrodes revealed ictal discharges at the right dorsomedial cingulate cortex, enclosing the contacts of electrodes H and G (Figure 1). After removing either electrode G, electrode H, or both electrodes, the source modeling still suggested significant neural currents at the dorsomedial cingulate cortex.

Source modeling on all ictal SEEG data from Patient 2 showed significantly increased neural current starting at the right hippocampus (Figure 2). Without using the data from the electrode inserted to the right hippocampal head (electrode RMT), the source modeling still reported similar focal ictal activity at the right hippocampus but with a lower amplitude.

Both patients became seizure-free after the brain resection at the right dorsomedial cingulate cortex (Patient 1) and right hippocampus and temporal pole (Patient 2). Resection brain regions overlapped areas with significant neuronal current estimates in both patients.

Conclusions: Our results suggested that distributed source modeling can provide accurate and sensitive localization of ictal activity, even when SEEG electrodes were not at the center of the epileptogenic zones. Surgical outcomes validated the source modeling results. In summary, distributed source modeling of SEEG can confidently estimate the ictal onset zone without additional electrode implantation to reduce surgical complications.

Funding: Please list any funding that was received in support of this abstract.: Natural Sciences and Engineering Research Council of Canada (NSERC; RGPIN-2020-05927); Canada Foundation for Innovation (38913 and 41351); MITACS (IT25405); Academy of Finland (298131).