Abstracts

Rationale: Successful surgical outcome for patients with intractable partial epilepsy and no radiological abnormalities rests entirely on the localization of the epileptogenic focus during the intracranial implant stage. In our experience, 40-50% of these patients fail localization and are explanted without surgery. In our series of 29 nonlesional frontal neocortical implant patients, we attempted to identify electrographic seizure onsets that are predictive of successful localization, and ultimately, of excellent post-resection prognosis.Methods: 40 consecutive Mayo Clinic patients implanted for cryptogenic epilepsy of frontal lobe semiology were selected. 11 patients were excluded because of either subtle MRI abnormalities discovered retrospectively, or due to non-frontal localization. In the remaining 29 patients, spatiotemporal, spectral, and morphological patterns of seizure onset were identified and evaluated against the endpoints of either “success” or “failure” of localization of seizure focus as judged by expert Epileptologists. Statistical analysis was used to quantitate the spatial extent of the seizure onset zone, and spectral analysis was used to verify visual approximation of the initial discharge frequencies.Results: In 15 patients with successful localization, a characteristic pattern of onset was seen in 10, consisting of repetitive spiking (RS) followed by a slow wave (SW) evolving into high-frequency-low voltage (HFLV) activity lasting up to 10 seconds, after which either regional rhythmic hypersynchronous (RH) activity or periodic spiking (PS) heralded either generalization or spontaneous termination of a seizure. In the remaining 5 patients, there was a variable absence of either SW or RH, or the SW was diffuse, but focal HFLV activity was preserved. In 14 patients with failed localization, this pattern was not seen, or seen multifocally, with RH, RS or PS activity at clinical seizure onset. The patients in the “localized” cohort also had fewer electrodes involved in the discharge at onset (1-3), with the same electrodes being most active interictally, and had excellent surgical outcomes, with pathology revealing cortical dysplasia or gliosis. The “failed localization” patients had multiple electrodes involved at onset, less of a spatial concordance of interictal discharges with seizure onset activity, and worse outcomes in 3 cases where surgery was performed.Conclusions: Successful intracranial seizure localization is associated with an onset pattern featuring HFLV activity in the interictally “hot” electrodes. Other types of onset are not predictive of localization, and may represent seizure spread patterns. Focal HFLV onsets are highly associated with good outcomes and with cortical dysplasia. Therefore, the absence of an HFLV onset in an intracranial recording of a nonlesional FLE patient’s typical seizure should raise concern about the location of the implant with respect to the epileptogenic zone, and possibly warrant reimplantation, in order to avoid localization failure and the medical and emotional burdens it carries for the patient.