Abstracts

Rationale: Despite the advances in imaging and electrophysiology, surgical outcome in non-lesional temporal lobe epilepsy remains modest. We reasoned that an analysis of functional connectivity from intracranial EEG would improve the localization of the epileptogenic zone and hence, the surgical outcome. Measures of functional connectivity have not yet been directly applied in predicting seizure outcome after epilepsy surgery Methods: All patients with intractable temporal lobe epilepsy who underwent temporal lobectomy between January 2009 and January 2012 after an invasive EEG evaluation using stereo EEG with at least 6 months of follow up (n=15) were included in the study. The temporal lobe was parcellated into 9 standardized regions and one electrode contact from each region was selected for analysis of EEG signals. EEG was sampled at 0.5 kHz and filtered into 7 different bands: alpha, beta, theta, delta, gamma, and High-Frequency-Oscillations (HFO). Functional connectivity was determined using a novel covariance-based measure, defined as follows. For each pair of recording channels, the maximum absolute value of the cross-correlogram was taken as the value of functional connectivity between the pair. The time lags of the cross-correlogram ranged from -100 to +100 ms. Each cross-correlogram was normalized so that at zero lag its value equals the cross-correlation coefficient of the pair. The functional connectivity values are thus bounded between 0 and 1. We investigated the entire matrix of channel pairs over all of the frequency bands and compared them with seizure recurrence after surgery. A favorable outcome was defined as complete seizure-freedom since surgery Results: Eight patients (53.3 %) had recurrence of seizures within 6 months of temporal lobectomy, including two whose seizures occurred in the setting of missed antiepileptic medications and sleep deprivation, and two who had persistence of their typical pre-operative auras. The high rate of seizure recurrence reflects the expected low rates of seizure-freedom in patients whose non-lesional, poorly localized, intractable epilepsy necessitates evaluation using invasive electrodes. Functional connectivity as defined above was higher in the temporal regions in 5/7 patients who were seizure free after surgery, while four of the 8 patients who had recurrence of seizures had lower functional connectivity among temporal regions. This suggests that our measure of functional connectivity may be a good parameter to find the epileptogenic zone. Indeed, a connectivity value greater than 0.5 in the beta band predicted well the surgical outcome with a sensitivity and specificity of 71.4 % and 50 % respectively for prediction of favorable seizure outcomes. Conclusions: These preliminary results support the potential for our novel measure of intracranial EEG functional connectivity in predicting surgical outcomes after temporal lobe epilepsy surgery. Prospective studies with a larger patient sample are needed to determine the applicability of this measure as a diagnostic tool