Abstracts

Rationale: Our group has described a measure of EEG synchrony, the Synchrony Index (SI), which measures the intensity of local coherence of two regions and their interregional interactions. The SI has been shown to increase in the region of focal seizure onset, and to distinguish between clinical and subclinical seizures. Prior analyses required selection of the involved electrodes by visual analysis. We test the hypothesis that measures of maximal synchrony can objectively identify and describe seizure foci independently from visual analysis. Methods: We analyzed intracranial EEG data (bilateral longitudinal hippocampal depth, bilateral subdural frontal and temporal electrodes) of seven seizures from two patients who underwent temporal lobectomy for refractory mesial temporal lobe epilepsy. SI was iteratively calculated for all electrode pairs in 1 second windows for 30 seconds starting at the time of visual electrographic seizure onset. A time series was thus created demonstrating the maximum SI value (SImax) for each electrode for each second, and the electrode with which it was generated. For each (a) single electrode and (b) electrode pair, we calculated two parameters: (1) the average SImax and (2) the percentage of time with maximal SImax. We compared the electrodes identified by these SImax parameters with the location of seizure onset identified by standard visual analysis. Results: Two subjects were evaluated with a total of seven hippocampal onset complex partial seizures. One subject had post-operative Engel class Id, and the other Engel Class II, after greater than 3 years of follow-up. The accuracy of localization by the SImax parameters was strikingly different between the two subjects. For the subject with seizure remission, each electrode pair calculation agreed with visual analysis for 4 of 4 seizures. The single electrode parameters were less accurate, each identifying the site of onset for only 1 of 4 seizures. For the subject without seizure remission, none of the SImax calculations ever correctly localized seizure onset (see Table). Conclusions: In this preliminary study, SImax correlates with visual identification of seizure onset in a subject with excellent post-surgical outcome but not in a subject with residual disabling seizures. Concordance of visual localization and maximum synchrony between pairs of electrodes may predict post-surgical outcome for patients with mesial temporal lobe epilepsy. Further validation in a larger series of subjects is planned.