Abstracts

Rationale: Epilepsy surgery planning can be complicated when intracranial EEG monitoring reveals ictal patterns with a broad regional neocortical onset, particularly if the onset region includes eloquent cortex. In some cases, however, a regional ictal pattern may arise from ictal activity that is propagated from other structures that might be safer to resect. This is most evident when a lesion is present in nearby neocortex or in mesial temporal structures, but may also occur in the absence of a lesion. Recordings from both intracranial depth electrodes and subdural electrode grids can provide crucial information about the propagation of seizure activity between mesial temporal and neocortical networks, and advanced analyses of the propagation of ictal activity between these structures and others may improve localization of the ictal onset zone when planning resection. Methods: Intracranial EEG recordings obtained during multiple seizures (3-7) were parameterized using multivariate autoregressive models (MVAR), and analyzed with short-time direct directed transfer function (SdDTF), which estimates the direction of flow between electrodes (sites), as well as the intensity, spectral content, and temporal evolution these flows. Seizures were marked at their electrographic onset, and the propagation of activity was analyzed across a wide range of frequencies (0-235 Hz). Results: SdDTF analyses showed propagation of ictal activity between a limited number of contacts in mesial temporal regions and a larger number of neocortical sites, as well as less prominent flows among the neocortical sites themselves. These findings suggested a prominent role for mesial temporal structures in seizure generation. Conclusions: For patients undergoing intracranial monitoring for epilepsy surgery, analyses of the patterns of EEG activity propagation between mesial temporal and neocortical networks may provide additional information that can improve seizure localization, in turn optimizing post-surgical outcomes and minimizing post-operative impairments. This project was funded by NINDS R01 NS40596 and NS48222.