Surgery for medically intractable seizures is recognized to offer important benefits for appropriately elected individuals. The success of surgery depends upon the ability to localize seizure onset to a specific rain region that mad be safely rejected. Seizure localization is typically performed by visual analysis of video-EEG recordings of ,seizures. While some seizures can be sufficiently localized by scalp recordings, the onset of some seizures can be difficult to determine. In some instances leg. temporal lobe complex partial seizures) this is cause the seizures originate from deep mesial structures; recording with intracranial depth electrodes or subdural strip electrodes usually allows for good localization of these seizure onset. In some partial seizures, however, for instance those originating from the frontal lobe or lateral temporal lobe seizure localization can be difficult even when intracranial recordings from extensive subdural grids are available for analysis. Visual analysis of seizure onset often-reveals broad regional onset even when focal lesions are present. The predominant hypothesis to be tested in this proposal is that these partial seizures that are difficult to localize even with intracranial electrodes are in fact focal seizures with very rapid regional spread. Although seizure activity rapidly propagates to involve broad areas of cerebral cortex, the seizure focus is still often an important generator. Determination of these patterns of low of electrical activity, which are not apparent by visual inspection of the Eeg, can allow for localization of seizure onset. To test these hypotheses multichannel recordings of seizures obtained from intracranial electrodes will be analyzed utilizing the recently described directed transfer function (DTF) method. The directed transfer function method is a multichannel autoregressive analysis that allows for demonstration of patterns of flow of electrical activity. The DTF method will be applied to digitized recordings of seizures obtained from subdural gird and depth electrode arrays in patients who have been candidates for seizure surgery. Seizures beginning from the temporal lobe mesial and lateral) will be analyzed to determine patterns of seizure generation and propagation. Analyses will then be extended to seizures of extratemporal onset with particular attention to those seizures which begin with apparent broad regional changes. It is hoped that the applications of these methods to intracranial recordings will provide important insights into seizure generation and spread that can ultimately better guide the surgical resections in these patients.
