Abstracts

Rationale: Approximately 30% of patients with epilepsy will be drug-resistant. In these individuals, we consider surgical intervention for treatment of their epilepsy. Intracranial electroencephalography (EEG) is the gold standard for identification of the epileptogenic focus, usually by identifying the seizure onset zone as well as interictal activity like high frequency oscillations, sharp waves, and spikes. However, it is likely possible to identify the epileptogenic focus on the basis of participation or non-participation in the networks that support cognition. Methods: Subjects undergoing intracranial EEG and stereo-EEG for the surgical management of epilepsy were tested on an object-in-place memory task. Analysis of the recorded activity while performing this task evaluated the pair-wise coherence of all electrodes in the patient's brain, in an attempt to identify the networks that support performance of this task. Principal component analysis (PCA) identified those electrodes and bands of coherence ?" and hence the brain networks --- that tended to co-vary during performance. Results: Electrodes with inter-ictal discharges and those electrodes within the seizure onset zone tended to have lower coefficients for coherence when subjected to PCA. This indicates that they systematically fail to participate with the networks that support cognition. Conclusions: Network participation as identified by pair-wise coherence constitutes a valuable tool for the identification of the epileptogenic focus in focal epilepsy. The observation of non-participation in networks supporting cognition has important implications for the pathophysiology of epilepsy. Funding: Leon Levy Foundation Neuroscience Fellowship, NINDS R25 Research Residency