Abstracts

Rationale: Patients suffering from pharmaco-resistant epilepsy may undergo intracranial implantation of stereo-EEG electrodes to localize epileptogenic tissue. Identification of electrodes adjacent to grey matter is critical, as these are most likely to delineate the seizure onset zone.Whereas electrode localization based on MRI and CT can be very helpful, these imaging methods may introduce subtle inaccuracies due to MRI field inhomogeneity or image co-registration. In the present study we propose to characterize electrophysiological differences between white and grey matter, with the aim of improving the ability to identify the tissue type surrounding electrode. Methods: The 3-dimensional location of each S-EEG electrode was determined from the post-op CT and pre-op MRI and each S-EEG electrode was classified as belonging to grey or white matter.S-EEG activity was compared across three patient states: asleep, watching TV, or involved in an active discussion. In addition to analysis of local measures such as signal variance and power spectrum, functional connectivity indices were also computed between all electrode pairs (amplitude correlation and phase-locking). Finally, cortico-cortical evoked potentials (CCEPs) were measured between all electrode pairs to provide a complementary directed functional connectivity measure. Results: First, analysis of the raw signal showed that activity recorded in grey matter presented larger variance than in white matter, independently of patient state. Furthermore, signal variance was found to be negatively correlated with patient state. Second, spectral analysis revealed difference in power between white and grey matter. When patients were sleeping, power in high frequency bands was stronger in the grey matter. When the patients were awake this difference was observed for lower frequencies as well. Next functional connectivity measures revealed that amplitude correlation and phase synchrony were stronger between electrodes located in white than in grey matter. Further analysis showed that amplitude correlation was more prominent within white or grey matter than between them. Interestingly, it appeared that phase synchrony tended to be more modulated across patient states than amplitude correlation but only for electrodes pairs located in grey matter. Finally, analysis of CCEPs demonstrated that white matter presents larger in-degree and out-degree values compared to grey matter. That is, when electrically stimulated, electrodes located in white matter were more likely to elicit an evoked response at another recorded site. Conclusions: Analysis of intracranial electrophysiological signals revealed several differences between white matter and grey matter. From a clinical perspective, we determined that signal variance can be used to discriminate whether electrodes are recording primarily grey or white matter. In addition, our functional connectivity results suggest that electrodes located in white matter show stronger functional connectivity than the ones in grey matter, and that in grey matter phase synchronization differentiates between activity states at the network level.