1000 Propagation of Seizures Throughout the Brain can be Predicted With Patient Specific Seizure Propagation Pathways

INTRODUCTION: The propagation of seizures throughout the brain is well-characterized at the cellular level, though the propagation at the network and whole brain level remains incompletely understood. To this end, the resting state connectivity and structural connectivity of the brain have been characterized and correlated with SEEG investigations and MRI imaging. METHODS: We collated ictal and inter-ictal data from eighteen patients undergoing SEEG investigations for seizure onset zone localization. For each investigation, the location of SEEG contacts were identified in patient space. The ictal onset at each SEEG channel was determined with the creation of a permutation entropy model. Using patient specific diffusion imaging, the probability of axonal connection and the white matter path length between each SEEG contact identified. In addition, the Euclidean path distance between each SEEG contact was identified as well. Linear correlations were identified between time of ictal onsets and the three propagation pathways metrics. RESULTS: SEEG investigation from 18 patients, with 94 distinct seizure events were analyzed. All patients and seizure events demonstrated good correlation between ictal onset time and 1) probability of connection, 2) path length, and 3) Euclidean distance between the clinically determined seizure onset zone contacts and the remaining, non-SOZ contacts. The correlations between the ictal onset times and the different propagation pathways ranged from ∼0.5 to ∼0.8 based on patient and seizure event. The dominant propagation pathway, as determined by the pathway with best correlation, differed between patients and seizure events. CONCLUSIONS: The propagation of seizures throughout the brain can be in part predicted by patient specific seizure propagation pathways. Patient and seizure specific differences may suggest variability in seizure propagation modalities.