• Pierre Besson, S Kathleen Bandt, Timothée Proix, Stanislas Lagarde, Viktor K Jirsa, Jean-Philippe Ranjeva, Fabrice Bartolomei, and Maxime Guye.
• Aix Marseille Univ, CNRS, CRMBM UMR 7339, Marseille, France.
• Brain. 2017 Oct 1; 140 (10): 2639-2652.

AbstractSee Bernasconi (doi:10.1093/brain/awx229) for a scientific commentary on this article. Drug-resistant localization-related epilepsies are now recognized as network diseases. However, the exact relationship between the organization of the epileptogenic network and brain anatomy overall remains incompletely understood. To better understand this relationship, we studied structural connectivity obtained from diffusion weighted imaging in patients with epilepsy using both stereo-electroencephalography (SEEG)-determined epileptic brain regions and whole-brain analysis. High resolution structural connectivity analysis was applied in 15 patients with drug-resistant localization-related epilepsies and 36 healthy control subjects to study structural connectivity changes in epilepsy. Two different methods of structural connectivity analysis were carried out using diffusion weighted imaging, one focusing on the relationship between epileptic regions determined by SEEG investigations and one blinded to epileptic regions looking at whole-brain connectivity. First, we performed zone-based analysis comparing structural connectivity findings in patients and controls within and between SEEG-defined zones of interest. Next, we performed whole-brain structural connectivity analysis in all subjects and compared findings to the same SEEG-defined zones of interest. Finally, structural connectivity findings were correlated against clinical features. Zone-based analysis revealed no significant decreased structural connectivity within nodes of the epilepsy network at the group level, but did demonstrate significant structural connectivity differences between nodes of the epileptogenic network (regions involved in seizures generation and propagation) and the remaining of the brain in patients compared to controls. Whole-brain analyses showed a total of 133 clusters of significantly decreased structural connectivity across all patients. One cluster of significantly increased structural connectivity was identified in a single patient. Clusters of decreased structural connectivity showed topographical preference for both the salience and default mode networks despite clinical heterogeneity within our patient sample. Correlation analysis did not reveal any significant findings regarding either the effect of age at disease onset, disease duration or post-surgical outcome on structural connectivity. Taken together, this work demonstrates that structural connectivity disintegration targets distributed functional networks while sparing the epilepsy network.

22 keywords...

hide…