Effects of seizure burden on structural global brain networks in patients with unilateral hippocampal sclerosis

Background and purpose: Temporal lobe epilepsy secondary to hippocampal sclerosis is related to epileptogenic networks rather than a focal epileptogenic source. Graph-theoretical gray and white matter networks may help to identify alterations within these epileptogenic networks. Methods: Twenty-seven patients with hippocampal sclerosis and 14 controls underwent magnetic resonance imaging, including 3D-T1, fluid-attenuated inversion recovery, and diffusion tensor imaging. Subject-specific structural gray and white matter network properties (normalized path length, clustering, and small-worldness) were reconstructed. Group differences and differences between those with higher and lower seizure burden (<4 vs. ≥4 average monthly seizures in the last year) in network parameters were evaluated. Additionally, correlations between network properties and disease-related variables were calculated. Results: All patients with hippocampal sclerosis as one group did not have altered gray or white matter network properties (all p >.05). Patients with lower seizure burden had significantly lower gray matter small-worldness and normalized clustering compared to controls and those with higher seizure burden (all p <.04). A higher number of monthly seizures was significantly associated with increased gray and white matter small-worldness, indicating a more rigid network. Conclusion: Overall, there were no differences in network properties in this group of patients with hippocampal sclerosis. However, patients with lower seizure burden had significantly lower gray matter network indices, indicating a more random organization. The correlation between higher monthly seizures and a more rigid network is driven by those with higher seizure burden, who presented a more rigid network compared to those with a lower seizure burden.