Functional connectome contractions in temporal lobe epilepsy: Microstructural underpinnings and predictors of surgical outcome

Sara Larivière,Yifei Weng,Reinder Vos de Wael,Jessica Royer,Birgit Frauscher,Zhengge Wang,Andrea Bernasconi,Neda Bernasconi,Dewi V Schrader,Zhiqiang Zhang,Boris C Bernhardt

DOI: https://doi.org/10.1111/epi.16540

IF: 6.74

Epilepsia

Abstract:Objective: Temporal lobe epilepsy (TLE) is the most common drug-resistant epilepsy in adults. Although it is commonly related to hippocampal pathology, increasing evidence suggests structural changes beyond the mesiotemporal lobe. Functional anomalies and their link to underlying structural alterations, however, remain incompletely understood. Methods: We studied 30 drug-resistant TLE patients and 57 healthy controls using multimodal magnetic resonance imaging (MRI) analyses. All patients had histologically verified hippocampal sclerosis and underwent postoperative imaging to outline the extent of their surgical resection. Our analysis leveraged a novel resting-state functional MRI framework that parameterizes functional connectivity distance, consolidating topological and physical properties of macroscale brain networks. Functional findings were integrated with morphological and microstructural metrics, and utility for surgical outcome prediction was assessed using machine learning techniques. Results: Compared to controls, TLE patients showed connectivity distance reductions in temporoinsular and prefrontal networks, indicating topological segregation of functional networks. Testing for morphological and microstructural associations, we observed that functional connectivity contractions occurred independently from TLE-related cortical atrophy but were mediated by microstructural changes in the underlying white matter. Following our imaging study, all patients underwent an anterior temporal lobectomy as a treatment of their seizures, and postsurgical seizure outcome was determined at a follow-up at least 1 year after surgery. Using a regularized supervised machine learning paradigm with fivefold cross-validation, we demonstrated that patient-specific functional anomalies predicted postsurgical seizure outcome with 76 ± 4% accuracy, outperforming classifiers operating on clinical and structural imaging features. Significance: Our findings suggest connectivity distance contractions as a macroscale substrate of TLE. Functional topological isolation may represent a microstructurally mediated network mechanism that tilts the balance toward epileptogenesis in affected networks and that may assist in patient-specific surgical prognostication.