Preoperative structural–functional coupling at the default mode network predicts surgical outcomes of temporal lobe epilepsy

Chunyao Zhou,Fangfang Xie,Dongcui Wang,Xiaoting Huang,Danni Guo,Yangsa Du,Ling Xiao,Dingyang Liu,Bo Xiao,Zhiquan Yang,Li Feng
DOI: https://doi.org/10.1111/epi.17921
IF: 6.74
2024-02-24
Epilepsia
Abstract:Objective Structural–functional coupling (SFC) has shown great promise in predicting postsurgical seizure recurrence in patients with temporal lobe epilepsy (TLE). In this study, we aimed to clarify the global alterations in SFC in TLE patients and predict their surgical outcomes using SFC features. Methods This study analyzed presurgical diffusion and functional magnetic resonance imaging data from 71 TLE patients and 48 healthy controls (HCs). TLE patients were categorized into seizure‐free (SF) and non‐seizure‐free (nSF) groups based on postsurgical recurrence. Individual functional connectivity (FC), structural connectivity (SC), and SFC were quantified at the regional and modular levels. The data were compared between the TLE and HC groups as well as among the TLE, SF, and nSF groups. The features of SFC, SC, and FC were categorized into three datasets: the modular SFC dataset, regional SFC dataset, and SC/FC dataset. Each dataset was independently integrated into a cross‐validated machine learning model to classify surgical outcomes. Results Compared with HCs, the visual and subcortical modules exhibited decoupling in TLE patients (p
clinical neurology
What problem does this paper attempt to address?