Abstract
The purpose of the current study was to assess the change characteristics of structural network connection strength (SNCS), functional network connection strength (FNCS), and structural-functional correlation (SFC) of preoperative disconnected networks in temporal lobe epilepsy (TLE) and the capability of these measures to predict postoperative seizure outcomes. This study enrolled 39 TLE patients with 29 healthy controls (HCs). Patients were divided into seizure-free (SF) and non-seizure-free (NSF) groups. Each individual network was divided into four-tier subnets. FNCS was significantly lower in the 1st level (p = 0.007/p = 0.009), 2nd level (p = 0.002/p < 0.001), and 3rd level disconnect networks (p < 0.001/p < 0.001) in SF and NSF(27.0 ± 8.4, 16 men; 30.9 ± 10.9, 8 men). SNCS was significantly lower in SF than HCs (33.3 ± 8.9, 15 men) and NSF (p = 0.008/p = 0.037) in the 1st level disconnected network. Further, significant SFC in the 1st level (p = 0.026/R²=0.169), 2nd level (p = 0.002/R²=0.283), and 3rd level disconnected networks (p = 0.011/R²=0.213) in HCs. Meanwhile, the SF group showed a positive SFC at the 2nd level disconnected network (p = 0.039/R²=0.159). Our study demonstrated that the consistency in SNCS and FNCS of the regions indirectly affected by surgery is associated with lower risk of post-surgical seizure. Preoperative multimodal magnetic resonance imaging may serve as an important tool to guide surgical planning.