A Preliminary Study Of The Effect Of Brain Tumor In Left Perisylvian Region On Language Network

Background: Language is one of the most important functions of human being. To protectneurological functionmaximallyis the ultimate goal of neurosurgery. So far little is known about the effect of brain tumor in left perisylvian area on brain network, especially on language network. We employed graphic theory to study the change of topological metrics of the whole brain network/language network derived from DTI tractography and investigated its association with language function.Methods: 1 .A group of patients with brain tumors in left perisylvian region (N = 33) and controls (N = 9) have been enrolled in the study. The patients were divided into two groups(LU=17, LD=16) according to preoperativelanguage function. Both deterministic and probabilistic fiber tracking methods were applied to construct brainnetwork. We used graph theory to calculate global and local metrics thencompared the difference between groups. We also used NBS analysis to detect disconnected edges. 2.We explored the longitudinal change of the whole brain / language FA networkfrom patients (N = 12) with brain tumor in left perisylvian area and investigated its association with language function.Meanwhile, we used tract-based analysis to obtain the mean FA of 5 language related tracts and explored the association with language function.Results: 1. All of the subjects showed small-world organization. We found a significant difference of the global efficiency metric ELambda and it was significantly correlated with WAB score in the deterministic network. No group difference was found for the nodal efficiency and nodal efficiency lateralization index. NBS analysis revealed some disconnected subnetworks. 2. ELambda was significantly correlated with both pre and post-operative WAB score. However, tract-based analysis revealed that the mean FA values of left SLF/ILF/IFOF were only significantly correlated with postoperative WAB score. No significant correlation was observed between the FA of language-related tractsand preoperative language function. Moreover, it showed that SLF lateralization index was significantly associated with postoperative WAB score.Conclusions: Brain tumors located in left perisylvian region not onlyaffect local brain connectivity, but also lead to change of topological organization of the whole brain/language network, which is closely associated with language function. Compared with language tractsanalysis,brain /language networkanalysismay serve as a promising tool for predicting language function in the future.