The Functional Reorganization of Language Network Modules in Glioma Patients: New Insights From Resting State fMRI Study

Abstract

BackgroundPrior investigations of language functions have focused on the response profiles of particular brain regions. However, the specialized and static view of language processing does not explain numerous observations of functional recovery following brain surgery. To investigate the dynamic alterations of functional connectivity (FC) within language network (LN) in glioma patients, we explored a new flexible model based on the neuroscientific hypothesis of core-periphery organization in LN.MethodsGroup-level LN mapping was determined from 109 glioma patients and forty-two healthy controls (HCs) using independent component analysis (ICA). FC and mean network connectivity (mNC: l/rFCw, FCb, and FCg) were compared between patients and HCs. Correlations between mNC and tumor volume (TV) were calculated.ResultsWe identified ten separate LN modules from ICA. Compared to HCs, glioma patients showed a significant reduction in language network functional connectivity (LNFC), with a distinct pattern modulated by tumor position. Left hemisphere gliomas had a broader impact on FC than right hemisphere gliomas, with more reduced edges away from tumor sites (p=0.011). mNC analysis revealed a significant reduction in all indicators of FC except for lFCw in right hemisphere gliomas. These alterations were associated with TV in a double correlative relationship depending on the tumor position across hemispheres.ConclusionOur findings emphasize the importance of considering the modulatory effects of core-periphery mechanisms from a network perspective. Preoperative evaluation of changes in LN caused by gliomas could provide the surgeon a reference to optimize resection while maintaining functional balance.

What Can Resting-State fMRI Data Analysis Explain about the Functional Brain Connectivity in Glioma Patients?

“…In particular, the first applications of deep learning methods have been showing that such an approach can achieve excellent outcomes for the specific purpose of reconstructing tb-fMRI-like networks, starting from the rs-fMRI data. Nevertheless, it is also significant to note that ICA is already widely acknowledged, so much so that it is also used to isolate the functional networks in studies that do not focus on their identification but rather aim at characterizing their properties in terms of FC [29][30][31][32].…”

“…The language network is reported to undergo an overall reduction in the FC [29] and network strength [42] in patients affected by gliomas, which was also observed to be associated with the tumor grade [42]. Interestingly, such alterations showed differences based on whether the lesion was located in the left or the right hemisphere: in the former case, the FC reduction emerged within both the hemispheres bilaterally [29], in particular between the Broca and Wernicke areas [43], whereas in the latter case, significantly lower values were observed only in the ipsilateral hemisphere [29]. In contrast to this, another research study, focusing on temporal gliomas, found that the LN only showed lower integration and higher segregation properties in the case of right-sided tumors, whereas those in the left mainly affected the visual network (VN) [44,56].…”

The dynamic functional connectivity fingerprint of high-grade gliomas

“…In studies investigating the LANG, an overall decrease of rs-FC 14,15 as well as network strength 16 in patients compared to healthy controls was reported.…”

“…While most of these studies have focused their attention on speci c RSNs, such as language network (LANG) [14][15][16] , default-mode network (DMN) 12,[17][18][19][20][21] and fronto-parietal network (FPN) [16][17][18] , only a few number of papers have investigated these alterations at the whole-brain level 19,[22][23][24][25][26][27] .…”

Resting-State Electroencephalography Functional Connectivity Networks Relate to Pre- and Postoperative Language Functioning in Low-Grade Glioma and Meningioma Patients

“…Resting state fMRI on the other hand is a technique, believed to depend on the neurovascular uncoupling causing changes in neurometabolites. Language network studies done using resting state (rs) fMRI have shown differences in the network organization before and after surgery in glioma patients ( Deverdun et al, 2020 ), and also, there is literature on the extent of the network involved in language function even in the contralateral hemisphere although impacting the core network to a lesser extent ( Jin et al, 2021 ). MEG studies has reported that FC network alterations in the theta, alpha, and gamma bands are associated with poorer performance on a verbal fluency test in glioma patients ( Van Dellen et al, 2012 ; Hu et al, 2013 ; Derks et al, 2019 ).…”