P06.02.B THE ASSOCIATION BETWEEN STRUCTURE-FUNCTION MULTILAYER NETWORKS AND EXECUTIVE FUNCTIONING IN GLIOMA

Abstract BACKGROUND Glioma patients often experience deficits in executive functioning (EF) that cannot simply be explained by tumor location. Network approaches have provided tools to investigate brain-wide disruptions to connectivity, providing novel insights into cognition in glioma. However, these have often been unimodal approaches, focusing either on structural (diffusion MRI) or functional (resting-state fMRI) connectivity, overlooking potentially cognitively relevant structure-function interactions. Multilayer network theory offers a framework to integrate structural and functional connectivity, and extract relevant multimodal network properties. Examples are multilayer network integration at the local (regional) and whole-brain level, which are thought to reflect the integrative role of particular areas of the brain in organizing information transfer, and the overall organization of cognitive processing across the entire brain, respectively. Here, we investigated the relationship between these multilayer network properties and EF in newly diagnosed glioma patients. METHODS We constructed patient-specific multilayer networks (n=41) incorporating structure-function connectivity based on dMRI and fMRI, and computed multilayer centrality (a measure of local network integration) of the EF-relevant frontoparietal network (FPN), normalizing values against matched controls (n=39). Additionally, whole-brain multilayer eccentricity (a measure of overall network integration) was assessed. These metrics were correlated with EF, taking relevant covariates such as tumor overlap with the FPN, age and molecular subtype into account. RESULTS Contrary to our hypothesis, lower FPN centrality did not correlate with poorer EF. However, higher whole-brain multilayer eccentricity was significantly associated with poorer EF (β=-0.375, p=0.018), particularly among patients with IDH-mutant gliomas (n=25, β=-0.440, p=0.028) but not IDH-WT gliomas (n=16, β=-0.287, p=0.281). CONCLUSION The overall organization of the brain’s multilayer network reflects EF in glioma patients. More specifically, in IDH-mutant glioma patients, multilayer networks with lower overall integration associate with poorer EF, hypothetically reflecting the longer distance that information needs to travel across the combined structure-function network. This could indicate that greater reorganization of the global brain network happens in these slower growing tumors relative to IDH-wildtype glioblastomas.