Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Language Network Brain Connectivity and Social Communication in Autism Spectrum Disorder
Poster Session D, Saturday, September 13, 5:00 - 6:30 pm, Field House
Fariea Bakul1, Elizabeth Valles-Capetillo1, Rajesh K. Kana1; 1The University of Alabama at Birmingham
Introduction: Autism spectrum disorder (ASD) is characterized by core challenges in social interaction and communication, often assessed using instruments such as the Social Responsiveness Scale (SRS)1. While language ability is an integral part of social communication, studies have not often addressed the relationship between the brain’s language network functioning and social communication. Several previous studies have reported uncorrected or non-replicable correlations highlighting the need for stratified, network-specific approaches2. In this study, we leveraged data from the Autism Brain Imaging Data Exchange (ABIDE-I) repository to examine how the brain’s functional connectivity within three core networks, Reading Network (RN), Language Network (LN), and Pragmatic Network (PN), affected either elevated (indicative of pronounced social-communication challenges) or diminished (indicative of fewer such challenges) SRS scores3. Methods: ABIDE I resting-state fMRI data from 704 participants were preprocessed using CPAC (Configurable Pipeline for the Analysis of Connectomes)4. Participants were grouped into two distinct subgroups: Cluster 1 (Low-SRS; n = 180 (23.68 ± 16.16) 147 M) and Cluster 2 (High-SRS; n = 126 (99.80 ± 23.87), 112 M)—via k-means clustering (k = 2). K-means clustering algorithmically determined the cut-off between low and high SRS groups (approximately mid-way between clusters at an SRS score around 60–65), rather than relying on a predefined threshold, allowing an empirically-driven distinction between participants with mild versus pronounced social-communication challenges. For the ROI-ROI, functional connectivity (ROIs selected from the Harvard-Oxford Atlas) was computed using the CONN toolbox5 and results were thresholded at a connection-level significance of p < 0.01, with a cluster-level correction of p < 0.05. Results: Group-level comparisons revealed differences in functional connectivity across all three networks. In the RN, the High-SRS group [F(4,696)= 5.95, p <.01] exhibited stronger connectivity between the bilateral precentral gyrus and Supplementary Motor Area (SMA) in addition to bilateral occipitotemporal cortices (iLOC, to ITG). In both the LN [F(4,696) = 8.13, p <.001] and PN [F(4,696)= 8.15, p <.001], the High-SRS group showed stronger connectivity between frontal (IFGoper, OFC, SFG, FP) and temporoparietal (pMTG, aMTG, TP, aSTG) areas. Furthermore, increased connectivity was also found in the LN for the IFGtri, AG, and pSTG regions. By contrast, the Low-SRS group showed no statistically significant connectivity within each network. Conclusion: This study provides novel evidence that differences in functional connectivity across core reading, language, and pragmatic networks may underlie the extent of social-communication difficulties in autism. Participants with higher (worse) SRS scores exhibited increased connectivity across a broader set of brain regions within these networks, potentially suggesting increased effort in processing pragmatic and socially nuanced information6 and compensatory recruitment during reading tasks with social demands7. In contrast, better SRS profiles aligned with more efficient, specialized network organization- possibly reflecting reduced social-cognitive load. These findings reveal how heterogeneity in sociability in ASD may be mirrored in the organization of social-linguistic brain systems. This also underscores the value of stratifying individuals by symptom severity. By linking specific social profiles to distinct neural patterns, the current findings underscore the need for subgroup-level inferences in neuroimaging research in autism.
Topic Areas: Disorders: Developmental,