Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Network Analysis of Hemispheric Language Dominance Reveals Enhanced Fronto-Temporal Connectivity in Atypically Lateralized Individuals
Poster D1 in Poster Session D, Saturday, October 26, 10:30 am - 12:00 pm, Great Hall 4
Ieva Andrulyte1, Laure Zago2, Gael Jobard2, Herve Lemaitre2, Francois M Rheault3, Simon S Keller1, Laurent Petit2,3; 1University of Liverpool, 2University of Bordeaux, 3Université de Sherbrooke
Introduction: The lateralisation degree of language functioning has been suggested to depend on hemispheric asymmetry in white matter (Verhelst et al, 2021; Westerhausen et al, 2006); however, the directionality of these findings remains unclear. This suggests that simple tractography techniques might not be sensitive enough to identify language dominance. Significant insights into the functional organization of the human brain may be achieved by considering networks and brain connectivity, providing more information about discrepancies in people with different hemispheric language dominance. In this study, we examined 283 healthy participants (aged 18-58 years) from the BIL&GIN database (Mazoyer et al., 2016) to compare their structural connectomes at the whole-brain level and determine the networks responsible for different laterality groups. Methods: All participants underwent functional and diffusion-weighted 3T MRI. Using a language production task, individuals were classified into three groups: 239 typical, 35 atypical, and 9 strongly atypical for left-asymmetrical language function (Mellet et al., 2014). Probabilistic tractography was employed to generate whole-brain tractograms (Theaud et al., 2020), and the Extractor_flow post-processing technique was used to extract white matter fibres according to anatomical Boolean guidelines (Petit et al., 2023). Connectivity matrices were generated using the Scilpy tool (https://github.com/scilus/scilpy), with nodes representing the homotopic regions of interest from the AICHA atlas corresponding to the 32 supramodal sentence areas from the SENSAAS language atlas (Labache et al., 2019), and edges weighted by fractional anisotropy (FA). Network-based statistics (NBS) tool was used to identify subnetworks that significantly differed between groups (Zelesky et al., 2010). We used an F-test to compare edge weights across the typical, atypical, and strongly atypical groups. This nonparametric approach is analogous to cluster-based correction and addresses the statistical challenge of massive multiple comparisons in whole-brain connectivity analysis by identifying the largest statistically significant fully connected network of suprathreshold edges. The p-value for a subnetwork was calculated as the proportion of permutations where the largest connected subnetwork was the same size or larger than the observed subnetwork, normalized by the number of permutations. Results: NBS analysis revealed one significant network comprising 29 nodes and 29 edges, showing group differences in edge weight associations among the three laterality groups. This network spanned bilateral temporal and frontal lobes, with particularly high nodal degree in the anterior insula and middle temporal gyrus bilaterally. The strongest connections were observed between bilateral temporal regions; from the left temporal lobe to the right frontal lobe; and between the right supplementary motor area and the precuneus. Post-hoc analyses indicated that atypical individuals exhibited stronger connectivity between the right anterior insula and right superior frontal gyrus, as well as bilateral temporal connections compared to left-lateralized individuals. Similarly, strongly atypical individuals had higher FA connectivity in bilateral fronto-temporal lobes; and between the right supplementary motor area and right precuneus compared to typical individuals. There was no significant difference in network connectivity between strongly atypical and atypical individuals. Conclusions: Both atypical and strongly atypical individuals showed stronger bilateral temporal connectivity compared to typically lateralized people, suggesting a potential compensatory mechanism.
Topic Areas: Language Production, Computational Approaches