Slide Slam N4
The topological properties of the left thalamus in the white matter network in children with reading disabilities
Chenglin Lou1,2, Alex Cross2,3, Lien Peters1,2, Daniel Ansari1,2,4, Marc Joanisse1,2,5; 1Department of Psychology, The University of Western Ontario, London, Canada, 2Brain and Mind Institute, The University of Western Ontario, London, Canada, 3Health and Rehabilitation Sciences, The University of Western Ontario, London, Canada, 4Faculty of Education, The University of Western Ontario, London, Canada, 5Haskins Laboratories, New Haven CT, USA
MRI studies of reading disabilities (RD) report anomalies in white matter tracts between the left thalamus and primary sensory cortex (Müller-Axt et al., 2017; Tschentscher et al., 2019), suggesting a potential role of thalamo-cortical connections in RD. In addition, recent studies report deficits of white matter tracts at the connectome level in RD (Bathelt et al., 2018; Lou et al., 2019, 2021; Lee et al., 2020). However, little is known about how the left thalamus and thalamo-cortical connections in the connectome affect reading performance. Here we investigated whether thalamo-cortical networks are related to reading performance in children with RD. 64 children (ages 8 -14 years, 31 boys) underwent diffusion-weighted imaging at 3T and were tested on a series of reading tasks including sight word reading, phonemic decoding, reading comprehension, and rapid automatized naming. The sample ranged in reading ability and included a subset of children with identified RD. Each participant’s whole-brain network was constructed, with number of streamlines, determined via whole-brain tractography, used as edge weights connecting 90 parcels. The topological properties of the left thalamus, including efficiency and routing cost (Avena-Koenigsberger et al., 2019), were computed based on both the whole-brain network and a subnetwork which was composed of connectome hubs and reading network nodes. Efficiency was evaluated by local efficiency (LE) and clustering coefficient (CC), and routing cost was evaluated by informational cost (IC) and transmission cost (TC). In addition, the number of streamlines between the left thalamus and all reading network regions per edge, as well as two reading-related regions (superior temporal gyrus, STG, and middle temporal gyrus, MTG) were extracted. Partial correlations were correlated between reading scores and topological metrics of the whole-brain network and the subnetwork, with gender and handedness as covariates. Another analysis examined whether the number of streamlines between the left thalamus and reading-related regions (STG, MTG, and all reading network regions) were associated with reading scores. Family-wise error was corrected with a 10,000-permutation Monte-Carlo simulation. Significant correlations between topological metrics of the left thalamus and reading scores were observed only in the subnetwork. CC (r = -.35, p(corr) = .019) and LE (r = -.40, p(corr) = .006) were correlated with reading comprehension. Similarly, CC (r = -.34, p(corr) = .025) and LE (r = -.32, p(corr) = .043) were correlated with rapid automatized naming. TC was positively correlated with phonemic decoding (r = .36, p(corr) = .021). In addition, the average number of streamlines between the left thalamus and reading network was correlated with reading comprehension (r = -.31, p(corr) = .045), and the number of streamlines between the left thalamus and left STG was correlated with phonemic decoding (r = -.35, p(corr) = .020). A control analysis of the right thalamus yielded no significant correlations. The present study demonstrated that reading subskills were associated with thalamo-cortical connections and topological properties of the left thalamus in the white matter network. This work highlights the role of the left thalamus and thalamo-cortical network in understanding skilled and impaired reading in children.