Subtly altered topological asymmetry of brain structural covariance networks in autism spectrum disorder across 43 datasets from the ENIGMA consortium

Abstract

Small average differences in the left-right asymmetry of cerebral cortical thickness have been reported in individuals with autism spectrum disorder (ASD) compared to typically developing controls, affecting widespread cortical regions. The possible impacts of these regional alterations in terms of structural network effects have not previously been characterized. Inter-regional morphological covariance analysis can capture network connectivity between different cortical areas at the macroscale level. Here, we used cortical thickness data from 1455 individuals with ASD and 1560 controls, across 43 independent datasets of the ENIGMA consortium's ASD Working Group, to assess hemispheric asymmetries of intra-individual structural covariance networks, using graph theory-based topological metrics. Compared with typical features of small-world architecture in controls, the ASD sample showed significantly altered average asymmetry of networks involving the fusiform, rostral middle frontal, and medial orbitofrontal cortex, involving higher randomization of the corresponding right-hemispheric networks in ASD. A network involving the superior frontal cortex showed decreased right-hemisphere randomization. Based on comparisons with meta-analyzed functional neuroimaging data, the altered connectivity asymmetry particularly affected networks that subserve executive functions, language-related and sensorimotor processes. These findings provide a network-level characterization of altered left-right brain asymmetry in ASD, based on a large combined sample. Altered asymmetrical brain development in ASD may be partly propagated among spatially distant regions through structural connectivity.

Fig. 1. Schematic workflow of this study.

A Flowchart of the procedure used in the current study. We first constructed intra-individual, intra-hemispheric structural covariance networks in each dataset using regional cortical thickness data. Then, for each individual, we computed graph theory metrics at the global and nodal levels using the intra-hemispheric networks. Finally, we calculated individual-level hemispheric differences for each metric, to examine case-control differences of topological network asymmetry. B Small-world network model. At the whole-hemisphere level, we estimated network integration and segregation using small-world parameters. A regular network is characterized by a high clustering coefficient and long shortest path length, corresponding to high local specialization and low global integration. In contrast, a random network has a low clustering coefficient and short shortest path length, corresponding to low local specialization and greater global integration. A small-world model reflects a balance between the extremes of local specialization versus global integration. C At the nodal level, we examined four graph theory measures: degree centrality and nodal global efficiency both measure global connectivity from/to a given node, whereas the cluster coefficient and nodal local efficiency reflect local connectivity from/to that node. Abbreviations: ASD autism spectrum disorder; HC healthy control; SD standard deviation.

Fig. 2. Cohen’s d effect sizes of ASD case-control associations for node-level topological asymmetries.

a Effect sizes from ASD case-control analysis of node-level topological metric asymmetries that reflect global connectivity of each node, i.e., degree centrality and nodal global efficiency. b Effect sizes from ASD case-control analysis of nodal-level topological metric asymmetries that reflect local connectivity of each node, i.e., the clustering coefficient and nodal local efficiency. Positive effect sizes (pink-red) indicate shifts towards greater leftward or reduced rightward asymmetry in ASD compared to controls, and negative effect sizes (blue) represent shifts towards greater rightward asymmetry or reduced leftward asymmetry in ASD compared to controls.

Fig. 3. Regions with altered average network-level asymmetries in ASD compared to separate region-by-region testing.

The color key is indicated in the figure. See the main text for the citation of the study that performed separate region-by-region testing.

Fig. 4. Altered asymmetry of connectivity linking to the nodes with significant alterations of degree centrality asymmetry in ASD.

a Altered asymmetry of connectivity linked to the fusiform in ASD. b Altered asymmetry of connectivity linked to the rostral middle frontal cortex in ASD. c Altered asymmetry of connectivity linked to the superior frontal cortex in ASD. The yellow nodes indicate the brain regions. Red indicates a significant edge-level, reduced rightward asymmetry of connectivity in ASD compared to controls, and blue indicates an edge-level, reduced leftward asymmetry of connectivity in ASD compared to controls.

Fig. 5. Cognitive functions associated with cortical regions showing altered connectivity asymmetry.

Meta-analyzed fMRI data were used to functionally annotate cortical regions showing altered connectivity asymmetry with the fusiform (a), rostral middle frontal (b) or superior frontal (c) cortex. Left panels indicate the regions showing alterations of lateralized connectivity, which were used as input masks to the decoder function of Neurosynth (see Methods). Middle panels show the brain co-activation maps corresponding to the input masks. Right panels show the cognitive terms corresponding to the co-activation maps, in word-cloud plots. The font sizes of the cognitive terms indicate their map-wide correlations with the co-activation maps (correlation coefficients are in Supplementary Table 17).