Abnormal wiring of the structural connectome in adults with ADHD

Abstract Current knowledge of white matter changes in large-scale brain networks in adult attention-deficit/hyperactivity disorder (ADHD) is scarce. We collected diffusion-weighted magnetic resonance imaging data in 40 adults with ADHD and 36 neurotypical controls and used constrained spherical deconvolution–based tractography to reconstruct whole-brain structural connectivity networks. We used network-based statistic (NBS) and graph theoretical analysis to investigate differences in these networks between the ADHD and control groups, as well as associations between structural connectivity and ADHD symptoms assessed with the Adult ADHD Self-Report Scale or performance in the Conners Continuous Performance Test 2 (CPT-2). NBS revealed decreased connectivity in the ADHD group compared to the neurotypical controls in widespread unilateral networks, which included subcortical and corticocortical structures and encompassed dorsal and ventral attention networks and visual and somatomotor systems. Furthermore, hypoconnectivity in a predominantly left-frontal network was associated with higher amount of commission errors in CPT-2. Graph theoretical analysis did not reveal topological differences between the groups or associations between topological properties and ADHD symptoms or task performance. Our results suggest that abnormal structural wiring of the brain in adult ADHD is manifested as widespread intrahemispheric hypoconnectivity in networks previously associated with ADHD in functional neuroimaging studies.


Supplementary figure
. Results of the NBS group analysis after removing six outlier participants (edge weight > 2 SD in two or more connections in the NBS networks) from the NT group.Compared with the main results (see the main text), the subcallosal gyrus was no longer part of the left-sided network.In the right-sided network, two nodes (the intraparietal sulcus and middle occipital sulcus/lunate sulcus) were not present after outlier exclusion, and the inferior occipital sulcus appeared as a new node.The left-sided network was identified only with intensity as a measure of network size.Results of the right-sided network are presented by using extent as a measure of network size.4. NBS results of the control analyses using the Schaefer parcellation.NBS identified decreased connectivity in the ADHD group as compared with the NT adults in a right-sided network largely overlapping with the network obtained with the Destrieux atlas.All p-values are FWE-corrected.With a threshold of 4.0, the network was identified only with intensity as a measure of network size.The other results are from using extend as a measure of network size.No statistically significant networks with ADHD > NT were identified (smallest p-value: t = 4.0, p = .11,FWEcorrected).No networks associated with the behavioural measures were found (smallest p-values: inattention: t = 3.0, p = .09;hyperactivity-impulsivity: t = 3.5, p = .36;commission errors: t = 4.0, p = .09;omission errors: t = 3.0, p = .33,FWE-corrected).

Supplementary figure 7.
Results of the NBS analyses after removing one outlier participant (total count of streamlines > 3 SD) from the ADHD group.Only networks with differences compared with the main results are shown.(A) In the group analysis after the outlier exclusion, the anterior lateral sulcus and thalamus were no longer present in the left-sided network.In turn, in the right-sided network, the middle occipital sulcus/lunate sulcus node was not present after the outlier exclusion, and the calcarine sulcus appeared as a new node.The left-sided network was identified only with intensity as a measure of network size.Results of the right-sided network are presented by using extent as a measure of network size.(B) In the network representing hypoconnectivity associated with higher amount of commission errors in CPT-2, the left superior frontal gyrus and right putamen were not present after the outlier exclusion, and the right inferior frontal sulcus appeared as a new node.

Supplementary table 1. Nodes and their center of mass MNI coordinates in the networks
identified with NBS using the Destrieux atlas.Note.M = mean, SD = standard deviation, df = degrees of freedom.

table 2 .
Nodes of the NBS networks distinguishing adults with ADHD from NT controls identified with t-statistic thresholds 3.5 and 4.0 using the Destrieux atlas.
Note.L = left, R = right, aMCC = middle-anterior part of the cingulate gyrus and sulcusSupplementary table 3. Group differences of global graph theory metrics obtained with the Destrieux atlas.Note.M = mean, SD = standard deviation, df = degrees of freedom.Supplementary

table 4 .
Correlations between ADHD symptoms, CPT performance and global graph theory metrics obtained with the Destrieux atlas.

table 5 .
Nodes and their center of mass MNI coordinates in the networks identified with NBS using the Schaefer parcellation.

table 6 .
Group differences of global graph theory metrics obtained with the Schaefer parcellation.