Table 1.
Demographic and clinical details of participants.
Figure 1.
Flowchart for the construction of structural cortical network.
For each subject, its original intensity image is first tissue-segmented, cortical-surface-constructed, thickness-measured, and surface parcellated into 68 ROIs. Then, it will serve as one entry of thickness matrix. A correlation matrix is finally obtained for each population group. The brain network is further built using ROIs as nodes and the correlations as edges, and modular organization is also computed.
Table 2.
List of 68 anatomical regions comprising the cortical networks.
Figure 2.
Modular organization of the cortical structural network of autistic children and typically developing children.
(A) The brain modular organization of autistic children. Module I: executive strategic, sensorimotor, and visual; Module II: spatial and auditory; Module III: recognition and self-awareness. (B) The brain modular organization of typically developing children. Module I: executive strategic; Module II: spatial, auditory, and visual; Module III: self-reference and episodic memory. Colors in each module represent different cortical regions.
Figure 3.
(A) Between-group difference in the network modularity. The network modularity of the autism (left) is significantly reduced when compared with the control (right). (B) Between-group difference in the intra-module connectivity (MC). Autism showed a significantly reduced connectivity in the modules I and II, while demonstrated a significantly increased connectivity in the module III. (C) Between-group differences in the inter-module connectivity (IMC). Autism showed a significantly increased connectivity between modules II and III, and between modules I and III. No significant difference was found for the connectivity between modules I and II.
Figure 4.
Regional roles of network modularity.
The between-group comparisons identified 16 and 9 regions with significant altered Participant Coefficient (PC) and intra-module degree (MD) values in the autistic brain, respectively. Positive values mean that autism has larger connectivity than healthy children. Stars indicate the regions with significant difference in the number of connections between groups in 10,000 permutation tests (p<0.01, FDR corrected).
Figure 5.
Illustration of connections with significantly altered regional correlations in the autism network.
Left shows the top view, and right shows the lateral view. Decreased correlations in autism were marked in red, and increased correlations were marked in blue. Visualization was implemented using BrainNet Viewer (http://www.nitrc.org/projects/bnv/).
Table 3.
Altered regional correlations in autistic and control networks (p<0.01 FDR corrected).