Distinct Functional Network Connectivity in Children with ASD, ADHD, and OCD: Data from the Pond Network

Background: Children with autism spectrum disorder (ASD) share genetic vulnerability and symptom domains with other neurodevelopmental disorders (NDs), such as attention-deficit/hyperactivity disorder (ADHD) and obsessive-compulsive disorder (OCD) despite distinct diagnostic criteria. A body of research of resting-state functional connectivity for each of the disorders shows that NDs exhibit some notable overlapping patterns of atypical functional connectivity in large-scale networks. However, the findings have been inconsistent across studies and there is no study to consider three different NDs within a single cohort. We compared large-scale whole brain fMRI networks directly across ASD, ADHD, OCD, and typically developing (TD) participants collected in a single cohort using graph theory to examine what the shared and disorder-specific functional organization abnormalities are in the three different NDs.

Objectives: We investigated a) if resting-state functional connectivity differs between NDs (including children with ASD, ADHD, and OCD) and TDs, and b) if the children with ASD show different functional connectivity profiles compared to children with ADHD or OCD.

Methods: The current study included 311 high-functioning children and adolescents (out of 354 subjects, excluding those with excessive motion), 156 with ASD, 68 with ASD, 43 with OCD, and 44 TDs (4-21 years of age, mean 12 years). T1-weighted and resting-state fMRI (5 minute duration) images were obtained on a Siemens 3T MRI using a 12-channel head coil. Subjects’ T1-weighted images were parcellated into 82 cortical and subcortical ROIs, and subject-specific networks were constructed by correlating the mean resting-sate time-series for each ROI pair. Local (degree, strength, local efficiency, and betweenness centrality) and global (small-worldness and global efficiency) network measures were obtained for each network over a range of sparsity thresholds, and the area under the threshold curve (AUC) was calculated. An analysis of covariance was used to compare the network measure AUCs of NDs with TDs, as well as ASD with OCD and ADHD, with age and sex as covariates. Non-parametric permutation testing was used to determine significance (p < 0.05).

Results: Children with NDs showed an altered connectivity profile compared to the TDs, with increased degree and strength in ROIs including subcortical structures and the insula. These ROIs, as well as ROIs such as the pars opercularis and caudal anterior cingulate gyrus, also showed significantly increased local efficiency in NDs compared to TDs. Significant age and sex interactions were found between the two groups. Within the ND group, ASD exhibited significantly decreased degree and strength in ROIs including the paracentral gyrus compared to ADHD, while this effect was found in OCD in ROIs such as the fusiform gyrus. ASD also showed increased degree and strength in subcortical structures compared to OCD. Betweenness centrality also differed amongst the ND groups. Age and sex interactions differed significantly between the NDs; details will be presented.

Conclusions: The NDs show increased graph theoretic measures in subcortical structures compared to TDs. Within the NDs, we found evidence for distinct and shared areas functional connectivity in ASD, ADHD and OCD. These findings will help us understand the heterogeneity across NDs.