Anterior Cingulate Connectivity in Children with Autism Spectrum Disorders

Background: Converging lines of evidence support models of autism spectrum disorders (ASD) as developmental dysconnection syndromes, characterized by increased short-range connectivity and reduced long-range connectivity in the brain. Initially based on findings of white matter abnormalities during development, such models have gained support from recent evidence of autism-related decreases in cortico-cortical functional connectivity (FC) in task-based studies. Two initial studies, using recently emerging resting state fMRI approaches to map FC, have shown FC abnormalities between the pregenual anterior cingulate cortex (pgACC) and medial prefrontal cortex in adults with ASD. However, to date no resting state fMRI research has been conducted in children with ASD.

Objectives: As a first step to characterize the pattern of FC in youth with ASD, we examined FC of the pgACC network in school-age children with ASD. The pgACC network was chosen based on both its role in social cognition and its consistent pattern of ASD-related hypoactivation in social processes.

Methods: Seven children with ASD (10.8 y ± 1.7; 3 girls; 6 with Autism and 1 with PDD-NOS) and seven age- and sex-matched NC completed a 6.5 min resting state fMRI scan (field strength= 3 T; TR = 2000 ms; voxel 3x3x3mm). All children and their parents signed an NYU IRB approved assent and consent form, respectively. Preprocessing included slice-time correction, motion correction, bandpass temporal filtering, spatial filtering, and spatial normalization. A spherical region of interest centered in the pgACC was selected as the seed region for FC analyses. For each participant, FC analyses were carried out using multiple regression (implemented in FSL’s FEAT) including the timeseries of the pgACC seed, and nine nuisance covariates as predictors (i.e., movement, white matter, global signal, and CSF). Group analyses using random effect models implemented in FLAME were carried out. Corrections for multiple comparisons were performed at the cluster level using Gaussian random field theory (min Z > 2.3; cluster significance: p ≤ 0.05, corrected). Further, we computed the number of short- and long range connections by measuring the Euclidean distance (cutoff=45 mm) between the center of the seed ROI and every other voxel that reached significance in the group-level thresholded Z-score maps.

Results: Children with ASD demonstrated higher degrees of local connectivity in the pgACC network. This pattern was quantitatively evident as an increased number of significant voxels connected within 45 mm Euclidean distance from the pgACC. Consistent with their young age, neither group showed a high degree of significant long-range FC between pgACC and the posterior component of the network (posterior cingulate cortex).

Conclusions: Our preliminary findings suggest that ASD are characterized by greater short-range connectivity in the pgACC network. Data collection is still ongoing to reach a sample size of 20 children per group. Attainment of our aims would further implicate the social-processing pgACC network in autism, as well as motivate translational studies on the neuronal developmental processes that drive increases in short-range connectivity and decreases in long-range connectivity.