Dimensional Analysis of Executive Dysfunction Comorbidity in ASD and ADHD

Background:   One challenge to understanding the neural basis of executive dysfunction comorbidity between Attention Deficit Hyperactivity Disorder (ADHD) and Autism Spectrum Disorder (ASD) is that executive function (EF) is multidimensional and extent of impairment highly heterogeneous within each disorder.  Further, EF abilities vary greatly among typically developing children as well.  Thus, which EF dimensions are most relevant to comorbidity is elusive. 

Objectives: We first applied a novel data-driven graph theory method, community detection, to identify relevant EF dimensions across ADHD, ASD, and control children, and then examined their neural basis using task-based functional magnetic resonance imaging (fMRI).  Our fMRI tasks were designed with the premise that controlling attention in an adaptive and flexible manner subserves EF.  

Methods:   We included 114 7-14 year-old children matched for age and IQ (35 ASD, 20 ADHD, 59 controls) and collected parent-report of 10 measures, Inattention and Hyperactivity/Impulsivity from ADHD Rating scale and 8 subscales selected from the BRIEF (Inhibit, Shift, Emotion Control, Initiate, Working Memory, Plan/Organize, Organization of materials, Monitor). Community detection was applied using a weight-conserving modularity algorithm not dependent on threshold, yielding 3 communities: #1 defined by worse planning and organizing but relatively better cognitive flexibility and hyperactivity/impulsivity (21 ASD, 6 ADHD, 29 Controls); #2 defined by worse cognitive flexibility but relatively better planning/organizing and hyperactivity/impulsivity (12 ASD, 4 ADHD, 16 controls); and # 3 defined by worse hyperactivity/impulsivity but relatively better planning organizing and cognitive flexibility (2 ASD, 10 ADHD, 14 Controls). Imaging was performed at 3T during 2 runs of a shape-classification task with right-hand response, in which distracters were flashed in the periphery. Runs differed in task-context, the first being stimulus-driven as distractors were irrelevant to the task, and the second requiring top-down control as a distractor was designated as target requiring left-hand response. Activation difference between the first to second run reflected attentional modulation by task-context, termed adaptive flexibility. Images were slice-time and motion-corrected, normalized to EPI template and resliced to 3mm, smoothed with 8mm FWHM and using stringent motion-criteria retained 88 subjects (51 controls, 26 ASD and 16 ADHD). We identified regions reflecting adaptive flexibility by Run x Distractor interaction in whole-brain ANOVA in SPM 8 (p < .05 Monte Carlo corrected; p<.001, k=28 voxels). For each cluster ROI, hierarchical linear regression tested whether clusters accounted for more variance than DSM-based groups.

Results:  

Significant Run x Distractor interaction was observed in visual attention regions, bilateral middle frontal gyrus, bilateral insula and putamen, bilateral SPL-TPJ and postcentral gyrus, dorsal ACC, SMA, precuneus, right MTG and left visual cortex and community-based grouping accounted for more variance in right TPJ (p=.04) and right MFG (p=.029) than DSM groups, with greatest flexibility-related activation for community #1, scoring best on measures of attention and flexibility and worse on planning/organizing. 

Conclusions:  

Our results suggest that executive dysfunction variability across ASD, ADHD and control children is better conceptualized by three distinct functional profiles associated with variability in frontal-parietal cortical engagement during flexible attentional control