Latent Profile Analysis Reveals Distinct Executive Function Profiles Across Children with ASD and ADHD

Background:  

Executive functions (EF), the mental control processes necessary to carry out goal-directed behaviors (Denckla, 1994), are impaired in both autism spectrum disorder (ASD) and attention deficit/hyperactivity disorder (ADHD). Although studies have attempted to delineate specific executive dysfunction profiles that discriminate ASD from ADHD (e.g., Sergeant et al., 2002), the high comorbidity of ADHD diagnoses in those with ASD (30%) complicates matters. As an alternative to considering DSM-based diagnoses, the scientific community is moving towards a neurobiological assignment of cognitive dysfunction in a manner consistent with the Research Domain Criteria (RDoc) put forward by NIMH (Insel, 2010). 

Objectives:  

The aim of the current study is to delineate subgroups of children based on patterns of EF strengths and deficits, or “EF profiles”. We investigated a mixed group of typically developing (TD, 45%) children, children with an ASD diagnosis (4%), children with an ADHD diagnosis (29%), and ASD with comorbid ADHD (22%). 

Methods:  

A latent profile analysis was calculated using MPlus, with 10 indicators of executive function (8 subscales from the parent-report of the Behavior Rating Inventory of Executive Function (BRIEF; Gioia, 2002), the statue subscale of the Developmental Neuropsychological Assessment (Korkman & Kemp, 1998), and the backward digitspan subscale of the Wechsler Intelligence Scale for Children- IV (Wechsler, 2004)). Together, these indicators measure inhibition, shifting, working memory, and planning/organizing. Participants included 207 children (Females: N = 45) ages 8-13 years (M = 9.98, SD = 1.22) with average full-scale IQ (M = 109.55, SD = 13.76). 

Results:  

The model that best fit the data contained three classes (entropy: .92; Lo, Mendell, Rubin LRT for 3 v 2 classes: 234.95, p = .02). The first class (N = 88) had overall above average executive functions (“superior”). The second class (N = 57) had slightly below average scores on all of the EF indicators (“middle”). The third class (N = 61) had overall the poorest executive function, which was below average for their age (“low”). Interestingly, these classes did not reproduce the groups based on clinical diagnosis. Most TD children were in the “superior” class (89%), while the majority of ADHD children were split between the “middle” (47%) and “low” (46%) classes. Similarly, children with ASD were primarily in the “low” class (63%), with 35% in the “middle”. EF classes predicted robust phenotypic differences between children. Specifically, the EF classes accounted for unique variance, over and above diagnosis, in anxiety and depression (R² = .56, R² change=.04, p=.001), social problems (R² =.73, R² change=.07, p < .001), attention problems (R²=.79, R² change=.08, p < .001), and aggressive behavior (R² =.58, R² change=.10, p < .001).

Conclusions:  

Using an RDoc framework, the present study examined EF in a mixed group of TD children and children with neurodevelopmental disorders. Importantly, the EF classes that emerged from latent profile analysis predicted variance in behavioral problems unique from diagnosis. Future studies should validate these EF classes biologically by investigating whether unique brain-based markers for EF dysfunction can be determined.