Executive Function Does Not Separate Autism From ADHD

What the research found

For most of two decades the field has assigned each condition its own cognitive signature. Trouble stopping an action already underway belonged to ADHD. Trouble shifting from one rule to a new one belonged to autism. The division was assembled indirectly, from studies comparing autistic children with typically developing children, set beside separate studies comparing children with ADHD with typically developing children.

Parker Townes and colleagues at the University of Calgary and the Hospital for Sick Children in Toronto asked what remains of that division when the comparison is made directly. Their systematic review and meta-analysis, published in the Journal of Attention Disorders in 2023, admitted only studies that tested all three groups at once, matched for age and sex, on the same tasks and by the same testers. Fifty-eight articles qualified, covering 8,259 participants under the age of 19, with a mean age of 9.5 years.

The analysis worked at the level of the individual measure. Forty-five performance metrics drawn from 24 tasks were pooled one at a time, wherever two or more studies had reported the same metric. The team refused to combine metrics into a single figure for each domain, on the grounds that two tasks both labelled working memory frequently measure different operations, so each domain amounts to a set of small separate results read together.

No domain distinguished autistic children from children with ADHD. Cognitive flexibility, the shifting that had been assigned to autism, was assessed through seven metrics across four tasks, and none of them separated the groups. Response inhibition, the stopping that had been assigned to ADHD, was assessed through fifteen metrics, of which thirteen showed nothing and the two that did point in opposite directions.

Where individual measures did diverge, the two steadiest were the Wechsler processing speed index, on which children with ADHD scored higher than autistic children by a standardized difference of 0.43 across three studies, and the Wechsler block design task, on which autistic children scored higher by 0.34 across four studies. Both of those belong to categories the authors themselves flag as not always counted as executive functions.

What the tasks registered instead was the distance between the diagnosed groups and everyone else. Both performed worse than their typically developing peers on attention, response inhibition, cognitive flexibility and visuospatial measures, with less consistent deficits in working memory and processing speed. Planning was the single domain in which neither group fell behind at all.

What this means for you and your child

Executive-function testing in a child of primary or secondary school age can, on this evidence, establish that executive difficulty is present and cannot arbitrate between the two diagnoses. The distinction rests on other ground, chiefly social communication and repetitive behaviour.

For a child who carries both diagnoses, the review offers less than a parent would want. Eleven of the 58 articles included a group with both conditions, and the review’s three-group design left no place to put them, so that group appears in none of the analyses. The authors are explicit that they cannot determine whether the shared executive profile exists because the two conditions draw on a common mechanism or because so many children in the autism samples also met criteria for ADHD, an overlap other work puts between 40 and 70 percent. For that child, what the review can offer is the absence of a distinct executive profile attached to either label, and therefore no separate set of strengths and weaknesses to plan around.

Several features of the evidence argue for caution about a shared profile. Many of the pooled comparisons rested on two or three studies, and the between-study variance estimate reached 1.73 on one inhibition measure, a value indicating that those studies were not measuring under comparable conditions. Girls made up 11 percent of the autism groups and 28 percent of the ADHD groups. Across pooled samples this small, a real difference of modest size would have gone undetected. What the review establishes is that the difference the field expected did not appear where it was looked for.

A child who cannot hold a three-step instruction, or who cannot let go of a rule that has stopped working, has a difficulty that appears on these tasks under both diagnoses, and it can be taught to directly, whichever label stands at the top of the report.

Whether the two conditions share an underlying mechanism, or whether the tasks currently in use are simply too coarse to detect the difference between them, is a question these fifty-eight studies could not decide. A child’s executive difficulties are real under either answer, and they are what these tests actually record.