Send to

Choose Destination
Mol Psychiatry. 2018 May;23(5):1198-1204. doi: 10.1038/mp.2017.75. Epub 2017 Apr 25.

Autism spectrum disorders: a meta-analysis of executive function.

Author information

Autism Clinic for Translational Research, Brain and Mind Centre, Central Clinical School, Faculty of Medicine, University of Sydney, Camperdown, Sydney, NSW, Australia.
School of Psychology, Faculty of Science, University of Sydney, Sydney, NSW, Australia.
Norment, KG Jebsen Centre for Psychosis Research, Oslo University Hospital and Institute of Clinical Medicine, University of Oslo, Oslo, Norway.


Evidence of executive dysfunction in autism spectrum disorders (ASD) across development remains mixed and establishing its role is critical for guiding diagnosis and intervention. The primary objectives of this meta-analysis is to analyse executive function (EF) performance in ASD, the fractionation across EF subdomains, the clinical utility of EF measures and the influence of multiple moderators (for example, age, gender, diagnosis, measure characteristics). The Embase, Medline and PsychINFO databases were searched to identify peer-reviewed studies published since the inclusion of Autism in DSM-III (1980) up to end of June 2016 that compared EF in ASD with neurotypical controls. A random-effects model was used and moderators were tested using subgroup analysis. The primary outcome measure was Hedges' g effect size for EF and moderator factors. Clinical sensitivity was determined by the overlap percentage statistic (OL%). Results were reported according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A total of 235 studies comprising 14‚ÄČ081 participants were included (N, ASD=6816, Control=7265). A moderate overall effect size for reduced EF (Hedges' g=0.48, 95% confidence interval (CI) 0.43-0.53) was found with similar effect sizes across each domain. The majority of moderator comparisons were not significant although the overall effect of executive dysfunction has gradually reduced since the introduction of ASD. Only a small number of EF measures achieved clinical sensitivity. This study confirms a broad executive dysfunction in ASD that is relatively stable across development. The fractionation of executive dysfunction into individual subdomains was not supported, nor was diagnostic sensitivity. Development of feasible EF measures focussing on clinical sensitivity for diagnosis and treatment studies should be a priority.

[Indexed for MEDLINE]
Free PMC Article

Supplemental Content

Full text links

Icon for Nature Publishing Group Icon for PubMed Central
Loading ...
Support Center