Associations Between Autism and Self-Reported Dimensions of Interoception.

Everything was measured at a single time point, so the study can show that these things go together but not which one comes first. Interoception was captured only by self-report questionnaires, and the developers' reliability data for the accuracy, attention and negative-evaluation scales come from general-population samples, not autistic ones. Participants reported their own diagnoses. Verification meant naming a clinic and clinician, and answers could in principle have been fabricated, though the authors note there was no payment difference to create an incentive. Mental health and neurodevelopmental conditions other than autism and anxiety were not verified at all, and every one of them was lumped into a single yes/no variable, so the study cannot say which specific conditions matter. The sample had high rates of depression, ADHD and PTSD. The autistic group (232 people) was larger than the non-autistic group (143), and comparing group averages hides the wide variation that exists within both groups. The anxiety measure includes questions about noticing anxiety-related bodily signals, and the attention and negative-evaluation scales may overlap, so adjusting for anxiety and treating the two as separate may not fully work. The split of attention questions into anxiety-related and non-anxiety-related signals has not been confirmed by factor analysis, so that categorisation may not be valid.

The lead author's post is funded by a University of Oxford Medical Sciences Graduate School studentship. The authors declared no competing interests. No company that sells a treatment or product was involved.

Do not conclude that autism is linked to poorer or heightened awareness of bodily signals. Higher autistic traits went with lower self-rated accuracy and higher attention in the general-population sample, but when participants were grouped by a verified autism diagnosis, neither accuracy nor attention differed from the non-autistic group. Only the negative interpretation of bodily signals was higher in the diagnosed group, and that difference was small.