Dissociations in the neural substrates of language and social functioning in autism spectrum disorder.

The ASD sample had average to above average IQ scores and excluded people with full-scale or verbal IQ below 80, so the findings may not apply to autistic people with intellectual disability. The study used self-ratings only, without proxy ratings or performance-based measures, so the results reflect participants' own perceptions rather than tested language or social communication ability. The ASD group did not self-report structural or pragmatic language scores worse than the standardization sample, even though pragmatic language impairment is a prominent characteristic of ASD, which may limit sensitivity to language difficulties. There was a group difference in time between the questionnaire and the brain scan, driven by 8 ASD participants; analyses were repeated without them, but this timing mismatch remains a limitation. The brain findings are associations between self-ratings and cortical thickness, not evidence that brain structure causes language or social difficulties, and the typically developing group showed inverse relationships. The social engagement cortical association did not remain significant after controlling for pragmatic language, so the separation between social and language correlates is not fully independent.

The disclosure sentence reports no financial disclosures or conflicts of interest for Jason Crutcher, Alex Martin, and Gregory Wallace. Publication types list NIH intramural research support.

Do not conclude that thinner cortex causes poorer language or social ratings, or that brain scans can diagnose autism. The study compared self-ratings and brain structure across groups, and the ASD group's language scores were within the normal range on average.