Atypical early neural responses to native and non-native language in infants at high likelihood for developing autism.

The final imaging sample was small, especially the typical-likelihood group (19 infants), and the original 63-infant sample was reduced by exclusions for scanner artifacts, motion artifacts, or no auditory response. The groups differed on sex, maternal education and birth order, which can affect language development, although post-hoc tests controlled for them. High-likelihood infants were analysed as one group, but they are developmentally heterogeneous: some later received an autism diagnosis, some had other concerns, and some did not. Sleep stage was not recorded during the scan, so it cannot be fully ruled out as a confound. The paper does not present a clinical screening tool or intervention effect.

Funding was from the Eunice Kennedy Shriver National Institute of Child Health and Human Development and the National Institute of Mental Health. The paper does not report author conflicts of interest.

Do not read these brain responses as a clinical test for autism. The high-likelihood group included infants who later had autism, language or other concerns, and not all developed autism; the imaging was done during sleep in a research setting and is not a bedside assessment.