Back

Incorporating special interests to investigate the language system in autism: A feasibility pilot fMRI study

Olson, H. A.; Johnson, K. T.; Nishith, S.; Gabrieli, J. D. E.; D'Mello, A. M.

2025-04-05 neuroscience
10.1101/2025.04.04.647117 bioRxiv
Show abstract

Most autistic individuals have sustained, focused interests in particular topics or activities. In some cases, these special interests have been shown to motivate communicative behaviors, a domain in which many autistic individuals experience challenges. We conducted a pilot study with 15 autistic children (ages 8.18 - 13.27 years, mean(SD)= 11.17(1.62), 3 female/11 male/1 nonbinary), comparing brain responses elicited by short narratives tailored to individuals special interests to responses elicited by generic, non-tailored narratives. Using functional magnetic resonance imaging (fMRI), we found that autistic children did not show typical language responses to generic narratives. However, they did show heightened responses to the narratives that incorporated their special interests relative to the generic narratives in language regions and in regions associated with reward and self-reference. Brain responses for personalized narratives were also more consistent across children than responses for the generic narratives. These results suggest that personalizing stimuli by incorporating special interests might be a promising approach for neuroimaging in autistic participants. LAY SUMMARYIn a pilot sample of autistic children, we found that listening to short narratives tailored to each childs special interest elicited higher responses in the brain than listening to non-personalized narratives. Brain responses across children were also more similar for the special interest narratives. Thus, personalizing stimuli to special interests may be a promising approach for neuroimaging studies of autism.

Matching journals

The top 9 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.