Interpersonal Coordination Deficits in Joint Action in Pre-school Children With Autism Spectrum Disorder: Evidence From fNIRS Hyper-Scanning and Machine Learning.
fNIRS hyperscanning shows ASD preschoolers have real, measurable brain-to-brain desync during joint play.
01Research in Context
What this study did
Sirao et al. (2026) watched the preschoolers play together. Half had autism, half were neurotypical.
Each pair wore soft fNIRS caps that tracked their brain oxygen at the same time. Kids did a tower-building game and a balloon-popping race.
What they found
Autistic pairs moved less in sync and their brain signals matched up 30 % less than neurotypical pairs.
The gap showed up in both games, so the trouble is not tied to one kind of play.
How this fits with other research
Simões et al. (2020) saw the same social gap in adults: autistic people stood farther away from VR avatars. The age and task differ, but the social-distance theme repeats.
Cheng et al. (2012) gave kids VR joint-attention lessons and saw gains. That paper proved training can help; Kaiyun now shows the baseline deficit is real and neural.
Garon et al. (2018) also found preschoolers with ASD scoring lower on simple EF tasks. Both studies use the same quasi design and age group, strengthening the idea that early social-cognitive gaps are measurable before school.
Why it matters
You now have a neural yardstick for joint play. If two ASD kids build a tower and their fNIRS signals stay out of step, you know the dyad needs extra scaffolding. Use turn-taking scripts or mediator toys and re-check synchrony after a few weeks. The tool gives you a before-and-after number that parents can see.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Pair two ASD learners, film a three-minute co-build, count synchronized blocks; if below 50 %, insert a visual turn card and re-film next week.
02At a glance
03Original abstract
PURPOSE: Whether children with autism spectrum disorder (ASD) exhibit interpersonal coordination deficits remains controversy. The study employed functional near-infrared spectroscopy (fNIRS) hyperscanning technology to investigate the performance of children with ASD in interpersonal cooperation and competition coordination tasks under dyadic gamed interaction situation. METHODS: 42 children (21 with ASD, 21 without ASD) performed a dyadic block-building task with collaborative intention (leading condition, following condition, turn-taking condition) (Experiment 1) and a competitive intention (Experiment 2) with an experimenter. Signal brain activation and interpersonal neural synchronization (INS) were analyzed and INS values further classified by machine-learning method and computed by using the SHAP toolkit. RESULTS: Children with ASD showed lower behavioral accuracy and behavioral synchrony in the tasks of the both two experiments; lower interpersonal neural synchronization(INS) in right temporoparietal junction (rTPJ)-right inferior frontal gyrus (rIFG), rTPJ-right inferior parietal lobule, and bilateral TPJ pairs in the three conditions under collaboration intention in experiment 1; and lower single-brain activation level in the right inferior parietal lobule (rIPL) under competition intention task in experiment 2. Finally, the children with and without ASD in Ex1 could be successfully discriminated against based on INS value by using Random Forest, AdaBoost, and XGBoost machine learning methods. CONCLUSIONS: From the interpersonal behavioral and neural synchronization, the current study demonstrated the interpersonal coordination deficits in joint action for children with ASD from cooperative and competitive intention; notably, INS is an effective neural indicator and the machine learning method provides a novel strategy for analyzing and validating INS for interpersonal coordination deficits in joint action for the population of ASD.
Journal of autism and developmental disorders, 2026 · doi:10.1093/scan/nsad031