Longitudinal associations between autistic children's anxiety and social communication differences: The moderating role of executive function behaviours.

Anxiety has been associated with social communication and interaction differences among autistic children. We sought to clarify the direction of these associations longitudinally, and test executive function as a moderator. Participants were autistic children ( = 157; 15% female, 85% male) engaged in a longitudinal study. Analyses focused on two timepoints during preadolescence ( ages 9.7 and 10.7 years). A cross-lagged panel model tested whether parent-reported anxiety at age 9 years predicted teacher-reported social communication and interaction differences at age 10 years, and vice versa. Next, multigroup analyses tested for similarity in cross-lagged pathways at different levels of teacher-reported metacognition and behavioural regulation (two facets of executive function). At each time point, respectively, 22% and 21% of children had anxiety exceeding a suggested clinical threshold. Longitudinal associations between parent-reported anxiety and teacher-reported social communication and interaction differences were not significant in the full sample. However, multigroup analyses found lower levels of parent-reported anxiety at age 9 years predicted greater teacher-reported social communication and interaction differences at age 10 years among participants with clinically elevated behavioural dysregulation. Findings indicate low levels of anxiety, coupled with behavioural dysregulation, may signify potential for increasing social communication and interaction differences observed by teachers among autistic children entering adolescence.Lay abstractAnxiety is a mental health concern affecting many autistic children, and has been linked to greater differences in social communication and interaction style. Executive functioning (i.e. the ability to direct and regulate attention and behaviour) plays an important role in autistic children's social-emotional development. We tested whether anxiety (reported by parents) predicts social communication and interaction differences (reported by teachers) over time or vice versa among autistic preadolescents. We also investigated whether the link between anxiety and social communication and interaction differed depending on children's EF abilities (reported by teachers). We found parent-reported anxiety predicted teacher-reported social communication and interaction differences a year later - but only for children who had heightened behavioural dysregulation (an aspect of executive functioning that includes impulse and emotion control). Our work suggests autistic preadolescents with behavioural dysregulation and limited anxiety may be at greater risk for social difficulties, and may need more support in this area. Executive functioning may be a useful mechanism to target in treatment for this group of children.