Walking Onset Predicts Rate of Language Growth in Children with Autism Spectrum Disorder

Saturday, May 17, 2014: 10:30 AM
Imperial A (Marriott Marquis Atlanta)
R. Bedford1, A. Pickles1 and C. Lord2, (1)King's College London, London, United Kingdom, (2)Weill Cornell Medical College, White Plains, NY
Background: Motor milestones such as the onset of walking are important developmental markers, not only for later motor skills but also for more widespread social-cognitive development. Early impairments in gross motor skills, including delayed walking onset, atypical gait and postural stability, have been shown in children with autism spectrum disorder (ASD) and it is possible that such difficulties have knock on consequences for language development.

Objectives: The aim of the current study was to test whether parent-reported onset of walking predicted the subsequent rate of expressive and receptive language development from 2 – 9 years in a large cohort of children with ASD controlling for non-verbal IQ and autism symptom severity.

Methods: We ran growth curve models for expressive and receptive language from the Vineland Adaptive Behavior Scales measured at 2, 3, 5 and 9 years in 209 autistic children. Walking onset (as reported in the Autism Diagnostic Interview-Revised; ADI-R) was included as a predictor of the slope of language development with non-verbal IQ (visual reception; VR, subscale of the Mullen Scales of Early Learning) and autism symptom severity (ADI-R total item score) as covariates. A high proportion of the children showed floor effects of the VR T-score. To characterize the variability in scores at this low end of the scale we also regressed the log of VR-T scores on the log of VR Age Equivalent scores and age at visit 2 and imputed these predicted values.

Results: Parent reported age of walking onset significantly predicted the subsequent rate of both receptive (ß = -0.09, S.E. = 0.04, p = 0.03) and expressive (ß = -0.1, S.E. = 0.04, p = 0.03) language development after controlling for VR T-score and severity of ASD symptoms. When the more stringent statistical model accounting for floor effects in VR T-scores was applied these relationships became non-significant (receptive: ß = -0.06, S.E. = 0.04, p = 0.16; expressive: ß = -0.06, S.E. = 0.04, p= 0.14), although the trends remained in the same direction. For both receptive and expressive language, scores were similar across early and late walkers when children were young but differed in the oldest children, with lower language scores in the late walkers.

Conclusions: As in typical development, results indicate that walking onset is related to subsequent receptive and expressive language development in ASD. Onset of walking appears to influence the rate of language learning, rather than causing an initial boost in language which then persists over time. This suggests that motor delays in young children with ASD can have a longitudinal cross-domain influence, potentially contributing, in part, to the linguistic difficulties which characterise ASD. Various mechanisms have been proposed to underlie this relationship in typical development, including increased joint attention episodes, having hands free to point and gesture, and cerebellar development. Future work is required in children with autism to understand the likely bidirectional interactions between motor and language pathways across development and to elucidate the underlying mechanisms.