Relations Among Parent-Reported and Spontaneous Gestures, Fine Motor, and Language in Young Children with ASD: A Structural Equation Model Approach

Background: Gestures and fine motor skills have been found to be related to language outcomes in young children with and without autism spectrum disorder (ASD). Evidence suggests some measurement variability between parent report and direct observation of gestures, resulting in research using either one mode of measurement or creating a standardized composite score for overall gesture use to predict language outcomes. To our knowledge, studies have not examined gesture use as an underlying, latent construct in a structural equation modeling framework, a powerful method of analysis that captures the variability and covariance of many individual components.

Objectives: This study examined the relationship among gesture, fine motor ability, and language in young children with ASD using structural equation modeling. 

Methods: A total of 197 children were classified into three groups: ASD (n=110), non-spectrum developmental delay (DD; n=35), and typical controls (TYP; n=52) based on an evaluation conducted between the ages of 12 to 48 months (M=34.55, SD=8.41). Behavioral assessments included the Mullen Scales of Early Learning (MSEL) and the Autism Diagnostic Observation Schedule (ADOS), and parent-report measures included the Communication and Symbolic Behavior Scales Developmental Profile Caregiver Questionnaire (CSBS-CQ), the Vineland Adaptive Behavior Scales, Second Edition (VABS-II), and the MacArthur-Bates Communicative Development Inventories (CDI). CSBS-CQ and ADOS gesture items, MSEL and VABS-II age equivalents for fine motor, receptive, and expressive language, and CDI words understood/produced were used in analyses.

Results: A series of confirmatory factor analyses (CFA) was first run that found individual gesture items on the CSBS-CQ and the ADOS loaded significantly onto underlying gesture use. Next, a model testing the correlations between the factor scores from the CSBS-CQ, ADOS, and Fine Motor (MSEL, VABS-II) across all children revealed a strong, positive correlation (r=.70) between fine motor and parent-reported gestures (CSBQ-CQ), and strong, negative correlations between spontaneous gesture use (i.e., fewer gestures result in higher scores in ADOS) and both fine motor (r=-.52) and parent-report of gestures (r=-.80) (Figure1). The fit indices for the overall model revealed good fit; however, when models were run separately for the ASD and non-ASD groups (DD+TYP), the results indicated good model fit only for the ASD group. Finally, a model was run to identify if the CSBS-CQ, ADOS, and Fine Motor domains loaded significantly onto one, latent construct of “gestures” to predict concurrent language outcomes (MSEL and VABS-II, CDI). For expressive language, the model did not fit across the entire sample, or in either grouping. However, for receptive language, the ASD-only model revealed good fit, resulting in better overall gesture use predicting better performance on receptive language measures (Figure2).

Conclusions: An underlying construct of gesture use was found among parent-reported and spontaneous gestures, as well as fine motor, which differs from the traditional method of analyzing each domain separately or as with composite scores. Overall gesture use predicted concurrent receptive language in ASD, but not expressive language. The predictive model was not found in the non-ASD group, which may be due to reduced variability among gesture use.