Communication Propels Growth in Early Speech Production for Young Children with Autism: A Mediation Model

Background: Language abilities of children with autism spectrum disorders (ASD) are heterogenous (Tager-Flusberg, 2004), with delays in gesture production (Iverson & Wozniak, 2007) and joint attention (Wetherby et al., 2007) aligning with concurrent and subsequent language development deficits (Iverson & Goldin-Meadow, 2005; Luyster et al., 2008). Distinct cognitive and linguistic phenotypes have been defined in prior within-disorder comparisons of language (Wittke et al., 2017), informing anticipated patterns of later language development. However, the unique contributions of social communication elements in driving these differential patterns of language growth remain unclear.

Objectives: This study investigates the unique roles of social communication elements including emotion and eye gaze, communication, and gesture, in mediating growth in spoken language production between toddlerhood and pre-school to better inform anticipated developmental trajectories for children demonstrating specific patterns of ASD behaviors.

Methods: This study utilized a subset of secondary data originally collected through the Autism Phenome Project (APP, N=55, 42 male, mean age at initial visit=33.9 months, SD=5.5), a longitudinal study completed through the MIND Institute at the University of California-Davis. The APP intends to define clinically meaningful ASD subtypes based on behavioral and biological data, specifically evaluating children immediately following ASD diagnosis at age 24-40 months and again at follow-up evaluations completed approximately 3 years later. We performed simple linear regression analyses, using Communication and Symbolic Behavior Scales, Developmental Profile (CSBS-DP; Wetherby & Prizant, 2002) Behavior Sample Social Composite scores (see Table 1) at the age of diagnosis to predict growth in mean length of utterance (MLU) approximately 3 years later.

Results: Performance on the Social Composite partially mediated the positive relationship between MLU at toddler age and MLU at preschool age, driven by performance on the Communication scale above and beyond performance on the Gesture scale and Emotion and Eye Gaze scale (see Table 2 for detailed linear regression results). That is, total effect of MLU at toddler age on MLU three years later while controlling for Communication scale performance increased when this mediator was included in the model. When models included Gesture or Emotion and Eye Gaze scales alone, statistically significant mediation effects were not indicated.

Conclusions: These findings distinguish social communication factors including rate, behavior regulation, social interaction, and joint attention as key elements facilitating growth in MLU during the early language acquisition period from toddlerhood to preschool age. Communication behaviors may lead to increased turn-taking and engagement opportunities with peers, suggesting that MLU growth is more child-input dependent than behaviors which summon adult-input like gesture and emotion and eye gaze. Next analytic steps for this research endeavor include analyzing direct and indirect mediational pathways, and assessing changes in these pathways with this sample, as well as a larger sample, when additional social communication elements are incorporated. Further, subsequent studies should incorporate analysis of additional spontaneous speech elements and Autism Diagnostic Observation Schedule (ADOS-G; Lord et al., 1999) performance at both time points in order to identify unique, clinically meaningful phenotypes based on language form, language function, lexical diversity, and ASD-behavior severity.