Language Outcomes for Children on the Autism Spectrum with Differing Language Development Profiles: What Is the Role of Nviq?
Objectives: We hypothesized that children with higher NVIQ scores with language delays or regression would have higher receptive and expressive skills. Additionally, we expected that children who developed language predominantly within normal limits would have the strongest receptive and expressive skills.
Methods: A sample of 473 children (77.6% male), aged 5-22 years (mean=9.8, SD=3.6) were included in analysis. All participants had a research confirmed ASD diagnosis using ADOS and ADI-R. Receptive and Expressive skills were assessed with Vineland Adaptive Behavior Scale, Second Edition and IQ was assessed with either the Mullen Scales of Early Learning or Differential Abilities Scales. We divided our sample into three groups: a) normal limits; b) delay; and c) regression. Assignment of groups was based on ADI-R questions regarding language development (Question Items 9, 10, and 11) . Pearson Correlations and One-Way ANOVAs were used to determine the relationship between NVIQ and receptive and expressive language skills. Additionally, Multiple Linear Regression Models were created for receptive and expressive language to assess the contribution of NVIQ and language group. All statistics were run using SPSS.
Results: Both receptive and expressive language skills were positively associated with NVIQ (receptive: r=0.183, p=0.002; expressive r=0.388, p<0.001). There were significant differences in receptive (p=0.05) and expressive skills (p<0.001) across the groups. The significant difference was only between the normal limit group and regression group (Receptive: p=0.03; expressive: p<0.001). Children in the normal limit group had the highest receptive and expressive language skills, followed by the delay group, and the regression group had the poorest language skills. Significant results were found for both regression models (Receptive: F(2,269)=4.911, p=0.009, R2=0.35; Expressive: F(2,270)=26.152, p<0.001, R2=0.162). The language category only significantly contributed the expressive language model (Receptive: Beta=-0.035, p=0.56, n.s; Expressive: Beta=-0.113, p=0.045), while NVIQ significantly contributed to both models (Receptive: Beta=0.178, p=0.004; Expressive: Beta=0.372 , p<0.001).
Conclusions: Results confirm the initial hypothesis. The only significant difference in language skills was seen between children with typical language development and children with language regression. NVIQ was a significant contributor to both regression models, while language category was only significant in the expressive language model. While NVIQ and language category accounted for some variance in language skills, a large percentage of the variance wasn’t accounted for.