Expressive Language and Social Functioning in Children with Specific Language Impairment Versus High Functioning Autism

Background:  The language phenotype of individuals with autism spectrum disorders (ASD) is heterogeneous, although most findings suggest subgroups of individuals with ASD experience language deficits. Evidence has emerged showing that cognitive and specifically, language profiles, between youth with ASD versus those with specific language impairment (SLI) are distinct. Yet, children with SLI share various social characteristics, including elevated problems with social communication and prosocial gestures.

Objectives: The purpose of this study was to characterize the language and social phenotypes in children with SLI versus high functioning autism (HFA), and to determine whether neural correlates of language processing are differentially associated with interpersonal functioning across the neurodevelopmental disorders.

Methods: Participants were 13 children with SLI, 13 with HFA, and 13 typically developing youth matched on age. The CELF-4, PPVT-3, EOWPVT, and Letter-Word Identification and Spelling subtests from the Woodcock Johnson, 3^rdEdition were administered to determine functioning across language domains. The Social Responsiveness Scale (SRS) was completed by caregivers to social functioning. Participants completed MRI and DTI scans to assess for any abnormalities in neural substrates (pars opercularis, pars triangularis, superior temporal gyrus) and circuits involved in language functions (superior longitudinal fasciculus).

Results:

Multivariate analysis of variance (MANOVAs) were performed to assess between-group differences in language domains, and DTI measures (mean diffusivity, MD) and cortical thickness/area, with whole brain fiber, cortical thickness, and area as covariates respectively. Results showed group differences across CELF-4 Receptive and Expressive Language, Letter-Word Identification, and Spelling, Fs> 6.05, ps<0.01, but not receptive and expressive vocabulary knowledge (PPVT-3, EOWPVT). SLI group performed more poorly than the other two groups in CELF-4 Receptive Language and Letter-Word Identification, but comparable to those with HFA in CELF-4 Expressive Language and Spelling. Group differences were observed across SRS subscales, including Awareness, Social Cognition, Social Communication, Motivation, and Autistic Mannerisms, Fs>7.96, ps<0.001. Generally, those with HFA showed more difficulties across each subscale than TD and SLI groups, with the exception of Social Awareness and Motivation, where HFA and SLI children were comparable. No group differences were observed in cortical volume/area, and MD.

Pearson correlations were computed per clinical group to assess associations between Total SRS score with language measures, and with MRI/DTI indices. No associations were observed in the TD and SLI groups. However, among children with HFA, greater social impairment was related to poorer performance on the EOWPVT and CELF-4 Receptive and Expressive Language. Additionally, among those with HFA, cortical thinning of the right superior temporal gyrus, r(13)=-0.70, p=0.008, and lower MD of the left superior longitudinal fasciculus to the parietal region was related to higher SRS score, r(11)=-0.61, p=0.04, reflecting greater social deficits.

Conclusions: In children with HFA, but not SLI or TD, social difficulties were related to their language performance, and structural abnormalities of neural substrates and circuits involved in language comprehension and visual-spatial attention. Results suggest that the etiology of the social phenotype associated with SLI versus ASD youth are likely distinct.