22767
Comparing Structural and Pragmatic Language Abilities in Subgroups of ASD Children with and without Comorbid Intellectual Disability

Friday, May 13, 2016: 5:30 PM-7:00 PM
Hall A (Baltimore Convention Center)
A. Whitten1 and J. W. Bodfish2, (1)Vanderbilt University, Nashville, TN, (2)Vanderbilt University School of Medicine, Nashville, TN
Background: Currently, the general consensus is that structural features of language (e.g., syntax, phonology) are intact while the social use of language, referred to as pragmatic language, is universally impaired in ASD. However, most of the research on language impairment in autism has been restricted to high-functioning individuals with autism. Consequently, it is not known whether a comorbid intellectual disability further affects language impairment in low-functioning individuals, and thus may require more extensive interventions. Furthermore, studies that directly compare structural and pragmatic language are necessary to advance knowledge of the language phenotype within all of ASD. 

Objectives: We aimed to compare both structural and pragmatic aspects of language across three groups: low functioning children with ASD (LFA), high functioning children with ASD (HFA), and an age-matched typically developing group (TD). 

Methods:  Using the Children’s Communication Checklist-2 (CCC-2), we compared mean scores between LFA (n=15, Nonverbal IQ < 80, mean age = 12 years, range = 5 – 17 years), HFA (n=57, Nonverbal IQ  > 80, mean age = 10 years, range = 6 – 17 years), and TD (n=52, mean age = 11 years, range = 6 – 17 years). We used one-way ANOVAs with planned post-hoc tests across 4 subscales measuring structural language (Speech, Syntax, Semantic, Coherence) and 4 subscales measuring pragmatic language (Initiation, Scripted Language, Context, Nonverbal Communication).

Results: There was a significant main effect of group for all four structural language subscales (F (2, 21) = Speech: 18.75, Syntax: 49.49, Semantic: 93.34, Coherence:101.80, all p < .001). Post hoc Tukey tests revealed the same pattern for all four structural subscales (p < .001), such that there were no significant differences between LFA and HFA, while both LFA and HFA groups scored significantly lower than TD. This same pattern was replicated across all four pragmatic subscales, with a main effect of group (F (2, 66) = Initation:  61.32, Scripted Language: 64.58, F(2, 121) = Context: 175.04, Nonverbal Communication: 170.12, all p < .001). Again, post hoc Tukey tests showed no significant differences between LFA and HFA, with both LFA and HFA scoring significantly lower than TD. Pearson correlations between IQ and structural and pragmatic CCC-2 subscales found no significant associations, with one exception. The pragmatic subscale ‘Context’ was significantly associated with IQ within the LFA group only (r = 0.60, p < .05). 

Conclusions: We found evidence that the language deficits do not differ between LFA and HFA groups. Furthermore, our results refute the idea that structural features of language are spared in ASD, as both LFA and HFA groups tended to score well below the TD group on all structural language subscales. Our results suggest that future studies on language impairment in ASD should no longer exclude low functioning individuals, and examine both structural and pragmatic aspects of language; such research will be vital to improving language interventions in ASD, especially for children with ASD whose language deficits do not respond to intervention.