Refining the Scope of Language Impairment in Individuals with ASD through the Assessment of Complex Sentences

Background: Language is developing atypically in all individuals with ASD, although there is considerable variation in exactly which aspects of language are affected and whether this results in a Language Impairment(LI, Kjelgaard & Tager-Flusberg, 2001). Recent cross-linguistic experimental work has shown that an atypical development of complex sentence comprehension emerges in individuals with ASD (Durrleman et al., 2016; Prévost et al., 2017; Riches et al., 2011), which is often related to a history of (potentially undetected) LI (Modyanova et al., 2017). Object-initial sentences, i.e., sentences that alter the subject-verb-object default word order of German, can be disambiguated by morphological (i.e., case and/or number) marking. Case-marking on nouns was shown to enhance the processing of object-initial sentences in neurotypical children before number does, but the co-occurrence of both case and number supports processing abilities in children with LI (Stegenwallner-Schütz & Adani, 2018).

Objectives: We provide a fine-grained assessment of complex sentence processing abilities in individuals with ASD and an exploration of which morphological properties may alleviate their expected processing difficulties. Towards this end, we manipulated number- and case-marking on two types of object-initial sentences in German: Object-Verb-Subject (OVS, e.g., The dog kicks the zebra [English translation], in which zebra is interpreted as the agent of the action expressed by the verb kick) and Object-Relative Clauses (ORC, e.g., Where is the dog that the zebra is kicking [English translation; notice that number- and case-marking are not reported in the English translation]).

Methods: German-speaking individuals with ASD (N=26 [2 girls]; mean age: 11;10 [y;m]; age range: 6;10-15;3) and their neurotypical age-matched controls participated in a picture-selection task (OVS-study) and in a referent identification task (ORC-study). Test and filler sentences were presented auditorily and the participant had to select the target picture/referent. Based on a detailed battery of psycholinguistic standardized tests tapping into expressive/receptive vocabulary, morphology and syntax, we identified a subgroup of individuals with ASD with co-occurring LI.

Results: The accuracy results reveal an impaired comprehension of OVS sentences (proportion correct: 0.63, SE: 0.04) and of ORC (proportion correct: 0.25, SE: 0.05) by individuals with ASD with co-occurring LI, compared to neurotypical controls. While case-marking enhanced the accuracy of ORC (both in the ASD and the neurotypical groups), number-marking revealed a facilitative effect neither on OVS nor ORC in the ASD group. The performance by the group of individuals with ASD without LI and the neurotypical controls did not differ on any of the comparisons. The reported effects are significant at the statistical significance level of p<0.05.

Conclusions: OVS and ORC sentences are complex linguistic structures, whose correct comprehension appears to be vulnerable in individuals with ASD, particularly when this condition co-occurs with LI. The apparent lack of deployment of number-marking during OVS processing also qualifies as a risk factor, which is not attested in neurotypical development. In contrast, the sensitivity to case can be identified as one of their strengths. Our results highlight the effectiveness of psycholinguistic investigations in the identification of risk and protective factors during language development in ASD.