GracLab at INSAR 2018

On May 13, 2018 by Dominika Slusna

This year GracLab seizes the opportunity to present its ongoing research on autism at Autism Research Meeting in Rotterdam (INSAR 2018) organized annually by Institute for Autism Research.

Our three pre-doctoral researchers – Kristen Schroeder, Elisabet Vila, and Dominika Slušná – prepared posters that summarize the results of their respective studies on Language and Communication in ASD. Find the abstracts further down on this page.

Comprehension of Interrogative Questions in Autism Spectrum Disorder. (E. Vila)

Individuals with Autism Spectrum Disorder (ASD) have difficulties to initiate and follow a
conversation. Despite the importance of asking and answering questions for social interaction, how
ASD individuals comprehend the language structures involved in them has been understudied.
Prosody, on the other hand, has a role in shaping questions to the point that there are languages such
as Spanish where yes/no questions differ from declaratives only prosodically. As for wh questions,
ASD children understand them much later than typically developing (TD) children (Goodwin et al.,
2012). They are also more difficult to answer than yes/no questions (Oi, 2010; Oi and Tanaka,
2011). The difficulty in answering wh questions in ASD should be clarified since the assessment of
the understanding of false-belief (Sally-Anne task) and other explicit Theory of Mind (ToM)
experiments is usually made through wh questions. In this study, we examined the comprehension of
wh and yes/no questions through visual and non-visual tasks, with and without false-belief content.
We selected 16 high-functioning ASD children from 7 to 12 years old (mean = 10.03 years old). All
of them were Spanish-Catalan bilingual with Spanish dominance. They were individually matched
by verbal mental age with a group of children in the same age range (mean= 9.79 years old). No
significant differences were found between the TD and the ASD group. ASD participants performed
better in Non-ToM than ToM tasks and in those with visual support. In addition, ASD participants
have greater difficulties to answer wh rather than yes/no questions.

Use of communicative gestures in school-age children and adolescents with non- or minimally verbal autism. (D. Slušná)

Language does not develop in either production or comprehension in at least 25{66237868edf11cf69973b9f41e73a2f4ef75e75b20cb8da5be86bd8152991cd8} of individuals on
the autism spectrum, who form its most severe and understudied end. We aimed to shed light on
language-gesture relationships by profiling production rates and types of nonverbal communicative
gestures in this population. Standardized measures of nonverbal IQ (as measured by Leiter-R) and
nonverbal symbolic cognition (as measured by ComFor) were also taken. Language status as a
recruitment criterion was confirmed in both production and comprehension through diagnostic
interviews (ADOS and ADIR), teacher informants, and standardized language tests. There was no
evidence of apraxia in this sample. Results showed the complete absence of declarative gestures
including distal pointing, and even distal imperative pointing was scarce at the group level, as were
iconic gestures. Overall gesture production rate was low as well, although in a study run in parallel,
mean-rank distributions of gesture rates in typically developing pre-verbal infants between 9 and 22
months were interestingly not significantly different. The proportion of bimodal communicative acts
(gesture + vocalization) was significantly lower than gesture-only acts. 79 {66237868edf11cf69973b9f41e73a2f4ef75e75b20cb8da5be86bd8152991cd8} of individuals scored
≤70 in nonverbal IQ. Neither language nor gesture scores correlated with levels of nonverbal
symbolic cognition, which in turn were significantly correlated with nonverbal IQ.

Linguistic Markers in the Narratives of Children on the Autism Spectrum (K. Schroeder)

Narratives are ecological means to elicit speech in clinical populations- both for assessing formal aspects of language as well as how language construes a shared reality of agents and events. Previous literature on narratives in autism spectrum conditions (ASC) have shown that referential anomalies distinguish children with ASC from clinical groups like ADHD or typically-developing peers (Banney et al., 2015; Rumpf et al., 2012) and may relate to ASC symptom severity (Suh et al., 2014). However, this previous work has not looked at reference fully systematically. In the present study, we catalogue referential constructions and errors though a linguistic lens and explore the relation between qualitative deficits in narratives and specific grammatical profiles.

Nineteen children with ASC without intellectual disability (mean age: 9.8, range: 7.3 – 12.6) and nineteen verbal-IQ and chronological age matched controls participated in the storybook narration task. The narratives were annotated for various grammatical constructions across nominal, verbal, and clausal domains as well as potential errors and rated for story completeness. We predicted that the ASC group would use proportionally less anaphoric devices and produce greater referential errors which would relate to qualitative storytelling measures. We also predicted that the narratives would be less grammatically complex and less complete.

Our results show that the children with ASC produced less complete narratives, however this didn’t relate to referential richness, rather relating to verbal IQ (r= .546; p=.016). This correlation was not found for the TD group (r=.270; p=.264). Children with ASC produced fewer definite anaphoric NPs, however this difference was not significant (p=.234). The ASC group produced more nominal errors (p=.015), being primarily driven by wrong lexical selection (p=.001), counter to our predictions. Children with ASC produced less grammatically complex constructions with fewer relative clauses (p=.026) but produced similar levels of mental state or emotional content (p=.525). These results show that even at the high-functioning end of the spectrum and controlling for verbal IQ, linguistic markers based on fine-grained grammatical measures can detect ASC against typically developing age-matched controls.

Leave a Reply

Your email address will not be published. Required fields are marked *