Objectives: We used eyetracking methodology to investigate ongoing sentence comprehension and determine the effect of sentence context on spoken word identification as a function of (a) participants' autism diagnosis and (b) their language ability.
Methods: We tested 24 adolescents meeting ADOS criteria for autism spectrum disorder and 24 non-autistic adolescents closely matched on language ability. Eye-movements were recorded as participants listened to sentences, responding when they heard a word that matched one of four objects on a computer display.
Results: Eye-movements were affected by the phonological overlap between the spoken words and the names of objects in the display. For example, on hearing the word “hamster”, participants looked more at a picture of a hammer than at other unrelated objects. However, this effect was significantly reduced if the sentence context made the ‘competitor’ (hammer) an unlikely referent, as in the sentence “Joe stroked the hamster”. Contrary to predictions, the effect of context on eye-movements was comparable across groups. Instead, regardless of diagnosis, individuals with poorer language skills showed reduced influence of context on eye-movements.
Conclusions: Our results contradict previous findings supporting the WCC account, suggesting that context-processing difficulties are related to language impairment rather than being autism specific. Together, these findings point towards a more refined version of WCC. This in turn has implications for theories regarding the neural basis of autism.