Abstract: Individuals with autism spectrum disorder (ASD) have been associated with impaired prediction, which affects the adaptation to the changing world. However, it is unclear whether this deficit is domain-general or -specific. Language and music provide an excellent way to investigate prediction, as both domains involve structured temporal sequences in which expectations are based on implicit learning of combinatorial principles. This study examined prediction in music and language in 31 Mandarin speakers with ASD and 33 matched controls. Participants were asked to produce the final note/word after hearing an unfinished melody/sentence in a melodic cloze task and a sentence completion task. Results indicated comparable performance between the two groups on the melodic cloze task, while controls produced more words that belonged to the most frequent responses than individuals with ASD. The predictive constraints of the melodic/sentence stems (i.e., probabilities of responses based on normative data) showed a positive effect on the produced final notes/words, with more frequent responses being produced following high-probability stems than low-probability stems. These findings suggest that the impairment in linguistic prediction in ASD may not be due to generalised problems with prediction in any type of complex sequence processing. Future studies are required to investigate how and why such a dissociation may occur in ASD and its implications for theories of prediction.
OSF does not support the use of Internet Explorer. For optimal performance, please switch to another browser.