Linguistic Prediction in Autism Spectrum Disorder

Area of Science:

• Neuroscience
• Psychology
• Linguistics

Background:

• Autism spectrum disorder (ASD) is linked to domain-general predictive ability impairments.
• Difficulties in predicting event outcomes affect navigating dynamic environments in ASD.
• The domain-specificity of these predictive deficits, particularly in language, remains under-researched.

Purpose of the Study:

• To investigate atypicalities in linguistic prediction in individuals with ASD.
• To determine if predictive impairments extend to language processing in ASD.
• To explore the relationship between autistic traits and linguistic prediction.

Main Methods:

• Utilized a cloze probability task to assess linguistic prediction in individuals with ASD and typically developing controls.
• Experiment 1 involved an offline cloze task with 33 individuals with ASD and 64 controls.
• Experiment 2 included a lab-based cloze task with 19 individuals with ASD and 22 controls, measuring voice onset time.

Main Results:

• No significant difference in cloze probability between individuals with and without ASD.
• Higher levels of autistic traits correlated with a tendency towards lower-probability (non-modal) cloze responses.
• Individuals with ASD exhibited significantly slower voice onset times (~200 ms) compared to controls.

Conclusions:

• Individuals with ASD demonstrate inefficiencies in linguistic prediction.
• Specific autistic traits are associated with these linguistic prediction inefficiencies.
• Slower speech production, indicated by voice onset time, is a notable finding in ASD.