Social Bayes: Using Bayesian Modeling to Study Autistic Trait-Related Differences in Social Cognition

Area of Science:

• Neuroscience
• Cognitive Science
• Psychology

Background:

• Autism spectrum is marked by social interaction impairments.
• Underlying subpersonal processes of social reciprocity deficits remain debated.
• Autism may involve altered brain inference of socially relevant sensory signals.

Purpose of the Study:

• Investigate the role of social cue integration in learning.
• Examine how autistic traits influence the processing of social and non-social cues.
• Utilize computational modeling to understand learning deficits in autism spectrum.

Main Methods:

• Employed a novel reward-based learning task.
• Integrated nonsocial and social cues with computational modeling.
• Assessed 36 healthy subjects using the Autism-Spectrum Quotient (AQ).

Main Results:

• Higher AQ scores correlated with poorer task performance (r = -0.39).
• Computational modeling revealed a learning deficit in high AQ scorers.
• High AQ scorers failed to integrate social context to adapt belief precision for nonsocial cues.

Conclusions:

• Pronounced autistic traits linked to misintegration of social cues.
• Performance differences are due to how social information infers nonsocial cues, not stimulus processing.
• Findings highlight the impact of autistic traits on social information integration in learning.