Increased variability but intact integration during visual navigation in Autism Spectrum Disorder.

Area of Science:

• Neuroscience
• Computational Psychiatry
• Developmental Psychology

Background:

• Autism Spectrum Disorder (ASD) is a neurodevelopmental condition impacting various functions.
• Computational models suggest aberrant Bayesian inference may underlie ASD phenotypes, but results are conflicting.
• Existing research often uses simplified tasks, limiting understanding of complex ASD behaviors.

Purpose of the Study:

• To investigate Bayesian inference in Autism Spectrum Disorder (ASD) using a naturalistic visual path integration task.
• To examine how continuous action and active sensing influence belief states in individuals with and without ASD.
• To determine the computational basis of spatial navigation errors in ASD.

Main Methods:

• Employed a naturalistic visual path integration task involving continuous movement and active sensing.
• Tracked participants' dynamic belief states during the task.
• Analyzed spatial errors, velocity estimation, speed priors, and belief variability in control and ASD groups.
• Assessed the impact of performance feedback on error reduction.

Main Results:

• Both control and ASD groups exhibited overshooting biases in radial distance and angular eccentricity.
• Errors in both groups were primarily driven by misestimated velocity signals due to nonuniform speed priors, not integration deficits.
• No significant difference in speed priors was observed between groups.
• The Autism Spectrum Disorder (ASD) group displayed heightened variability in belief states compared to controls.
• Endpoint variance and trajectory irregularities correlated with ASD symptom severity.
• Performance feedback significantly reduced variance in the ASD group, bringing them closer to control performance.

Conclusions:

• Naturalistic tasks are crucial for understanding Autism Spectrum Disorder (ASD) phenotypes.
• Heightened variability, rather than altered priors, may characterize computational differences in ASD during path integration.
• These findings suggest a broader computational perspective is needed to fully understand ASD pathology.
• Interventions improving belief state stability could benefit individuals with ASD.