Abstract
P150. Computational Characterization of Social Inference Deficits Associated With Autism Traits During Observational Learning
Biological psychiatry (1969), Vol.91(9 Supplement), pp.S147-S147
05/01/2022
DOI: 10.1016/j.biopsych.2022.02.384
Abstract
Background
One's ability to infer the goals and intentions of others is crucial for social interactions and, more generally, to function in society at large. This skill set varies broadly across individuals, and more severe deficits are commonly associated with autism spectrum disorders (ASD), as well as other comorbidities such as social anxiety. However, despite the importance of social inference, the underlying computational principles and the link between their disruption and individual variability remain poorly understood.
Methods
Here, we provide a computational characterization of social learning via a large-scale online study in which adult participants from the general population (N=901) completed an observational learning task that involves a tradeoff between two strategies: imitation (repeat the observee’s most recent action) and emulation (infer the observee’s goal). Self-report questionnaires were collected to assess symptom dimensions relevant to psychopathology, including the Social Responsiveness Scale (SRS) as an index of autism-like traits. Using computational modeling, we quantified individual tendencies to rely on imitation versus emulation learning.
Results
We find that more severe autism symptoms were associated with a deficit in social goal inference (lower bias towards emulation; R=-0.152, P<0.001). Follow-up analyses indicate that this association holds when controlling for other model parameters, such as decision noise and heuristic strategies.
Conclusions
Our findings suggest that social inference deficits typically observed in autism may be related to a difficulty in inferring other people’s goals and intentions when learning from observation. Further investigations in diagnosed ASD patients are warranted to confirm the clinical relevance of this finding.
Details
- Title: Subtitle
- P150. Computational Characterization of Social Inference Deficits Associated With Autism Traits During Observational Learning
- Creators
- Caroline Charpentier - California Institute of TechnologyQianying Wu - California Institute of TechnologySarah Oh - California Institute of TechnologyJamie Feusner - University of TorontoReza Tadayonnejad - UCLA Semel Institute for Neuroscience and Human BehaviorJeffrey Cockburn - California Institute of TechnologyJohn O'Doherty - California Institute of Technology
- Resource Type
- Abstract
- Publication Details
- Biological psychiatry (1969), Vol.91(9 Supplement), pp.S147-S147
- Publisher
- Elsevier Inc
- DOI
- 10.1016/j.biopsych.2022.02.384
- ISSN
- 0006-3223
- eISSN
- 1873-2402
- Language
- English
- Date published
- 05/01/2022
- Academic Unit
- Psychological and Brain Sciences
- Record Identifier
- 9984696834502771
Metrics
1 Record Views