Cerebellar contributions to social learning in autistic and nonautistic adolescents
Abstract
The role of the cerebellum in social learning remains poorly understood. Although the cerebellar posterior lobe (CPL) has been implicated in social decision-making and autistic traits, its contribution to social prediction and prediction-error (PE) processing during learning is unclear. We focused on adolescence, a critical developmental period for social decision-making and learning. 71 autistic (AUT) and 75 non-autistic (N-AUT) participants completed social and non-social learning tasks during fMRI. Preregistered hypotheses predicted that the CPL would be involved in prediction and PE processing, with effects differing by diagnostic status and age.
We found no group differences in social learning. Across participants, PEs decreased more steeply during non-social than social learning; in the AUT group this difference between social and non-social learning was more pronounced. In the cerebellum, PE-related activity in the CPL (lobules Crus I/II, VI, and VIIb) showed group- and age-related differences. Contrary to our hypothesis, PE-related activity during social learning was greater in AUT than N-AUT participants in right Crus II. Across several other CPL regions, PE-related activity increased with age in N-AUT participants but showed the opposite pattern in AUT participants. Mirroring the marginally steeper behavioral PE reduction in the non-social condition in the AUT vs. the N-AUT group, PE-related brain activity was greater for non-social than social PEs in left Crus I and right lobule VI in the AUT group. Exploratory searchlight analyses showed greater PE-related connectivity between left Crus II and prefrontal cortex in AUT than N-AUT participants.
These findings suggest that AUT adolescents may engage CPL differently during learning, particularly for social versus non-social information. Greater CPL activity and connectivity during social learning in AUT may reflect compensatory mechanisms supporting comparable behavioral performance to N-AUT youth.
Related
- Do Autistic Adolescents Talk Differently in Same-Sex Versus Mixed-Sex Conversations?
- Relationship between autistic traits, emotional symptoms and adverse life events in Chinese adolescents: mediating effects of resilience
- Online gaming for autistic youth: Exploring the impact on participation from the perspectives of autistic adolescents and their caregivers.
- Quality over quantity: association between autistic traits and daily-life social interactions in adolescents and young adults from the TwinssCan cohort
- Coordinated cerebellar-midbrain activity for predictive signaling during fear learning in mice
BibTeX
@misc{chen2026cerebellar,
title = {{Cerebellar contributions to social learning in autistic and nonautistic adolescents}},
author = {Yen-Wen Chen and Shannon Cahalan and Treece, Nicole Grace and Ann Clawson and Alyssa Verbalis and Christoph Korn and Gabriela Rosenblau and Allison Jack},
howpublished = {PsyArXiv},
year = {2026},
month = oct,
doi = {10.31234/osf.io/n9fau_v2},
url = {https://osf.io/preprints/psyarxiv/n9fau_v2/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Cerebellar contributions to social learning in autistic and nonautistic adolescents
AU - Yen-Wen Chen
AU - Shannon Cahalan
AU - Treece, Nicole Grace
AU - Ann Clawson
AU - Alyssa Verbalis
AU - Christoph Korn
AU - Gabriela Rosenblau
AU - Allison Jack
PY - 2026
DA - 2026/10/07
PB - PsyArXiv
DO - 10.31234/osf.io/n9fau_v2
UR - https://osf.io/preprints/psyarxiv/n9fau_v2/
AB - The role of the cerebellum in social learning remains poorly understood. Although the cerebellar posterior lobe (CPL) has been implicated in social decision-making and autistic traits, its contribution to social prediction and prediction-error (PE) processing during learning is unclear. We focused on adolescence, a critical developmental period for social decision-making and learning. 71 autistic (AUT) and 75 non-autistic (N-AUT) participants completed social and non-social learning tasks during fMRI. Preregistered hypotheses predicted that the CPL would be involved in prediction and PE processing, with effects differing by diagnostic status and age. We found no group differences in social learning. Across participants, PEs decreased more steeply during non-social than social learning; in the AUT group this difference between social and non-social learning was more pronounced. In the cerebellum, PE-related activity in the CPL (lobules Crus I/II, VI, and VIIb) showed group- and age-related differences. Contrary to our hypothesis, PE-related activity during social learning was greater in AUT than N-AUT participants in right Crus II. Across several other CPL regions, PE-related activity increased with age in N-AUT participants but showed the opposite pattern in AUT participants. Mirroring the marginally steeper behavioral PE reduction in the non-social condition in the AUT vs. the N-AUT group, PE-related brain activity was greater for non-social than social PEs in left Crus I and right lobule VI in the AUT group. Exploratory searchlight analyses showed greater PE-related connectivity between left Crus II and prefrontal cortex in AUT than N-AUT participants. These findings suggest that AUT adolescents may engage CPL differently during learning, particularly for social versus non-social information. Greater CPL activity and connectivity during social learning in AUT may reflect compensatory mechanisms supporting comparable behavioral performance to N-AUT youth.
ER -