Probabilistic Learning of Cue-Outcome Associations is not Influenced by Autistic Traits.

This research looked at whether people with more autistic traits have trouble learning patterns where outcomes aren't always predictable (like learning that a grey cloud sometimes means rain). The study tested adults with varying levels of autistic traits and found no differences in their ability to learn these patterns. This challenges theories that suggest autistic people struggle with this type of learning.

This study examined whether autistic traits influence the ability to learn probabilistic cue-outcome associations, testing predictions from Bayesian/predictive coding models of autism. Across two experiments with adult participants varying in autistic traits, researchers assessed learning when participants were explicitly instructed or implicitly exposed to complex cue-outcome relationships requiring inference of relevant cues. Contrary to theoretical predictions, no differences in probabilistic learning accuracy were found based on autistic traits levels. The findings challenge current Bayesian/predictive coding models of autism, which suggest autistic individuals should experience difficulties with probabilistic learning in uncertain environments.

Results suggest probabilistic learning abilities may not differ based on autistic traits, potentially informing educational and therapeutic approaches. Findings challenge theoretical assumptions about learning differences in autism, warranting reconsideration of intervention strategies based on predictive coding models.

Sample size not reported. Study limited to adults, unclear generalizability to children or diagnosed autistic individuals. Study design type not specified. Findings based on trait measures rather than clinical diagnosis.

According to Bayesian/predictive coding models of autism, autistic individuals may have difficulties learning probabilistic cue-outcome associations, but empirical evidence has been mixed. The target cues used in previous studies were often straightforward and might not reflect real-life learning of such associations which requires learners to infer which cue(s) among many to track. Across two experiments, we compared adult learners with varying levels of autistic traits on their ability to infer the correct cue to learn probabilistic cue-outcome associations when explicitly instructed to do so or when exposed implicitly. We found no evidence for the effect of autistic traits on probabilistic learning accuracy, contrary to the predictions of Bayesian/predictive coding models.

Implications for the current Bayesian/predictive coding models are discussed.