We are proud to congratulate Laurina Fazioli on successfully defending her doctoral dissertation.
Title:
Investigation of perceptual decision-making in autism spectrum disorder
Advisors:
Dr. Amit Yashar and Pr. Bat-Sheva Hadad
Summary:
Autism Spectrum Disorder (ASD) is a neurodevelopmental condition affecting various aspects of behavior, including difficulties in social communication, repetitive behavior, and restricted interests. Atypical sensory processing has been increasingly recognized as a core feature, yet the mechanisms underlying atypical perception remain unclear. Previous research largely focused on perceptual skills (e.g., sensitivity), overlooking perception as an integrative process. Perceptual decision-making—the process of making inferences based on sensory information—lies at the interface between sensory processing and behavior, and provides a systemic model to investigate perception at multiple levels: (1) perceptual inference (first-order decision), and (2) confidence in this inference (second-order decision, reflecting metacognitive abilities). However, perceptual decision-making in autism has received little interest.
Here, we aimed to investigate perceptual decision-making in autism within the Bayesian perception framework—positing that perception results from combining sensory uncertainty, prior knowledge, and reward information. Specifically, we asked: (1) To what extent does first-order decision in autism incorporate decision components (i.e., sensory uncertainty, prior, and reward)? (2) To what extent is higher-level metacognitive decision-making in autism contingent on first-order decision components?
In three experiments—each manipulating one Bayesian information—autistic (n = 59) and non-autistic (n = 83) participants performed an orientation categorization task, reporting stimulus category (first-order task) and decision confidence (second-order task). We manipulated sensory uncertainty, prior, and reward by varying stimulus contrast (Experiment 1), category probability (Experiment 2), and points per correct answer (Experiment 3). Using psychophysics and computational approaches, we quantified the effects of Bayesian information on first- (i.e., sensitivity, decision criterion), and second-order (i.e., decision confidence) decisions.
Both groups showed comparable first-order performances—as they exhibited similar sensitivity and criterion shift—demonstrating a suboptimal, but typical integration of each Bayesian information during first-order perceptual decision-making. However, while non-autistic participants displayed constant metacognitive abilities across experiments, autistic metacognitive abilities depended on the Bayesian information biasing their perceptual decision. Specifically, autistic participants demonstrated enhanced metacognitive abilities when first-order decisions were adjusted from sensory evidence alone.
Contrary to dominant views suggesting atypical first-order perceptual inferences in autism, our findings indicate that qualitative differences in higher-order—rather than lower-level—perceptual processes may constitute a core component of autistic perception, and shape the way autistic individuals engage with sensory input. These findings have critical implications in the understanding of core mechanisms of autism, with relevance beyond autism research, such as diagnostic and rehabilitation domains.
Congratulations Laurina!
Yashar Lab 