Atypical sensory perception is widely observed in individuals with autism spectrum disorder (ASD). Differences in perceptual integration, in which sensory information is combined over time, have been particularly well documented. While several models of altered sensory processing have been proposed to account for differences in perceptual integration in ASD, the underlying cognitive mechanisms are unknown.

In this project, Benjamin Scott plans to apply recent developments in computational neuroscience, game-based psychophysics and cognitive psychology to characterize differences in perceptual integration across people with ASD and to distinguish between alternative cognitive models. Their approach involves the application of a game-based task in which sensory information is presented in brief, randomly timed pulses. This game is inspired by pulse-based accumulations tasks, which were developed in animal models to distinguish perceptual errors that result from noisy sensory input, atypical working memory and other aspects of the integration process. Using this task, the researchers plan to collect a large behavioral data set from children and adolescents with ASD as well as neurotypical individuals.

An important element of this online game-based task is the incorporation of a feedback-based training pipeline. This feature will enable Scott and his collaborators to evaluate performance in minimally verbal people with ASD who are traditionally underrepresented in psychophysical studies. To gain mechanistic insight, the research team will characterize perceptual integration in each individual using a combination of signal detection theory and sequential sampling models of the decision process. These models can be fit to each individual’s behavioral data and will provide both population-level and individual-specific estimates of latent parameters such as sensory noise, memory stability and decision thresholds that govern perceptual decision making.

Together, these analyses will be used to distinguish among distinct cognitive mechanisms that could hypothetically underlie perceptual differences related to ASD and to understand how those mechanisms vary with other sensory and behavioral symptoms across the ASD population. This research program will provide a computational model-based framework for evaluating perceptual integration across a range of people with ASD and a quantitative approach for comparing perceptual differences in rodent models of ASD in future studies.