Abstract
Prediction is a fundamental brain function. We use past information to predict events ranging from positions of moving objects to social interactions. Recent work postulates impaired prediction as a global trait in autism spectrum disorder (ASD; Sinha et al., 2014). However, empirical evidence for this hypothesis is limited. Here, we test the integrity of motion prediction abilities in ASD. To better understand potential mechanisms of impairment, we comprehensively examined the factors known to be associated with motion prediction performance: the central-tendency bias and the relationship between smooth pursuit eye-movements and prediction errors. METHODS: Participants were 26 children and adolescents (ages 9-17) with ASD and 20 typically developing (TD) controls. The task was to predict the time-to-arrival of a moving object that disappeared behind an occluder. Across trials, we sampled the speed (constant; 10-20°/s), visible (8-18°) and occluded (0.5-20°) distances from uniform distributions, which together resulted in a range of visible (0.4-1.8s) and occluded (0.025-2s) durations. Prediction performance was evaluated using the bias and variability in responses. Smooth pursuit quality was characterized by pursuit gain (eye-/object-velocity). RESULTS: Individuals with ASD exhibited early prediction bias, where they responded, on average, 30ms earlier than the actual hitting time. This was most pronounced at longer occluded durations, and was not explained by the central-tendency bias observed in TD. They also showed atypically increased pursuit gain compared to controls during stimulus occlusion. Notably, pursuit quality was differentially linked to prediction performance in the two groups. Better smooth pursuit was associated with reduced prediction variability in TD and counterintuitively related to earlier prediction bias in ASD. CONCLUSIONS: Results demonstrate atypicalities in how individuals with ASD predict motion. The findings provide empirical support for the proposal of impaired prediction in ASD, and further suggest differential use of extra-retinal signals for motion prediction in ASD.
Meeting abstract presented at VSS 2018