Objectives: Here, we tested the hypothesis that the relationship between illusion susceptibility and systemizing is limited to illusions that are caused by global distortions of the observer's reference frame, using an illusion (the Rod-and-Frame Illusion, or RFI) that is known to have two variants. When shown a large tilted frame, the observer's global perception of vertical is distorted, causing a misperception of the orientation of an enclosed rod. In contrast, a smaller tilted frame has no effect on perceived vertical, but causes a misperception of rod orientation via a local contrast effect between the rod and the nearby contours of the frame.
Methods: Typically developing college students with normal vision (n=54) performed three tasks involving the RFI in counter-balanced order. The Perception task was a measure of the overall effect of the RFI. Participants were asked to make a judgment (“clockwise” or “counterclockwise”) about the perceived tilt of a rod presented within a tilted frame. The Saccade-to-Vertical task was a measure of the global distortion of the observer's reference frame. Participants were asked to make a vertical saccade, within the context of a tilted frame, to the top of an outer circle. The Saccade-to-Target task was a measure of the local contrast effect. Participants were asked to make a saccade, within the context of a tilted frame, to the point on the outer circle where they believed that an imaginary extension of the rod would intersect with the circle. Participants also completed the questionnaires of the Systemizing Quotient, the Autism Quotient, and the Empathizing Quotient.
Results: As hypothesized, the global distortions of perceived vertical associated with large frames were negatively correlated with systemizing, as well as the attention-to-detail subscale of the AQ. Surprisingly, systemizing was also found to be correlated with the local contrast effects associated with small frames, but here the correlation was a positive one: higher levels of systemizing were associated with an increased susceptibility to the local contrast effects.
Conclusions: In sum, these findings indicate that while autism is not simply associated with a decreased reliance on contextual cues, it is associated with a more complex shift from a general reliance on global contextual cues to an exaggerated reliance on local contextual cues. The complexity of these relationships in terms of illusion susceptibility may have confounded earlier attempts to understand the perceptual effects of autism.