Biological motion perception is differentially predicted by Autistic trait domains

Abstract

We tested the relationship between biological motion perception and the Autism-Spectrum Quotient. In three experiments, we indexed observers' performance on a classic left-right discrimination task in which participants were asked to report the facing direction of walkers containing solely structural or kinematics information, a motion discrimination task in which participants were asked to indicate the apparent motion of a (non-biological) random-dot stimulus, and a novel naturalness discrimination task. In the naturalness discrimination task, we systematically manipulated the degree of natural acceleration contained in the stimulus by parametrically morphing between a fully veridical stimulus and one where acceleration was removed. Participants were asked to discriminate the more natural stimulus (i.e., acceleration-containing stimulus) from the constant velocity stimulus. Although we found no reliable associations between overall AQ scores nor subdomain scores with performance on the direction-related tasks, we found a robust association between performance on the biological motion naturalness task and attention switching domain scores. Our findings suggest that understanding the relationship between the Autism Spectrum and perception is a far more intricate problem than previously suggested. While it has been shown that the AQ can be used as a proxy to tap into perceptual endophenotypes in Autism, the eventual diagnostic value of the perceptual task depends on the task's consideration of biological content and demands.

Figure 1

(a) Mean duration thresholds for discriminating between left and right directions of the structural biological motion, kinematics biological motion, and random-dot motion stimulus in Experiment 1. (b) Mean naturalness discrimination thresholds (% naturalness) for Experiment 2. Error bars represent +/1 standard error of the mean.

Figure 2

(a) Overlaid individual-subject scatterplots and regression fits between the AQ domain imagination and thresholds for the three (structural biological motion, kinematics biological motion, random-dot motion) stimuli in Experiment 1. (b) Individual-subject scatterplot and regression fit between the AQ domain attention switching and thresholds on the biological motion naturalness discrimination task in Experiment 2.

Figure 3

(a) Mean duration thresholds for discriminating between left and right directions for the two biological motion stimuli (structural and kinematics) and the random-dot stimulus in Experiment 3. (b) Mean naturalness discrimination thresholds (% naturalness) for the same group of participants. Error bars represent +/1 standard error of the mean. (c–e) Individual-subject scatterplots and regression fits between thresholds for each of the structural biological motion, kinematics biological motion, random-dot motion stimuli and the AQ imagination scores. (f) Individual subject scatterplot and regression fit between the AQ domain attention switching and thresholds in the biological motion naturalness discrimination task.