Abstract
Background. Autism-spectrum characteristics are often obscured by rehearsed cognitive compensation strategies. We aim to devise tasks that avoid triggering compensatory behaviors, so that latent differences can be revealed. For example, the tendency for high-AQ subjects to avoid looking at eyes could be more reliably measured when subjects were occupied by an "orthogonal" task to avoid looking at the mouth (Shimojo et al., VSS 2012). Here, we test another method: use a task that challenges the pro-social impulses in low-AQ subjects, giving high-AQ subjects the advantage. We ask subjects to avoid looking at face stimuli. Method. Unscreened student participants (N=11; AQ 11-46; one diagnosed AS) viewed a series of paired face and flower images (FaceGen / photo) while their gaze was monitored (Eyelink2). The image pairs were arranged diagonally upper-right to lower-left, to minimize generic spatial biases in viewing. Subjects inspected the images under three task conditions: 1) Don’t look at the flower, 2) Don’t look at the face, 3) Free view. Later, they filled out the Baron-Cohen Autism Quotient (AQ) questionnaire. Each condition was assigned to a separate block of 36 trials, randomly ordered within subject. Results. Percentage of initial gaze shifts in the direction of the face was inversely correlated with the AQ scores in the "Don’t look face" and the "Free view" (p<.001;p<.05), but not in the "Don't look flower" conditions. The "Don’t look face" revealed a clearer and steeper slope of correlation. ANOVA revealed a main effect of task (P<.001), and an interaction between task and AQ category (high/low) (P<.05). Discussion. People with low AQ scores showed difficulty suppressing a tendency to look at eyes and faces. As AQ scores increase, this difficulty declines. This reversed diagnostic strategy may aid the design of simple measures, to complement the existing methods based on complex social tasks.
Meeting abstract presented at VSS 2013