Preference Choices and Gaze to Faces In High-Functioning Autism

Background: Preference and gaze interact in a positive feedback loop to produce a phenomenon known as the ‘gaze cascade’ effect (Shimojo, Simion, Shimojo, & Scheier, 2003).  When observers are shown stimulus pairs and instructed to choose which of the two they find more attractive, their gaze is equally likely to be on either picture. However, in the few seconds before a decision is made, a gaze bias occurs toward the stimulus that is eventually chosen. This gaze bias is especially robust in tasks that involve face preference decisions. Recent research suggests individuals with autism spectrum disorders (ASD) have deficits in evaluating and making social judgments about faces, including inattention to faces and direct gaze aversion. 

Objectives: The present study was set up to examine whether these known aberrations in visual face processing interfere with preference choice decision making in ASD, reflected in a deviant gaze cascade pattern.

Methods: 4 ASD subjects and 4 age and gender matched healthy controls (HC) performed a 2-alternative forced-choice task while their eye-gaze was tracked.  The planned subject group size is 10 each of HC and ASD.  The task was to select the stimulus they prefer by pressing a button under a free viewing condition. Stimulus types consisted of faces and natural scenes.

Results: First, we were able to replicate the findings of a gaze cascade in the HCs, already with this temporary group size.  Interestingly, the known gaze aversion for faces in ASD did not interfere with the gaze bias toward the to-be-chosen picture at decision time, independent of stimulus type. Reaction time analysis showed that there was a main effect of group. ASD subjects responded faster than HCs in the trials where a decision about facial attractiveness had to be made, but not in trials where natural scenes were presented.  Further analysis of the gaze patterns showed that autistic subjects had an increased gaze bias toward the selected picture compared to HCs, especially in trials involving decisions about facial attractiveness.

Conclusions: These findings implicate that while gaze is clearly involved in preference formation in ASD subjects, the psychological process that leads to the decision may differ from that of HCs.  The course of viewing behavior in ASD clearly deviated from that of the HCs and is not in agreement with the typical gaze cascade.  In light of the reduced reaction times in ASDs for facial stimuli and the pattern of their gaze behavior we suggest that the subjects have an abnormal preference decision process when confronted with facial stimuli.