Successfully Directing Eye Gaze Does Not Improve Face Recognition Ability in Children with ASD

Thursday, May 12, 2016: 11:30 AM-1:30 PM
Hall A (Baltimore Convention Center)
S. J. Thompson, S. R. Leekam, G. Powell and C. Jones, School of Psychology, Cardiff University, Cardiff, United Kingdom
Background:  Individuals with ASD show unusual patterns of eye gaze to the face and perform poorly on face processing measures. However, whether prompting individuals with ASD to look at optimal face regions will improve accuracy is unknown.  In previous research we found that when told to look at the mouth, neurotypical adults showed difficulty inhibiting an unconscious preference for first looking to the eye region. In addition, prompting instructions improved accuracy. What remains unknown is: (i) whether individuals with ASD show a first look eye preference (ii) whether directing eye gaze can improve their accuracy.

Objectives:  (1) To establish whether typically developing (TD) and ASD children have a preferred first look location that is difficult to inhibit. (2) To establish whether directing eye gaze improves accuracy.

Methods:  Fourteen children with ASD (8–13 years) and 24 matched TD children participated.  Using a forced-choice face recognition paradigm, participants had to distinguish the previously seen target face from an alternative image, where the eyes or mouth were digitally altered. This was first conducted with no looking instructions (unprompted condition) and then with instructions to look to the changed region (prompted condition).  Eye movements and task performance were measured. A control task using houses was also included.

Results:  Both child groups showed the same eye tracking patterns found in neurotypical adults. With respect to objective 1, first looks were significantly more often made to the eyes than mouth in the unprompted condition by both groups. With respect to objective 2, when prompted to look to the mouth, the first look for both groups was as often to the mouth as to the eyes, whereas when directed to the eyes, first looks reflected task instruction.  Analysis of dwell time indicated no group differences, with both groups spending more time looking at the prompted region. However, the groups differed on task performance.  Unlike the TD group, the ASD group did not show better face recognition ability when prompted where to look than in the unprompted condition, and did not show better performance when the eyes were the changed region compared to the mouth.  Performance on the control task was comparable between groups.

Conclusions:  Children with ASD showed similar looking patterns to TD children; they had a bias for first looking to the eye region. Further, this bias was difficult to inhibit even when it was an optimal strategy for task performance.  Both groups overcame this initial bias and spent most time looking at the prompted region. The data indicate that it is possible for children with ASD to show typical patterns of looking to the face, albeit in a controlled laboratory environment. However, unlike the TD group, children with ASD could not benefit from the prompts and their behavioural performance did not improve.