Self-Referential Gaze Judgements Are Impaired in Autism

Background: Impaired social cue processing is characteristic of autism. Gaze direction is a vital social cue that can serve several purposes; direct gaze may signal communicative intent, whilst averted gaze directs an observer’s attention to an object of mutual interest. Direct and averted gaze may be processed differently and one hypothesis is that impairment in gaze processing in autism is limited to direct gaze judgements. In typical development judgements of gaze direction are also influenced by the emotional expression of the face, such that happy faces are more likely to be judged as looking at the observer. This effect has not previously been tested in children with autism.

Objectives: This study disambiguated accuracy for geometric gaze discrimination from self-referential gaze discrimination in children with autism and assessed the influence of emotion expression on these judgements.

Methods: Participants were 22 children with autism and 21 age and IQ matched control children. Experiment 1 presented emotional face photographs at 2°, 4° and 6° deviations from direct gaze and required participants to judge if the face was averted to the left or right (geometric gaze judgement). Experiment 2 presented emotional face photographs at 0° (direct gaze) and 4° and required that participants judge if the face was ‘looking at them’ (self-referential gaze judgement).

Results:

Experiment 1: The main effect of gaze angle was statistically significant (F= 52.22, p< .01, ηρ²=.56) indicating that as gaze angle increases, accuracy of gaze judgments also increase. A significant interaction between gaze angle and emotional expression was also present (F= 5.75, p< .01, ηρ²=.12). No other main effects or interactions were significant, critically there was no main effect of Group, nor did Group interact with any of the within participant measures (all p>.05).

Experiment 2: The main effect of Group was statistically significant (F= 4.46, p<.05, ηρ²= .10), and indicates that participants with autism made more errors than controls. The main effect of angle was also significant (F= 14.13, p<.01, ηρ²= .26), however, this was qualified by a small but statistically significant interaction between emotional expression and gaze angle (F=3.18, p<.05, ηρ²=.07). The interaction was investigated by testing the difference in accuracy between direct and averted gaze for each emotion. This was greater for happy emotional expressions than neutral emotional expressions (t=2.623, p<.05), and also shows a trend to be greater than fearful expressions (t=3.933, p= .06). There was no difference between neutral and fearful expressions (t=.06, p =.95).

Conclusions: Judgement of gaze direction is impaired in autism but this impairment is limited to self-referential gaze judgments and cannot be explained by a general weakness in gaze discrimination abilities. Furthermore, children with autism are equally as sensitive as controls to the affect of emotional expression on gaze direction judgements. We suggest that this impairment may be related to atypical development of the cognitive processes determining egocentrism in autism.