An Eye-Tracking Study: The Effect of Task on Visual Attention to Faces in Autism

Background: Individuals with autism demonstrate impaired ability to interact in the social world, and researchers have hypothesized that deficits in face perception may contribute to general difficulties in social interaction.  Indeed, it has been found that individuals with autism distribute their visual attention to faces in an atypical manner (Klin et al, 2002; Pelphrey et al, 2002; Jones et al, 2008) and that atypical face processing predicts level of social impairment (Klin et al, 2002; Jones et al, 2008).   Many studies have shown that compared to typically developing individuals those with autism display increased looking to the mouth and decreased looking to the eye region of the face.  Questions remain as to whether the face viewing behavior of typical controls and individuals with autism varies by task.          

Objectives: The current study investigates the effect of task demands on the face viewing strategy utilized by typically developing individuals and those with high functioning autism.

Methods: Participants were typically developing children and adults and those with high functioning autism.  Each participant completed two tasks, an emotion recognition (ER) task and a gender identification (GI) task.  In the ER task, participants viewed 80 stimuli of subtle emotional expressions.  Participants selected their answer from a forced choice list that followed each stimulus.  Stimuli were created from Ekman photographs (Young et al., 2002); they exhibited angry, afraid, sad, and disgusted expressions.  In the GI task, participants judged the gender of 72 randomly presented faces by pressing keys labeled man and woman.  Eye-tracking data were collected to explore fixation length in two areas of interest: the eye region and the mouth region. 

Results: Regardless of task, typically developing individuals spent more time looking at the eye region of the face than participants diagnosed with autism.  Across tasks, individuals with autism spent more time than typical controls fixated on the mouth region than the eye region of the face.  That being said, a significant interaction of task (ER task versus GI task) and diagnostic group (autism versus typical control) was found (F=5.409; p=0.023).  During the GI task both groups spent a greater amount of looking time to the eye region of the face than they did during the ER task, however individuals with autism spent less time looking to the eyes than typical controls.  During the ER task both groups spent more time viewing the mouth region of the face than they did during the GI task, but this effect was much more dramatic for the participants with autism. 

Conclusions: While individuals with high functioning autism look more at the mouth and less at the eye region across tasks, it appears that their face viewing strategy does shift when asked to make judgments about different facial information.  Like typically developing individuals, those with autism look more at the eye region when completing a GI task and more at the mouth region when completing an ER task.  The difference lies in the degree to which individuals with autism attend to these regions.