Eye-Tracking Study Investigating Performance Differences Between ASD and Non-ASD Groups in Simple and Complex Emotion Recognition Tasks

Background:  Individuals with Autism Spectrum Disorders (ASD) have a difficulty of emotion recognition. The performance is worse with complex emotions than simple emotions (Golan et al.,2006).

Objectives:  The-eye tracking study was designed to investigate the influence of task complexity on the performance differences between ASD and non-ASD groups in simple and complex emotion recognition tasks. 

Methods: The participants were 25 adults (12 with ASD and 13 without ASD) with normal IQ. The twenty-seven movie stimuli were constructed (three semantically neutral sentences in each of three vocal emotions with each of three facial emotions). Three emotions were happy, angry and sad. The nine stimuli were congruent and the eighteen stimuli were incongruent between the facial emotion and the vocal emotion. The participants were given two simple tasks and two complex tasks. The two simple tasks were the auditory task and the visual task. The auditory task was to recognize vocal emotion in each of the movie stimuli with audio-only presentation. The visual task was to recognize facial emotion in each of the movie stimuli with video-only presentation. The two complex tasks were the emotion labeling task and the congruent-incongruent task. The emotion labeling task was to label one of three emotions to each of the movie stimuli. The congruent-incongruent task was to judge whether the facial emotion and the vocal emotion in each of the movie stimuli were congruent or not. Three measurements were used: percentage of correct responses, mean reaction time and fixation behavior. In the emotion labeling task, percentage of correct responses was not recorded because the task has no correct answers except the congruent movie stimuli. In the auditory task, fixation behavior was not recorded because no video was presented in the task. The Tobii 1750 eye-tracker system was used to record eye movements and analyze the fixation behavior. 

Results: The percentage of correct responses in the congruent-incongruent task was significantly higher in the non-ASD group than in the ASD group. In the auditory and the visual tasks, the two groups were not significantly different. The mean reaction times in both the auditory and the visual tasks were longer in the ASD group than in the non-ASD group. In the emotion labeling and the congruent-incongruent tasks, the two groups were not significantly different. The fixation behaviors in both the emotion labeling and the congruent-incongruent tasks were significantly different between the two groups. In both tasks, the ASD group fixated significantly more in the mouth region and significantly less in the eye region than the non-ASD group. In the visual task, the two groups were not significantly different.

Conclusions: The findings suggested as follows: the ASD group (1) performed as well as the non-ASD group but needed more time to recognize emotion in the simple emotion recognition task with a single modality, (2) performed worse and changed their fixation behavior in the multi-modal complex emotion recognition task.