Do Individuals Diagnosed with Autism Spectrum Disorders Have an Advantage in Real-World Visual Search Tasks?

Background: People diagnosed with Autism Spectrum Disorders (ASD) often perform better than typically developing (TD) individuals in lab-based visual search tasks. This ability is often attributed to superior bottom-up processing (an ability to discriminate between a target and distractors, related to weak central coherence). However, research has also suggested that individuals with ASD show difficulties with top-down processing (integrating contextually relevant information to aid search), which may cause problems in real-world situations. Little is known about whether these differences in the performance of individuals with ASD are also represented in more ecologically valid visual searches and how this might influence their everyday functioning. 

Objectives: We report findings from a real-world scene search study to compare ASD, TD, and Anxious control (ANX) participants. We used eye tracking data to the record the task performance and search strategy for contextual and non-contextual targets. 

Methods: Twenty-five individuals with ASD, 27 TD individuals, and 28 ANX individuals completed two simple visual search tasks. For both tasks, participants had to find a named object within a real-world scene. In one task, objects were located in contextually-relevant locations; in the other task, the objects were superimposed upon the picture in a non-relevant position. Each participant completed 41 trials in each condition, each with a unique scene. Eye movements were recorded using an SR Research EyeLink 1000 eyetracker. 

Results: All groups performed better in the contextual target condition than the non-contextual condition. In the non-contextual search, the ASD group were less accurate than the other two groups. However, preliminary analyses indicate that this difference was not significant. For the contextual search, the ASD group were also less accurate. Group means were significantly different with the ASD group being lower than the other two groups. Reaction time (RT) data suggest that the ASD group were also significantly slower than the TD and ANX groups in both conditions.  

Conclusions: Our preliminary results suggest that a previously proposed advantage for ASD in locating targets without contextual information may be lost when these targets are contained within real-world scenes. Our ASD group appears to benefit less than both typical and anxious comparison groups, when targets are contextually relevant to these scenes. Previous research has also suggested that ASD groups may perform faster in some visual search tasks. However, these RT data suggest that, even though their accuracy was comparable to TD and ANX individuals when finding the non-contextual targets, they took longer to do so. This suggests that the presence of complex real-world information, whether related to the target they are searching for or not, may interfere with simple visual search and increase difficulty in navigating daily physical and social interactions.