Purchase this article with an account.
Arjen Alink, Ian Charest; Preferential use of local visual information in individuals with many autistic traits. Journal of Vision 2018;18(10):406. doi: https://doi.org/10.1167/18.10.406.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Individuals with an autism spectrum disorder (ASD) diagnosis are often described as having an 'eye for detail'. This observation, and the finding that individuals with ASD tend to 'see the trees before the forest' when performing the Navon task, has led to the proposal that ASD is characterized by a bias towards processing local image details. However, it is unclear if this bias is restricted to abstract stimuli - like the Navon stimuli - or whether day-to-day vision is dominated by such object details in ASD. If the latter is true, then we expect individuals with a high number of autistic traits to rely more on image details for object recognition. To assess this, we developed a new psychophysical method for determining the relative contribution of low-level image features to recognition. In short, we asked participants to perform a cat vs. dog recognition task based on images containing a randomly selected subset of the original image's features (Gabor wavelets) and used reverse correlation to measure the importance of each feature for recognition. We used this method to demonstrate that object recognition depends more on high-spatial frequency visual features in individuals with an above-median number of autistic traits. Intriguingly, this enhanced reliance on high-spatial frequency information was best predicted by autistic traits related to impaired social skills and least predicted by traits related to attention to detail. The findings of this study suggest that the bias towards processing image details ubiquitously affects vision in individuals with a high number of autistic traits, which raises the possibility that such a bias underlies a wide range of real-life abilities and difficulties associated with ASD.
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only