Assessing Hypersensitivity to Difference in Autistic and Non-Autistic Adults

Abstract

Current literature on autism has tended to examine individual traits and characteristics, or the link between two or three traits. This thesis will outline previous literature on the cognitive and perceptual theories of autism, as well as research on anxiety, sensory processing, intolerance of uncertainty and the need for predictability, and schema development in this group; a model is presented that proposes hypersensitivity to difference as a central unifying trait from which these autistic traits flow. A change-blindness task (Chapter 2) did not prove to be a useful measure of hypersensitivity to difference. It appeared to provide evidence that autistic participants were not affected by context; however, this was not conclusive, given that non-autistic participants performed in a similar way. Ratings of similarity and difference (Chapter 3) by autistic and non-autistic participants revealed a subgroup of autistic participants who made very low similarity ratings for pairs of items that only shared a thematic relationship, which highlighted a reduced use of context in this group. Protocols showed that autistic participants made more reference to difference compared to non-autistic participants, although this was not reflected in the difference ratings. A card-sorting task (Chapter 4) showed that autistic participants performed similarly to non-autistic participants in creating more taxonomic categories over thematic ones. The performance of autistic participants demonstrated that they were able to form categories based on featural similarities. A second sorting task using novel stimuli showed that both sets of participants made a comparable number of categories, but autistic participants generated more unique labels, possibly suggesting narrow, more specific categories. A small-scale pilot study (Chapter 5), using virtual reality technology to increase the ecological validity of a change detection task, raised interesting questions about the potential usefulness of this with autistic participants. Overall, there was sufficient evidence to suggest that the model is worth further exploration.