Prototypical Category Learning Intact In Adolescents and Adults with High-Functioning Autism

Background: Categorization is something we all do with little effort. Some researchers have claimed that this basic ability is impaired in individuals with autism spectrum disorder (ASD). According to a recent neural theory of atypical cognitive development in autism (Grossberg and Seidman, 2006), “hypervigilant” category learning in some persons with autism may cause an impaired ability to learn prototypes for abstract general categories, with consequences for deficits in attention and related behavioral symptoms of autism (e.g. difficulty generalizing learned concepts, over focus on fine detail). Several studies have observed atypical prototype learning in autism; others failed to find any impairments.

Objectives: The present study utilized a classic prototype formation task to examine if individuals with autism would (1) struggle to learn categories, and (2) have difficulty generalizing from learned categories to new members of those categories. 

Methods: Twenty high-functioning individuals with autism and matched controls participated in a classic prototype learning paradigm. Two prototype dot patterns were generated; a set of dot patterns with varying similarity (high, medium, or low) to these prototypes was also created. During training, participants categorized medium-similarity patterns into two categories until criterion was reached. At test, participants were shown training items and a novel set of dot patterns with varying similarity to the unseen prototypes (high-, medium-, and low-similarity to unseen prototypes).     

Results: Individuals with autism learned more slowly and less well; the ASD group (1) required significantly more blocks to reach criterion, and (2) had significantly lower accuracy in both training and test. However, there were no group differences in pattern of performance during the test phase; individuals with autism showed a typical pattern for normal prototypical category learning (greatest accuracy for high-similarity dot patterns, equal accuracy for training and test medium-similarity dot patterns, lowest accuracy for low-similarity dot patterns).

Conclusions: These data indicate that high-functioning individuals with autism do not manifest gross deficits in prototypical category learning. A possible theoretical interpretation of these data is given in terms of underlying brain mechanisms.