Background: There are two accounts of categorization performance in autism: that there is an impairment in prototype formation (Klinger & Dawson, 2001) and that there is an impairment in processing features held in common between stimuli (Plaisted, O'Riordan, & Baron-Cohen, 1998). These accounts, together with central coherence theory (Frith, 1989; Frith & Happé, 1994), imply a reduced or absent prototype effect in autism.
Method: Children with autism or Asperger syndrome (n = 15) matched on age, gender, and verbal mental age with typically developing children (n = 15) completed a picture recognition task (Experiment 1). These participants also studied categories of cartoon animals possessing either an average prototype structure (Experiment 2) based on (Younger's 1985) stimuli or a modal structure (Experiment 3) based on (Hayes and Taplin's 1993b) stimuli. Following the study phases, participants completed recognition tests comprising prototypes and other exemplars with varying degrees of similarity to the prototypes.
Results: For both participant groups, recognition memory appeared intact (Experiment 1) and a full prototype effect in recognition memory was observed in both Experiment 2 and Experiment 3.
Conclusions: The present studies fail to support predictions of impaired prototype effects in autism. The discussion focuses on key methodological differences between these studies and those that support claims that central coherence, prototype formation, and common feature processing are impaired in autism.