This study focuses on one of the most basic and critical aspects of information processing-categorization. The term categorization is used in a broad sense to mean not only the grouping of similar objects, but also a process basic to core perceptual recognition abilities. Categorization is a process that begins within the first few months of life. Infants quickly begin to categorize their world, significantly reducing demands on memory. Once the infant can categorize and have a central representation of similar objects such as dogs, it is not necessary to remember every encountered instance unless the instance is unusual. The ability to categorize is also critical to language development, and most theories assume it is a pre-requisite for learning words. Categorizing or prototyping is also central to face recognition abilities and problem solving. This project will extend current research on high functioning children and adults with autism (HFA) on processing categorical information about objects and facial information including gender, emotional expression, recognition and attractiveness. Studies will take a more in depth look at the mechanisms that underlie development of categorical expertise. Extensions of this research will look into understanding individual differences in performance of HFA participants, and will provide important behavioral indices of autism that will be used by Projects II, III and IV as they attempt to characterizethe neuroanatomy of autism as reflected in brain structure and connectivity. Importantly, this research will be extended to two populations critical to understanding autism spectrum disorders (ASD) including its origin, early diagnosis, and the development of early intervention strategies. First, this project will longitudinally study a group of 6- to 16- month old infant siblings of older children diagnosed with autism. Second, the project will study a group of newly diagnosed ASD toddlers that are low functioning. Both of these populations will be tested with nonverbal methodologies originally designed for studying typically developing infants and toddlers. These methods will allow examination of critical early information processing abilities such as attention, memory, and categorization. These methods represent well-designed experimental procedures that could easily be translated to clinical practice for early diagnosis if results support their use. These methods allow study of underlying information processes currently not addressed by more standard observation studies of at-risk infants. These basic information processing mechanisms are more effective targets for changing outcome than behavior.