The proposed research is designed to extend our understanding of the neuropathology underlying one of the core behavioral characteristics of autism, deficits in emotional understanding. Autism is a severe and pervasive developmental disorder characterized by marked impairments in social relatedness and understanding, an integral component of which is emotional understanding and responsiveness. Although this has been well characterized behaviorally, previous empirical research on neurobiological substrate of these deficits has been limited by the available methodologies. Recent advances in functional imaging technology, however, allow for a more direct examination of the neural systems implicated in processing emotion, systems which are presumably damaged in individuals with autism. fMRI will be used to measure in vivo, neural activity while participants watch film clips judged to depict examples of positive and negative emotions. These measurements will be obtained from 30 adolescents and adults with high-functioning autism and 30 matched, non-autistic controls. It is hypothesized that (1) in the non-autistic sample, emotional stimuli will elicit distinct patterns of neural activation, involving the limbic association cortex, and (2) positive, approach-related emotions will elicit activation distinct from negative, withdrawal-related emotions, involving hemispheric asymmetries. It is further hypothesized that (3) autistic participants will exhibit a pattern of neural activation distinct from non-autistics when presented with emotional stimuli, and (4) that autistic subjects will have particular difficulty with more complex, cognitively mediated emotions, such as embarrassment, than non-autistic subjects. This research is designed to add to our understanding of the underlying pathophysiology of autism, and to provide a possible extension to neurological models of the understanding and processing of emotion in normal development.