The quantity and quality of research into autism and related conditions have increased dramatically in recent years. Consequently we selectively review key accomplishments and highlight directions for future research. More consistent approaches to diagnosis and more rigorous assessment methods have significantly advanced research, although the boundaries of the 'broader phenotype' remain to be defined and the validity of Asperger's disorder as a discrete syndrome remains controversial. Recent epidemiological studies have shown that Autism Spectrum Disorders are common, but there continues to be debate about the causes of the increase in the frequency with which autism is diagnosed. Psychological research has helped to develop new developmental models for the disorder and there have also been significant advances in the molecular genetics of autism and understanding of the underlying neurobiological processes. Areas important for future research include the study of autism as it first develops, i.e., in infants and very young children, and of specific processes (psychological and neurobiological) which underlie the disorder. Significant challenges lie ahead in evaluating the growing number of treatments for autism and in integrating the results of research into treatment and educational settings.