Autism constitutes one of the best validated child psychiatric disorders. Empirical research has succeeded in delineating the key clinical phenomena, in demonstrating strong genetic influences on the underlying liability, and in identifying basic cognitive deficits. A range of neurobiological abnormalities has also been found, although the replicability of specific findings has not been high. An understanding of the causal processes leading to autism, and accounting for the marked variability in its manifestations, requires an integration across these different levels of enquiry. Although this is not yet possible, a partial integration provides a useful strategy for identifying key research questions, the limitations of existing hypotheses, and future research directions that are likely to prove fruitful. The research findings for each research level are critically reviewed in order to consider how to move towards an integration across levels.

One reason for looking at a person's eyes may be to diagnose their goal, because a person's eye direction reliably specifies what they are likely to act upon next. We report an experiment that investigates whether or not young normal infants use eye contact for this function. We placed them in situations in which the adult's action toward them was either ambiguous or unambiguous in its goal. Results showed that the majority of normal infants and young children with mental handicap made instant eye contact immediately following the ambiguous action but rarely after the unambiguous action. Young children with autism, in contrast, made eye contact equally (little) in both conditions. These results are discussed in relation to the function of eye contact, to our understanding of infant cognition, and to the theory of mind hypothesis of autism.

This study aimed to establish whether or not young children and young people with autism can understand the mental state of intention. Participants were exposed to personal experience of unintended outcomes, to test if they could distinguish intended vs. unintended actions. Recognizing accidental outcomes was more difficult for normal 4-year-olds than 5-year-olds, and more difficult for young people with autism, compared with comparison groups. Such findings suggest that the theory of mind deficit observed in people with autism is not restricted to understanding epistemic states, but also extends to understanding intention. The results are also compatible with an action-monitoring deficit. Future research needs to test these two accounts against each other.

Details are given of a new advanced theory of mind task, developed to approximate the demands of real-life mentalizing in able individuals with autism. Excerpts of films showing characters in social situations were presented, with participants required to answer questions on characters' mental states and on control, nonsocial questions. When compared with control participants, adults with high-functioning autism and Asperger syndrome were most impaired in their ability to answer the questions requiring mind-reading ability. Although the present findings have implications for task modification, such naturalistic, dynamic stimuli are held to offer an important means of studying subtle difficulties in mentalistic understanding.