Virtual environments for social skills training: comments from two adolescents with autistic spectrum disorder

People with autism experience profound and pervasive difficulties in the social domain. Attempts to teach social behaviours tend to adopt either a behavioural or a 'theory of mind' (ToM) approach. The beneficial aspects and limitations of both paradigms are summarized before an examination of how virtual reality technology may offer a way to combine the strengths from both approaches. This is not an exhaustive review of the literature; rather, the papers are chosen as representative of the current understanding within each broad topic. Web of Science ISI, EMBASE and PsycInfo were searched for relevant articles. Behavioural and ToM approaches to social skills training achieve some success in improving specific skills or understanding. However, the failure to generalize learned behaviours to novel environments, and the unwieldy nature of some behavioural methodologies, means that there is a need for a training package that is easy to administer and successful in promoting learning across contexts. Virtual reality technology may be an ideal tool for allowing participants to practise behaviours in role-play situations, whilst also providing a safe environment for rule learning and repetition of tasks. Role-play within virtual environments could promote the mental simulation of social events, potentially allowing a greater insight into minds. Practice of behaviours, both within and across contexts, could also encourage a more flexible approach to social problem solving. Virtual environments offer a new and exciting perspective on social skills training for people with autistic spectrum disorders.

The potential of virtual environments for teaching people with autism has been positively promoted in recent years. The present study aimed to systematically investigate this potential with 12 participants with autistic spectrum disorders (ASDs), each individually matched with comparison participants according to either verbal IQ or performance IQ, as well as gender and chronological age. Participants practised using a desktop 'training' virtual environment, before completing a number of tasks in a virtual café. We examined time spent completing tasks, errors made, basic understanding of the representational quality of virtual environments and the social appropriateness of performance. The use of the environments by the participants with ASDs was on a par with their PIQ-matched counterparts, and the majority of the group seemed to have a basic understanding of the virtual environment as a representation of reality. However, some participants in the ASD group were significantly more likely to be judged as bumping into, or walking between, other people in the virtual scene, compared to their paired matches. This tendency could not be explained by executive dysfunction or a general motor difficulty. This might be a sign that understanding personal space is impaired in autism. Virtual environments might offer a useful tool for social skills training, and this would be a valuable topic for future research.