Linking Aetiology With Social Communication In A Virtual Stroke Patient

This paper describes an approach to building a virtual stroke patient which allows learners to visually explore connections between different stroke aetiologies and social behaviour. It presents an architecture that links a parametric model of aetiology to verbal and non verbal behaviour which can be manipulated in realtime. We believe that this design has the potential to consolidate understanding by allowing learners to systematically explore variations in clinical presentation. To the best of our knowledge, this is the first Intelligent Virtual Agent (IVA) that uses a parameterised behaviour model to provide an interactive examination and diagnosis of a stroke patient.