A Computational Model of Emotion and Personality: Applications to Psychotherapy Research and Practice

VR applications in psychotherapy are gaining prominence in the research community and acceptance among practitioners, with particular successes in the area of phobia treatments. However, a promising technology remains unexplored: computational models of cognition and emotion. Cognitive models (also termed cognitiveor agent-architectures) aim to emulate cognitive processing such as attention, perception, and decision-making, and are used by cognitive scientists to advance understanding of the mechanisms and structures mediating cognition. These models are also used in applied settings to improve the realism of training and assessment environments, and to improve human-computer interaction and system design. Recently, architectures have been developed that explicitly represent emotions: both emotion appraisal processes, and effects of emotions on cognition. Such computational models of cognition and emotion have two types of applications in cybertherapy. First, they can enhance the realism of a synthetic agents used in VR assessment and treatment environments (e.g., avatars used in social phobia treatment). Second, they have the potential to advance our understanding of the etiology and treatment of a variety of affective disorders, by enabling the modeling of the mechanisms of cognitiveaffective interactions that play central role in these disorders, and in their treatments. In this paper I first describe a cognitive-affective architecture capable of modeling the dynamic generation of emotions (affect appraisal), and selected effects of emotion on cognition. I then describe possible applications of this architecture to psychotherapy practice and research.