Abstract
Understanding the mechanisms underlying psychopathology is a central goal in clinical science. While existing theories have demonstrated high clinical utility, they have provided limited quantitative and experimentally falsifiable explanations of psychopathological mechanisms. Previous computational modelling studies have primarily focused on isolated factors, posing challenges for advancing clinical theories. To address this gap and leverage the strengths of clinical theories and computational modelling in a synergetic manner, it is crucial to develop a quantitative model that integrates major vulnerability factors within a comprehensive psychopathological model. In this study, using social anxiety disorder (SAD) as an example, we present a novel approach by combining active inference modelling, an innovative computational approach that elucidates human cognition and action, with cognitive-behavioural theory (CBT), a well-established clinical framework. This CBT-informed quantitative account integrates multiple vulnerability factors to elucidate the psychopathology of SAD. Through a series of simulations, we systematically delineate their effects on pathological behaviours. This resultant model inherits the conceptual comprehensiveness of CBT and the quantitative rigor of active inference modelling, revealing previously elusive pathogenetic pathways and enabling the formulation of specific predictions for empirical studies. Furthermore, this resultant model provides a strong theoretical foundation for precision medicine in SAD by allowing for individual differences in the symptom development trajectory. Overall, this research presents the first quantitative model of SAD that unifies major pathogenesis factors proposed by CBT. It highlights the feasibility and potential of integrating clinical theory and computational modelling to advance our understanding of psychopathology.