Anxiety refers to an emotional state that is typically experienced when an important goal, such as survival or social approval, is under threat. Whereas in some cases it serves a purpose as a negative reinforcer that encourages action to remove the threat, many individuals suffer from disproportional levels of anxiety in their daily lives. Despite decades of scientific efforts, surprisingly little progress has been made towards understanding the behavioral and cognitive mechanisms of anxiety. Two explanations are often given in the current literature. First, the laboratory tasks used in experimental psychopathology may be too simplistic and abstract to probe the mechanisms that underlie anxiety in real life. Second, patient status often hinges on the diagnostic criteria of the Diagnostic and Statistical Manual of Mental Disorders (DSM), which has been much criticized for assuming categorical differences between patients and non-patients, as well as comorbidity being the rule rather than the exception. In short, there are conceptual problems on both sides of the equation when relating clinical anxiety to behavior and cognition in an experimental setting. This study aims to address both these shortcomings. Patients and non-patients will be recruited to participate in complex, naturalistic tasks in virtual reality where they need to take risks to survive under various levels of threat certainty. We will use the ‘Hierarchical Taxonomy of Psychopathology’ (HiTOP) framework, which has recently been developed to overcome the issues of the DSM framework, to characterize participants along a complete set of orthogonal symptom dimensions and higher-order constructs.