This project will develop a novel computational model and test to measure the mechanism of exposure therapy, which is used to treat threat-based psychopathology. Threat-based psychopathologies, such as anxiety disorders, are the most common form of mental illness and the sixth-leading cause of disease-related disability worldwide. Exposure therapy is the most widely used psychological treatment for threat-based disorders. During exposure therapy, a person's invalid, pathologic threat cue-associations are identified, and the person is trained against a pathologic threat response by exposure to the cue. Exposure therapy has strong theoretical underpinnings in learning. However, its mechanisms have been conceptually formulated but not formally defined and tested. This project takes the first step in addressing the gap between a theory of exposure therapy mechanism and ability to measure it, by developing a formal expression of theoretical mechanisms in a computational model. This formal specification and validation with human subjects may improve translation between animal and human studies, provide targets for neurophysiological investigations, and contribute to increasing the efficacy and efficiency of exposure-based treatments. To achieve this goal, we will extend CompAct (Competitive Activation), a computational model describing the interactive dynamics of associative learning and attention, or attention learning. The principles of attention learning have been well established in both humans and other species. CompAct has demonstrated excellent predictive performance in tasks designed to probe attention learning. To measure individual differences in threat processing, we will develop a novel, clinically relevant attention-learning task. By measuring affective cue salience, affective cue associations, and their changes during learning, the model can explain both extinction (reduction of acquired threat associations) and habituation (reduction of salience of threat cues). Extinction and habituation are two primary mechanisms that have received empirical support in exposure therapy. In addition to development, we aim to validate the novel task and CompAct. In two large normative samples (N's=1000), we will parametrically characterize individual differences in the dynamics of affective cue salience and associations. We will then associate these differences to individual levels of anxiety symptoms, which are continuous with anxiety disorders, a threat-based psychopathology for which exposure therapy is indicated.
Exposure therapy is the most widely used method to treat threat-based disorders such as anxiety. Despite its widespread use, its mechanism is unknown. This research lays the foundation for formally defining promising mechanisms by which exposure therapy may work by using a computational model. This work will lay the foundation of future human- and animal-based work to improve exposure therapy by targeting its mechanisms.
