Neural Markers of Treatment Mechanisms and Prediction of Treatment Outcomes in Social Anxiety

PROJECT SUMMARY/ABSTRACT Social anxiety disorder (SAD) is one of the most common mental disorders. For unknown reasons, many patients do not respond to existing treatments. Treatment guidelines and systematic reviews often recommend CBT as the first line treatment, and then to start an SSRI adjunctively for patients who show no or only partial response to initial CBT. A major advance and step toward personalized medicine would be to identify reliable treatment predictors and to clarify the neuromechanism of treatment change. One promising approach toward improving patient outcomes is to examine the key neurocircuitry of SAD that may also serve as neuromarkers to predict treatment response. We have gathered convincing pilot data pointing to such neuromarkers to predict response to CBT for SAD. The next translational step and our primary aim is to apply state of the art computational psychiatry approaches to further establish the evidence of these neuromarkers, in line with moving psychiatry toward precision medicine. This aim will be efficiently achieved by collecting multimodal data to better elucidate key neurocircuitry in SAD compared to controls with state-of-the art neuroimaging in a well powered sample, as well as differential treatment related changes in neural circuitry (target engagement). The ultimate goal is to effectively treat all patients, not only a few and without knowing why, and to illuminate the brain circuitry associated with effective treatments in order to inform psychopathology, nosology, and therapy of common mental disorders. For these reasons, we propose to recruit a large number of patients with SAD (n = 190) and healthy controls (n = 50) to examine differences in relevant neurocircuitries that will also be used as neuromarkers of treatment response. Patients with SAD will first receive CBT group therapy. Those who show no or only partial response will then receive individual and tailored CBT plus SSRI. In addition to MRI measures, we will examine EEG and behavioral measures to determine whether there may be less expensive correlates of neuropredictors that can be easily implemented in clinical practice. We have assembled a team of skilled researchers with complementary expertise at the Massachusetts Institute of Technology (MIT; John D. E. Gabrieli, Ph.D.), Boston University (BU; Stefan G. Hofmann, Ph.D.), and McLean Hospital (Daniel Dillon, Ph.D.), as well as outstanding consultants in neuroimaging analysis (Northeastern University: Susan Whitfield- Gabrieli, Ph.D.) and machine learning applications in psychiatry (McLean Hospital: Christian Webb, Ph.D.).