While knowledge on the neurobiological signatures of fear and anxiety disorders and, in particular, their association with treatment outcome is accumulating, clinical translation still awaits empirical proof of evidence. Exposure-based cognitive-behavioural therapy (CBT) is a first-line treatment, but clinically significant change is only seen in approx. 50-65% of patients. Patient stratification is a powerful option to increase treatment response; however, developing prognostic markers suitable for single-patient predictions still is in its infancy and crucially requires external cross-validation embedded within an a priori prediction approach - a procedure yet largely missing in the field of biomarker research. Employing a bicentric strategy, we will test the hypothesis that a priori prediction of treatment outcome based on neurobiological measures is possible in a second, independent sample. We will build upon findings from previous mechanistic studies of the CRC and incorporate them into the development of a predictive pattern comprising fear-relevant genotypes and molecules targeting neuropeptides (NPSR1, MAOA, SCL6A4/5-HT, CRHR1, BDNF Val66Met, OXTR), related epigenetic signatures as well as neurofunctional activation patterns associated with fear circuitry functions, and clinical data. Pre-treatment neurobiological signatures will be tested for their potential as a predictive response marker towards behavioural exposure in a model disorder of fear circuitry dysfunctions (spider phobia). Multivariate pattern analyses employing a machine learning framework will be used to generate predictions on the individual patient level and to cross-validate markers in a second, independent sample. We expect this project to further bridge the translational gap between basic and clinical research and to bring stratified medicine approaches into reach as one of the long-term goals of this CRC.