236. Developing fMRI-Based Biomarkers for Pain and Emotions

For the last two decades, functional Magnetic Resonance Imaging (fMRI) revolutionized how we study pain, emotions, and their psycho-patho-physiology. However, as evidence accumulates, many of the brain-function mappings from fMRI studies appear to be flawed due to their poor sensitivity, specificity, and reproducibility. A new emerging paradigm, which we termed predictive mapping, has a potential to resolve these issues, building better fMRI-based biomarkers for pain and emotions. This new approach is based on specific uses of machine learning combined with experimental designs optimized for prediction, yielding well-defined neuroimaging signatures of brain-outcome relationships that can be prospectively tested in new individuals, studies, and translational applications.