The "risk matrix": Predicting financial risk taking with fMRI Brian Knutson Stanford University Consistent with "greed" and "fear" moving markets, growing FMRI evidence suggests that brain activity associated with positive arousal promotes risk taking, but brain activity associated with negative arousal promotes risk avoidance. New theory-free classifier solutions verify that a localized, robust, and generalizable set of neural features can predict financial risk taking. These findings have implications for predicting nontraditional (e.g., skewed) as well as traditional (e.g., variable) types of risky choice.