In pursuit of preventive medicine, neuroscientists are mining nucleic acids, blood, saliva, and brain images to uncover biomarkers to estimate risk of brain disorders like psychosis and dementia. Though this science is young, the prospect of bioprediction is elevating controversy about how (or whether) to integrate its imperfect probabilistic information into clinics, courtrooms, or legislatures. Medicine, law, and policy, after all, have protocols for the presence of disorders, not the risk of them. This book raises the idea that biomarkers seem problematic because they reveal the extent to which moral and legal structures embrace a flawed assumption that the concept of disorder is categorical (sick versus well) and thereby elide important questions about the types and magnitudes of probabilities that matter. Progress in the neuroethics of biomarkers, therefore, requires the rejection of a binary concept of disorder in favor of a probabilistic one based on risk of harm, which the book conceives as a generalized bio-actuarial 'Probability Dysfunction'. This reconceptualization of disorder enables a clear framework through which to interrogate the revision of the definition of mental disorder in the DSM-5, disagreement in the nosology of risk of psychosis and dementia, conditions in which biomarkers might enhance moral responsibility, reduce legal responsibility, or challenge current structures for distributive justice. By arguing that biomarkers can often be morally useful through enabling moral agency, better assessment of legal responsibility, and fairer distribution, the chapter contests the principle that the acceptability of bioprediction turns primarily on whether it enables better medical treatment.