Several methodological barriers impede discovery of early illness pathways in schizophrenia, including small samples, elongated study periods, and failure to integrate procedures and data across prodromal and first episode projects. A compounding factor is the tendency for single-site studies to focus narrowly on schizophrenia risk factors, rather than exploring vulnerability mechanisms that may cut across DSM-IV boundaries. To address these concerns, we discuss the merits of an integrated multisite approach to research that promotes large-scale investigation into the earliest phases of serious mental illness. The distinctive characteristics of this collaborative approach to early serious mental illness research could include (1) subject recruitment across several sites; (2) a broad diagnostic focus; (3) a core clinical and neuroscience assessment protocol; (4) longitudinal evaluation of subjects through a range of outcomes; and (5) an iterative approach to psychopathology research. This model represents a method for exploring prodromal phenotypes, for discovering causal risk mechanisms, and for investigating the biological and environmental interactions that define the early course of several disorders, including schizophrenia, bipolar illness, and borderline personality disorder. This strategy could speed discovery of clinical tools most relevant to the earliest stages of serious mental illness; i.e., better methods of screening, diagnosing, and treating mental disorders before symptoms and impairments solidify into chronic disabilities.