Brain imaging studies in schizophrenia have typically involved single assessment and cross-sectional designs, while longitudinal studies rarely incorporate more than two time points. While informative, these studies do not adequately capture potential trajectories of neurobiological change, particularly in the context of a changing clinical picture. We propose that the analysis of brain trajectories using multiple time points may inform our understanding of the illness and the effect of treatment. This paper makes the case for frequent serial neuroimaging across the course of schizophrenia psychoses and its application to active illness epsiodes to provide a detailed examination of psychosis relapse and remission.