Brain structure in schizophrenia: what do we know and what next?

Meta-analyses and mega-analyses of structural MRI and diffusion MRI studies in thousands of people with schizophrenia have shown reduced brain volumes and structural connectivity relative to healthy control participants. Some of these structural differences probably present pre-morbidly. Severity of schizophrenia symptoms and cognitive impairment consistently correlate. Convergent findings with different analytical methods show that the results are not artefactual. Post-mortem studies add validity and point to a pathophysiology of reduced neuronal size and dendritic arborisation, which is multifactorial. Structural MRI changes probably occur between being in the at-risk state and the first episode of psychosis, and ongoing reductions in brain volume might be associated with poor outcomes. By the standards of observational epidemiology, some findings are convincing or highly suggestive, but many are only suggestive or weak, and more studies are needed. Progress in scientific understanding and evaluating clinical applications will require clear hypotheses, adequately powered studies, and standardised methods and reporting.