Classification of schizophrenia patients and healthy controls from structural MRI scans in two large independent samples

► We used VBM (sMRI) to separate individuals with schizophrenia from healthy controls. ► A support vector machine model trained on a large sample (N = 239) gave 71% accuracy. ► Application of the model to a replication sample (N = 277) validated this result: 70%. ► Larger amounts of training subjects gave higher accuracy and more stable models.