Antonucci LA, et al. - Researchers performed this study to assess whether patients with schizophrenia (SCZ) could be discriminated from healthy controls (HC) using a classification approach encompassing risk factors, cognition, and their associations. They assumed that greater HC-SCZ classification accuracy would be achieved with cognition and that the discrimination performance of cognition would be improved with combined gene-environment stratification. Using support vector classification and repeated nested cross-validation, they trained GWAS-based genetic, environmental, and neurocognitive classifiers to separate 337 HC from 103 SCZ. Patients with SCZ were discriminated from HC via the cognitive classifier with a Balanced Accuracy (BAC) of 88.7%, followed by environmental (BAC = 65.1%) and genetic classifiers (BAC = 55.5%). The results support cognitive deficits as core features of the phenotype of SCZ and that cognitive features alone provide the greatest amount of information for SCZ classification. Consistent with their being risk factors, they observed modulation of HC-SCZ classification performance of cognition by gene-environment stratification, perhaps presenting another target for improving early identification and intervention strategies.