Previous analyses of grey and white matter volumes have reported that schizophrenia is associated with structural changes. Deep learning is a data-driven approach that can capture highly compact hierarchical non-linear relationships among high-dimensional features, and therefore can facilitate the development of clinical tools for making a more accurate and earlier diagnosis of schizophrenia.

To identify consistent grey matter abnormalities in patients with schizophrenia, 662 people with schizophrenia and 613 healthy controls were recruited from eight centres across China, and the data from these independent sites were used to validate deep-learning classifiers.

We used a prospective image-based meta-analysis of whole-brain voxel-based morphometry. We also automatically differentiated patients with schizophrenia from healthy controls using combined grey matter, white matter and cerebrospinal fluid volumetric features, incorporated a deep neural network approach on an individual basis, and tested the generalisability of the classification models using independent validation sites.

We found that statistically reliable schizophrenia-related grey matter abnormalities primarily occurred in regions that included the superior temporal gyrus extending to the temporal pole, insular cortex, orbital and middle frontal cortices, middle cingulum and thalamus. Evaluated using leave-one-site-out cross-validation, the performance of the classification of schizophrenia achieved by our findings from eight independent research sites were: accuracy, 77.19–85.74%; sensitivity, 75.31–89.29% and area under the receiver operating characteristic curve, 0.797–0.909.

These results suggest that, by using deep-learning techniques, multidimensional neuroanatomical changes in schizophrenia are capable of robustly discriminating patients with schizophrenia from healthy controls, findings which could facilitate clinical diagnosis and treatment in schizophrenia.

Schizophrenia is a complex mental disorder with high heritability and polygenic inheritance. Multimodal neuroimaging studies have also indicated that abnormalities of brain structure and function are a plausible neurobiological characterisation of schizophrenia. However, the polygenic effects of schizophrenia on these imaging endophenotypes have not yet been fully elucidated.

To investigate the effects of polygenic risk for schizophrenia on the brain grey matter volume and functional connectivity, which are disrupted in schizophrenia.

Genomic and neuroimaging data from a large sample of Han Chinese patients with schizophrenia (N = 509) and healthy controls (N = 502) were included in this study. We examined grey matter volume and functional connectivity via structural and functional magnetic resonance imaging, respectively. Using the data from a recent meta-analysis of a genome-wide association study that comprised a large number of Chinese people, we calculated a polygenic risk score (PGRS) for each participant.

The imaging genetic analysis revealed that the individual PGRS showed a significantly negative correlation with the hippocampal grey matter volume and hippocampus–medial prefrontal cortex functional connectivity, both of which were lower in the people with schizophrenia than in the controls. We also found that the observed neuroimaging measures showed weak but similar changes in unaffected first-degree relatives of patients with schizophrenia.

These findings suggested that genetically influenced brain grey matter volume and functional connectivity may provide important clues for understanding the pathological mechanisms of schizophrenia and for the early diagnosis of schizophrenia.