According to the results of research published in the Schizophrenia Bulletin, a generalizable structural imaging signature for schizophrenia based on magnetic resonance imaging studies was generated with a good predictive accuracy of 76%.
Using pooled data from 5 case-control imaging studies of participants both with schizophrenia (n=440) and without schizophrenia (n=501), researchers used advanced multivariate analysis tools to create a neuroanatomical signature for schizophrenia. The robustness and reproducibility of the model were assessed using regional volumes, voxelwise measures, and complex distributed patterns.
The model revealed that patients with schizophrenia had reduced gray matter volume relative to patients without schizophrenia. Regions with reduced volumes included the prefrontal cortex, temporal cortex, parietal cortex, insula, and amygdala. Patients with schizophrenia also had enlarged ventricular regions and palladium.
The predictive accuracy of the model according to multivariate analysis was 76% (area under the curve, 0.84). Cross-validation of the model performed on specific study sites resulted in accuracy of 72% to 77%, suggesting good generalizability according to the study authors.
The neuroanatomical patterns in the model were significantly associated with the negative symptom scores of participants (P <.001), but not positive symptom scores.
In conclusion, the study authors explained that the results “provide among the most robust evidence to date of structural brain abnormalities in adults with schizophrenia. Furthermore, these results emphasize that such signals can be used to derive highly accurate multivariate models that allow for discrimination at the level of individual patients, thereby providing a robust neuroanatomical signature of schizophrenia.”
Rozycki M, Satterthwaite TD, Koutsouleris N, et al. Multisite machine learning analysis provides a robust structural imaging signature of schizophrenia detectable across diverse patient populations and within individuals [published online November 24, 2017]. Schizophr Bull. doi:10.1093/schbul/sbx137