Machine Learning and Pattern Analysis of MRI in Neuropsychiatric Disorders
Neuropsychoatric disorders, and in particular schizophrenia which has been of emphasis in our research, are associated with subtle but consistent structural and functional deficits. Although conventional approaches, such as ROI- or voxel-based analyses have shown robust group differences, they have been insufficient for single-subject classification. This is due to the very subtle and spatially distributed nature of brain changes. Since 2005, our group has been involved in research applying state of the art image analysis, as well as machine learning methods to schizophrenia and other neuropsychiatric disorders [1-6]. In our original study  we found that SVM-based machine learning was able to provide quite promising single-subject classification, based on a complex pattern of neuroanatomical changes:
(LEFT) ROC curve of schhizophreina vs. control classification, using cross-validation; (RIGHT) effect size map of group differences, and average template after warping all MRIs to a standardized space in order to measure regional neuroanatomical patterns.
Inter-site, Cross-Ethnicity Integrative Study of Schizophrenia
In 2016 we embarked at an inter-site mega analysis, pooling data from multiple sites, initially in the USA, Germany, and China, aiming to obtain robust and reproducible across sites and ethnicities classifiers of schizophrenia. Results from n analysis of approximately 1,000 individuals provided ROC curves and regional effect size maps shown below:
(LEFT) ROI curve for individual classification between schizophrenia patients and controls; (RIGHT) Effect size map of group differences
- Satterthwaite, T.D., et al., Structural Brain Abnormalities in Youth With Psychosis Spectrum Symptoms. JAMA Psychiatry, 2016.
- Zhang, T., et al., Heterogeneity of Structural Brain Changes in Subtypes of Schizophrenia Revealed Using Magnetic Resonance Imaging Pattern Analysis. Schizophr Bull, 2015. 41(1): p. 74-84.
- Koutsouleris, N., et al., Individualized differential diagnosis of schizophrenia and mood disorders using neuroanatomical biomarkers. Brain, 2015. 138(Pt 7): p. 2059-73.
- Kambeitz, J., et al., Detecting neuroimaging biomarkers for schizophrenia: a meta-analysis of multivariate pattern recognition studies. Neuropsychopharmacology, 2015. 40(7): p. 1742-51.
- Fan, Y., et al., Unaffected family members and schizophrenia patients share brain structure patterns: a high-dimensional pattern classification study. Biol Psychiatry, 2008. 63(1): p. 118-24.
- Davatzikos, C., et al., Whole-brain morphometric study of schizophrenia revealing a spatially complex set of focal abnormalities. Archives of General Psychiatry, 2005. 62(11): p. 1218-1227.