This R03 proposal aims at developing a methodological framework for nonlinear multivariate analysis, using a non-linear pattern classification method, to identify and quantify subtle and spatially-complex brain abnormalities in unaffected family members of schizophrenia patients. Compared to conventional methods based on regions of interest (ROI) that rely only on a few pre-selected manually-labeled ROIs, this automated method will consider all brain regions simultaneously for identification of complex patterns of brain abnormality that cannot be necessarily summarized by a few pre-defined ROIs. Accordingly, a major methodological challenge to be addressed in this project is the development of statistical image analysis and data mining methods for estimating the collection of brain regions or networks that jointly form patterns to characterize schizophrenia as uniquely as possible. Patterns determined by comparison of confirmed schizophrenia patients and healthy controls will be examined on unaffected family members of patients, to test whether individuals that are genetically related to patients display, to some extent, endophenotypes of schizophrenia. Also, unaffected family members will be further compared with patients to identify the morphological profiles that seem directly associated with the phenotype of schizophrenia. The performance of the proposed nonlinear pattern analysis and classification method will be tested on an existing schizophrenia dataset in the Schizophrenia Research Center at the University of Pennsylvania. Although not an immediate goal of this specific proposal, the long-term objective of the proposed work is to apply this automated methodology to various studies, including (1) a large-scale genetic study of schizophrenia involving informative samples, in order to examine questions that cannot be addressed with conventional ROI-based analysis; (2) the quantification and recognition of endophenotypes of schizophrenia in unaffected individuals, which can potentially pave the way for the detection of adolescents that possess brain endophenotypes that put them at risk; (3) clinical studies investigating the added value of quantifying endophenotypes of schizophrenia for clinical diagnosis, especially in difficult cases or outside environments with well-trained psychiatrists, in which unbiased computer-based methods can potentially greatly assist in clinical evaluation and diagnosis. ? ? ?
| Wee, Chong-Yaw; Yap, Pew-Thian; Zhang, Daoqiang et al. (2012) Identification of MCI individuals using structural and functional connectivity networks. Neuroimage 59:2045-56 |
| Wee, Chong-Yaw; Yap, Pew-Thian; Li, Wenbin et al. (2011) Enriched white matter connectivity networks for accurate identification of MCI patients. Neuroimage 54:1812-22 |
| Yap, Pew-Thian; Fan, Yong; Chen, Yasheng et al. (2011) Development trends of white matter connectivity in the first years of life. PLoS One 6:e24678 |
| Shi, Feng; Fan, Yong; Tang, Songyuan et al. (2010) Neonatal brain image segmentation in longitudinal MRI studies. Neuroimage 49:391-400 |
| Tang, Songyuan; Fan, Yong; Wu, Guorong et al. (2009) RABBIT: rapid alignment of brains by building intermediate templates. Neuroimage 47:1277-87 |
| Shen, Dinggang (2009) Fast Image Registration by Hierarchical Soft Correspondence Detection. Pattern Recognit 42:954-961 |
