Title Page
Abstract
Contents
1. INTRODUCTION 11
1.1. White matter disruptions representing the brain dysconnectivity in schizophrenia 11
1.2. The various results of previous DTI studies regarding WM disruptions in schizophrenia 12
1.3. Tract-based spatial statistics 14
1.3. The aim of the current study 16
2. METHODS 18
2.1. Study population and data collection 18
2.1.1. Asan Medical Center (AMC) 18
2.1.2. Center of Biomedical Research Excellence (COBRE) and Neuromorphometry by Computer Algorithm Chicago (NMorphCH) 19
2.1.3. University of California Los Angeles Consortium for Neuropsychiatric Phenomic LA5c Study (UCLA) 20
2.2. Image processing 20
2.2.1. Quality control and preprocessing MRI image 20
2.2.2. Harmonization 21
2.3. Generalized TBSS analysis 22
2.4. Statistical analysis 24
3. RESULTS 26
3.1. The performance of harmonization 26
3.2. Baseline characteristics 28
3.2. Aberrant registration results in original TBSS 30
3.3. Group Differences in White Matter Parameters in voxel-wise analysis 32
3.3. Group Differences in White Matter Parameters in ROI analysis 38
3.4. Controlling for average FA, MD, RD and AD 41
3.5. Voxel-wise correlation analysis between White Matter Parameters and symptom severity 41
3.6. ROI-based correlation analysis between White Matter Parameters and symptom severity 53
4. DISCUSSION 62
5. Conclusions 68
References 69
Table 1. Comparison of matching variables between the reference site and each target site 27
Table 2. Demographic and clinical characteristics of the subjects 29
Table 3. Significant TBSS whole brain FA group differences (HC > SZ) 36
Table 4. Significant TBSS whole brain AD group differences (HC > SZ) 37
Table 5. Significant TBSS whole brain RD group differences (HC < SZ) 37
Table 6. Group comparisons of mean FA (ROIs with significant difference are shown) 39
Table 7. Group comparisons of mean RD (ROIs with significant difference are shown) 40
Table 8. Group comparisons of mean AD (ROIs with significant difference are shown) 40
Table 9. Location of Significant positive correlation Between PANSS positive and FA in schizophrenia 50
Table 10. Location of Significant positive correlation Between PANSS negative and FA in schizophrenia 50
Table 11. Location of Significant negative correlation Between PANSS positive and MD in schizophrenia 51
Table 12. Location of Significant negative correlation Between PANSS negative and MD in schizophrenia 51
Table 13. Location of Significant negative correlation Between PANSS negative and AD in schizophrenia 52
Table 14. Location of Significant negative correlation Between PANSS negative and RD in schizophrenia 52
Table 15. Positive Correlation between FA and PANSS positive in schizophrenia (ROIs with significant correlation were shown) 55
Table 16. Positive Correlation between FA and PANSS negative in schizophrenia (ROIs with significant correlation were shown) 56
Table 17. Negative Correlation between MD and PANSS positive in schizophrenia (ROIs with significant correlation were shown) 57
Table 18. Negative Correlation between MD and PANSS negative in schizophrenia (ROIs with significant correlation were shown) 58
Table 19. Negative Correlation between RD and PANSS positive in schizophrenia (ROIs with significant correlation were shown) 59
Table 20. Negative Correlation between RD and PANSS negative in schizophrenia (ROIs with significant correlation were shown) 60
Table 21. Negative Correlation between AD and PANSS positive in schizophrenia (ROIs with significant correlation were shown) 61
Table 22. Negative Correlation between AD and PANSS negative in schizophrenia (ROIs with significant correlation were shown) 61
Figure 1. Flow-chart of generalized TBSS analysis. Flow-chart showing the detailed procedures of generalized TBSS analysis. Abbreviations: MNI, Montreal Neurological Institute 23
Figure 2. Results of DTI images registration of several subjects in original TBSS. In the first row, normal registration images are shown. Aberrant registration results are illustrated in the... 30
Figure 3. Group differences of FA in original TBSS analysis. Significance threshold was set to P<0.05, FWE corrected. Significant clusters are shown in red-yellow. The WM... 31
Figure 4. Significant decreases in FA in schizophrenia patients. Significance threshold was set to P<0.05, FWE corrected. Significant clusters are shown in red-yellow. The WM skeleton is... 33
Figure 5. Significant decreases in AD in schizophrenia patients. Significance threshold was set to P<0.05, FWE corrected. Significant clusters are shown in red-yellow. The WM skeleton is... 34
Figure 6. Significant increases in RD in schizophrenia patients. Significance threshold was set to P<0.05, FWE corrected. Significant clusters are shown in red-yellow. The WM skeleton is... 35
Figure 7. Regions showing positive correlation between FA and PANSS positive in schizophrenia patients. Significance threshold was set to P<0.05, FWE corrected. Significant... 44
Figure 8. Regions showing positive correlation between FA and PANSS negative in schizophrenia patients. Significance threshold was set to P<0.05, FWE corrected. Significant... 45
Figure 9. Regions showing negative correlation between MD and PANSS positive in schizophrenia patients. Significance threshold was set to P<0.05, FWE corrected. Significant... 46
Figure 10. Regions showing negative correlation between MD and PANSS negative in schizophrenia patients. Significance threshold was set to P<0.05, FWE corrected. Significant... 47
Figure 11. Regions showing negative correlation between AD and PANSS negative in schizophrenia patients. Significance threshold was set to P<0.05, FWE corrected. Significant... 48
Figure 12. Regions showing negative correlation between RD and PANSS negative in schizophrenia patients. Significance threshold was set to P<0.05, FWE corrected. Significant... 49