Diffusion Tensor Imaging-based Cerebral White Matter Microstructure Analysis Methods And Application

The cerebral white matter is an infrastructure that connects the cerebral cortex and subcortical areas,and is a medium for efficient communication between brain regions.Diffusion tensor imaging(DTI)is based on Gaussian model and uses anisotropic diffusion tensor to reconstruct the microstructure of brain tissue on the basis of single exponential diffusion imaging.Its four quantitative parameters,including fractional anisotropy(FA),mean diffusivity(MD),axial diffusivity(AD),and radial diffusivity(RD)are very sensitive to the white matter microstructure(WMM),so it is often used to detect the changes of the WMM.DTI-based WMM analysis methods are often used in the studies of brain development,aging,and cognitive function,as well as in the studies of the pathology of the neuropsychiatric disorders.However,the current WMM analysis methods are inconsistent and unsystematic,which leads to the limitations of research results,and it is also difficult to make horizontal comparisons between studies.The dissertation proposes a framework for the analysis of WMM against this situation and the work of the dissertation is shown below:1.The multi-scale and multi-parameter analysis method framework is proposed,which provides a systematic and complete methodology for the WMM analysis.The framework consists of three parts:First,the framework analyzes the parameters images at the regional scale to identify the changes in the regional scale of the WMM.Second,the framework analyzes DTI parameters images at the voxel level of the skeleton.The method has small registration errors and does not require smoothing.Finally,the framework analyzes DTI parameters images at the whole-brain voxel scale.The analysis method is sensitive to parameters conforming to the Gaussian distribution and can automatically detect the changes in the WMM of any local clusters.The regional-scale analysis method provided by the framework makes the regional-scale DTI parameters conform to the Gaussian distribution,and has the advantages of fewer multiple comparisons,high statistical power,and no registration errors.2.In the dissertation,the patterns of abnormal WMM in FED and RD are revealed from the perspective of multi-scales and multi-parameters,and the results show that there are different patterns.The right Superior/Middle Frontal Gyrus microstructural abnormalities were more prominent in FEMDD than in RMDD,and the MD,AD and RD parameters of Superior/Middle Frontal Gyrus were not correlated with the number of episodes in the whole patient group.The left Superior/Middle Frontal Gyrus WMM abnormalities were more prominent in RMDD than in FEMDD and the MD,AD and RD parameters were negatively affected by the number of episodes throughout the patient group.The result indicated that the focal brain regions of the WMM abnormalities were different between FEMDD and RMDD,with prominent right Superior/Middle Frontal Gyrus microstructural abnormalities in FEMDD and prominent left Superior/Middle Frontal Gyrus WMM abnormalities in RMDD.3.By the comparison experiment,we found that FA is more sensitive to regional scale and skeletonized voxel-level analysis,while MD,AD,and RD are more sensitive to whole-brain voxel-level analysis.The results from regional scale analysis and from voxellevel analysis complement each other and verify each other,and the use of framework makes the results more complete.The framework proposed in the dissertation describes and analyzes the WMM from the two dimensions of multi-DTI parameters and multi-geometrical scales,which overcomes the limitations of a single analysis method.Furthermore,the results obtained from the comparative study provide potential biological markers of depression from the perspective of WMM.The potential biological markers are expected to be used in the early screening of depression check.In addition,the WMM changes are closely related to cognitive function,so the WMM analysis method framework proposed in the dissertation can also be applied to cognitive impairment screening,cognitive status assessment and selection of specific personnel in the army.