Parkinson’s Disease Cases Data Analysis Based On High-order Diffusion Tensor Imaging Algorithm

Learning the Fractional Anisotropy(FA) and Mean Diffusion(MD) of substantia nigra based on Diffusion Tensor Imaging(DTI) model is a new method to stusy Parkinson’s disease(PD)pathology. Existing pathology study mathords have two main problems as DTI method cannot reconstructe complex fiber:one is traditional high tensor model have the problem of fiber estimation, and the second one is existing pathology study lacks index. In this paper, the main work and achievements which based on high order tensor fiber oritation eatimation and pathology study index are as follows:Firstly, in view of the difficult problem of fiber direction extraction of DTI, the paper proposes non-negative constrainted higher order tensor deconvolution(HOT-SD). Make angular resolution, angle error and cross area contrast experiments between Cartesian Tensor fiber orientation distributions(CT-FOD) and Constrained Spherical deconvolution(CSD) with simulated data and clinical data. Results show that HOT-SD algorithm have advantages of angular resolution and robustness through the sparse representation of fiber orientation.Secondly, in order to get the reconsitution effect of HOT-SD, we use streamline algorithm.The contrast experiments results compared with Q-ball Imaging(Q-ball) through simulated data and clinical data show that the reconsitution of streamline based on HOT-SD have advantages of stability and reliability.Thirdly, in view of the lack of scientific pathological methods, the paper proposes Generalized Fractional Anisotropy(GFA) and General Fractional Anisotropy Peak(GFApeak)based on High Angular Rusolution Diffusion Imaging(HARDI), fiber Mean-squared Displacement(MSD) index and average density based on fiber bundle. The paper reconsitution the fiber of substantia nigra base above-mentioned method, and each index compared with DTI and Q-ball. Analysis results through hypothesis test show that patients with Parkinson’s disease have characteristics that FA, MSD and average density of fiber exist significant difference.As a brief, this paper proposes methods for the problems of fiber orientation estimation and pathological research. In order to test and verify the reliability of method, with data of 21 patients with Parkinson’s disease and 16 health people provided from Department of Neurology of Tongji Medical College Affiliated Union Hospital, Huazhong University of Science and Technology the paper uses hypothesis-testing tools to analyze the results of each index.Experimental results show the advantages of angular resolution and robustness, the reliability of the studying index, can provide reliable basis for pathology of Parkinson’s disease.4. In this paper, HOT-SD, GFA and GFApeak based on HARDI, fiber MSD index and average density based on fiber bundle which can improve the quality of imaging.