Navigation and service

Development of Postprocessing Methods for HARDI Data Analysis

The main goal of the project is to overcome the limitations of the rank-2 diffusion tensor model concerned with the assumption that the DW signal is Gaussian. The project is dedicated to the development of reconstruction methods, segmentation and tractography algorithms able to infer multiple fibre structure from High Angular Resolution Diffusion Imaging (HARDI) data. Images in the HARDI scheme are acquired with DW gradients applied along a large number (≥ 45) of uniformly distributed set of directions on the surface of a unit sphere. The HARDI scheme is particularly promising for clinical applications, because it contains useful information about the angular dependence of the DW signal. The HARDI technique may help to characterise selective fibre loss in diseases associated with white matter degeneration. Compared to the rank-2 diffusion tensor model, HARDI yields more accurate and detailed information about diffusion in the brain. It is probably the most time-efficient data acquisition strategy to reconstruct fibre orientations, which can benefit white matter tractography. Various reconstruction algorithms for HARDI data analyses can be divided into two categories: one based on the q-space theory (Q-ball, PAS-MRI, DKI, DOT, DSI, DPI) and the other based on the slow exchange limit (multi-tensor fitting, spherical deconvolution). While both methodologies are significantly different in recovering the angular information from the measured diffusion data, they both define a spherical continuous function within each voxel that encodes the anisotropy of diffusion.

Development of Postprocessing Methods for HARDI Data AnalysisADC profile approximation for different types of voxels: Spherical harmonic decomposition method was used to model ADC profiles for isotropic, one-fibre containing, and multi-fibre containing voxels from HARDI data.