When considering single voxel angular uncertainty, and SNR ≥60 was required for ☙° uncertainty this can have consequences for accurate fiber tracking. The SNR requirements for eigenvector directions were more stringent: SNR ≥25 and ≥45 was required for eigenvector angular deviations of ±4.5° uncertainty with muscle fractions of 1 and 0.5, respectively. In the latter paper, simulations were performed with six standard diffusion gradient directions, and identified the optimum b value to be between 400 and 700 mm 2/s, and SNR of 25 for the baseline image was required to obtain an accuracy of 5 % in the DTI indices when the voxel contained only muscle tissue. Simulation studies have been performed incorporating low T2 values, % muscle fraction, DTI indices (eigenvalues, and FA) typical values for muscle in order to obtain the optimum acquisition parameters and SNR requirements to measure DTI indices with a given accuracy (Damon 2008). However, as seen later, muscle tractography is still challenged by the lower FA values, lower SNR as well as the admixture of fat in muscle voxels. Muscle fibers do not cross and present simpler geometrical constructs compared to brain white matter fibers may be one of the reasons for the success of the simpler deterministic algorithms. In muscle fiber tracking, FACT and streamline tractography methods have been used in almost all the studies reported so far. Tractography is an area of active research and a recent review comparing different tractography algorithms is discussed by Lazar ( 2010). 2003) and tractography using advanced models for fiber crossings (Malcolm et al. More advanced methods in tractography include probabilistic tractography that provides a measure of the probabilities of connections (Behrens et al. Several computational methods are used to perform basic streamline tractography: Euler’s method (following the eigenvector or tangent for a fixed step size) and second order or fourth order Runge–Kutta method (where the weighted average of two or four points is used for each successive step) (Basser et al. Another related approach is streamline tractography which also works by successively stepping in the direction of the principal eigenvector. Tract selection (using anatomical ROIs) and seed placement are highly interactive resulting in a strong operator dependence of fiber tracts. ![]() Termination occurs when the FA value falls below a preset threshold or the orientation of the fiber changes by a larger angular threshold. In this algorithm, the algorithm starts from seed voxels defined by the user or by FA thresholds and follows the eigenvector direction in each voxel. ![]() Several tractography algorithms have been proposed to connect eigenvectors the most commonly used is known as fiber assignment by continuous tracking, FACT (Mori and van Zijl 2002 O’Donnell and Westin 2011). The ultimate goal is to obtain the fiber tracks which represent the physiological unit of the muscle fiber bundles. The eigenvectors are conveniently represented in color and the following conventions are used in DTI visualization: with the x-projections mapped to red, y-projections to green, and z-projections to blue.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |