Home > Author > >

" Diffusion tensor imaging (DTI), or tractography, is an in vivo MRI technology that uses water diffusion in brain tissue to visualize in stunning detail the brain's three-dimensional white matter anatomy. DTI is made possible by characterizing water diffusion in tissues by means of a mathematical tool called a tensor, based on matrix algebra: (1) a 3 x 3 matrix, called a diffusion tensor, is used to characterize the three-dimensional properties of water molecule diffusion; (2) from each diffusion tensor, the three pairs of eigenvalues and eigenvectors are calculated using matrix diagonalization; and (3) the eigenvector that corresponds to the largest eigenvalue is selected as the primary eigenvector. A 'streamline' algorithm then creates "tracts" by connecting adjacent voxels if their directional bias is above some treshold level. Does the orientation of the primary eigenvector coincide with that of the actual axon fibers in most white matter tracts ? Takahashi et al. (2011), for example, have demonstrated that radial organization of the subplate revealed via tractography directly correlates with its radial cellular organization, and G. Xu et al. (2014) were able to determine that transient radial coherence of white matter in the developing fetus reflected a composite of radial glial fibers, penetrating blood vessels, and radial axons. "

, Bioinspired Devices: Emulating Nature’s Assembly and Repair Process


Image for Quotes

 quote : Diffusion tensor imaging (DTI), or tractography, is an in vivo MRI technology that uses water diffusion in brain tissue to visualize in stunning detail the brain's three-dimensional white matter anatomy. DTI is made possible by characterizing water diffusion in tissues by means of a mathematical tool called a tensor, based on matrix algebra: (1) a 3 x 3 matrix, called a diffusion tensor, is used to characterize the three-dimensional properties of water molecule diffusion; (2) from each diffusion tensor, the three pairs of eigenvalues and eigenvectors are calculated using matrix diagonalization; and (3) the eigenvector that corresponds to the largest eigenvalue is selected as the primary eigenvector. A 'streamline' algorithm then creates