Calculating X,Y,and Z diffusion values from tensor?

1

Hello,

I am interested in calculating the DTI-ALPS metric in some data using mrtrix. This essentially consists of getting the x, y, and z diffusion diffusion values into standard or template space and then extracting these tensor values out of ROI’s in specific tracts.

Would I be right in assuming that I could use the tensor2metric command with the -vector/-value flag to get an x,y,z diffusion images? What other flags or specifics of the command would I need to get x,y, and z, or V1,V2, and V3 from a tensor.mif file?

Thanks,
Daniel

2

Hi @CallowBrainProject,

Assuming you’re after the eigenvectors & eigenvalues, yes. You can use the -num option to get specific ones, or all of them (e.g. -num 1:3) – they’ll then be concatenated in the output file. You can also have the eigenvectors modulated differently using the -modulate option – by default, they’re modulated by the FA, but you may prefer to get unit eigenvectors (-modulate none) or eigenvectors scaled by the corresponding eigenvalue (-modulate eigval).

The eigenvectors are computed with respect to the scanner / world coordinate system, by the way. Not sure whether that matches what you call ‘standard space’ though…

All the best,
Donald.

3

DTI-ALPS metric requires diffusivity along x/y/z-axis, which is not the eigen vectors but the intersection of tensor with axis. They are the first three elements of your tensor.mif image, as dwi2tensor doc states “The tensor coefficients are stored in the output image as follows: volumes 0-5: D11, D22, D33, D12, D13, D23”

1 Like