[maedoc]

After some playing around, it seems tracks are saved in t1 space, not diffusion space. As a interesting twist, one has to apply the inverse of the affine w/o translation, i.e. just the 3×3 not 4×4 affine.

Knowing this, obtaining voxel coords & label values is easy.

It would be nice if this were documented somewhere in the file formats section.

[Anonymous]

Hi Maedoc, diffusion modeling outputs (diffusion tensor, ODFs etc.) are always saved in T1-space. So, usual tractography would results into tracks being generated and saved in T1-space. An exception to this would be if diffusion outputs from diffusion_coord_outputs directory is used for tractography (this directory is only generated if --output-diffusion-coordinate flag was used).

As a interesting twist, one has to apply the inverse of the affine w/o translation, i.e. just the 3×3 not 4×4 affine.

Above statement is not entirely accurate. BDP supports wide variety of data transformation across T1 and diffusion space, and so users do not have to do it themselves. Please take a look at “TRANSFORMATION FLAGS” section on flags documentation, in particular flags: --transform-t1-volume, --transform-interpolation, --transform-data-only. This is the recommended way to transfer data from one space to another.

Flag documentation: http://brainsuite.org/processing/diffusion/flags/#fullflaglist

[maedoc]

I was aware of the transformation flags which is how I ruled out that the tracks were in the DWI space.

My question was, how can I recover voxel coordinates of tracks? If the coordinates of the curves in the dft file are in T1 space, normally one inverts the 4×4 T1 affine and applies it to the curve coordinates, but this produced a translation that was not correct. Rather, I had to invert take the inverse of affine(1:3, 1:3) and apply it to curve coordinates.

I verified this by plotting a slice of the T1 space label image with imagesc (transposed & with upside down) and plotting the seeds of all the curves within that slice. The seed points line up nicely with the image, whereas this fails completely with the full 4×4 affine.

I repeat, it would be nice to have confirmation on this as I did not read enough of the source code to confirm 100%.

[Anonymous]

The coordinates in the BrainSuite’s surface-space (which is same space as the curves’ space) are related to voxel-coordinates simply by scaling with voxel-dimension (and not by 3×3 inverse affine-matrix). For x-coordinates this would be: voxel_coord_x = curve_coord_x/pixdim[1]; and likewise for y and z — Note that this expression uses c-based indexing, i.e. array (and voxel) indexing start from 0.

Another similar question: https://forums.brainsuite.org/forums/topic/surface-mesh-space/

[Author]

Posts