if you refer to an affine transformation from e.g. coregistration,
realignment, normalisation (affine parts only) - yes and no. The
Diffusion toolbox will take care of this information, but it is
currently designed to reorient gradient information before you do
tensor estimation, not to correct tensor data itself.
To do preprocessing with the Diffusion toolbox, you would
1) add gradient and b value information to each of your DW and b0
images using "(Re)Set DW information"
2) do any spatial processing you want, including reslicing
3) run "Copy/Reorient DW information" to transfer DW information from
the non-resliced images to the resliced/normalised ones
4) estimate the tensor, using the reoriented gradient system.
This will work for any direction scheme - provided it has enough
directions to compute a tensor.
If you want to reorient tensor maps, then you'd have to set up 6
imcalc's computing the rotation formula element by element.
Hope this helps,
Quoting kenichi oishi <[log in to unmask]>:
> Hi all,
> Is there a good way to apply a transformation matrix of SPM5 to a diffusion
> tensor field with appropriate tensor reorientation?