DWI already registered to T2w?


(Hope this is the correct place to post questions about dHCP data.)

I am looking at the diffusion data in dHCP. I found that the data documentation say “V. Diffusion data are aligned to high-resolution structural (T2-weighted) space using boundary-based registration8​ , 9 ​on the average attenuation volume for the b=1000 s/mm2​ ​ shell (i.e. b1k/b0). This transformation is combined with a non-linear registration1​ 0​ of the T2w volume to the 40 weeks template1​ 1​ to allow transformations between diffusion and atlas spaces.

I wonder if this means that the provided dMRI data is already in the T2w space, e.g. XX_desc-preproc_dwi.nii.gz and XX_desc-restore_T2w.nii.gz.

If this was correct, while for most subjects the DWI and T2w visually looking to be aligned well, some subjects are not, e.g. sub-CC00060XX03_ses-12501.

Otherwise, could you let me know if there are additional warping steps needed.

Thank you!!!


Dear Fan,

Thanks for your message. Yes, this is the right place to post dHCP-related queries.

No, preprocessed data are in native diffusion space. To bring them to T2w space, you can apply the rigid body transformation matrix (i.e., *_from-dwi_to-T2w_mode-image.mat) using FSL FLIRT (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FLIRT).

Hope this helps, cheers,

Hi Matteo,

Thank you for the information. That works very well!