Alignment of data from SUMA and diffusion data in TORTOISE

I have diffusion data that I have processed through TORTOISE, using a T1 processed in Freesurfer as a structural reference. When I use the Make_SPEC_FS command through SUMA, giving the Freesurfer output to this command, I am noticing that the SUMA outputs are in a different space from the Freesurfer data. I am trying to view my TORTOISE processed diffusion data results as a overlay onto the SUMA generated surface. I have also tried using the structurals (.nii format) generated by SUMA as references for TORTOISE processing as an alternative approach. The T2 output from TORTOISE is registered to the SUMA processed T1, but the output diffusion data is not in the same space.

Just to verify - are you using the Freesurfer T1 as the "structural" or the "reorientation" volume?

Also what version of TORTOISE are you using?

I am doing a rigid registration of the fat sat T2 weighted dataset to the SUMA output T1.nii. I am then using the registered T2 weighted dataset as the structural image. I am not using the reorientation option. I am using TORTOISE 3.2.0 on Biowulf (the DIFFPREP module only)

Thanks for the info! Would you mind sharing your code command for doing the registration from T2 to T1?

Yes, I am using FLIRT in FSL: flirt -in T2.nii -ref T1_SUMA.nii -dof 6 -cost mutualinfo -out T2toT1_SUMA

The T1_SUMA.nii volume is the T1.nii produced by SUMA. When I view the output of the T2 registration overlaid onto the SUMA based T1:

This is the output of the DMC structural output and the registered T2, and they seem to align well:

The corrected DWIs produced by TORTOISE however are not in the same space:

Thanks. Where are you applying the transformation from flirt to the corrected DWIs?

From what you posted I see:

T2 --Aligned --> T1
DMC Structural --Aligned--> "Registered T2" (assuming output flirt T2toT1_SUMA.nii)
Corrected DWI --Aligned--> Registered T2

I am not directly applying the transformation from FLIRT to the corrected DWIs. I am using the FLIRT transformed T2 as my structural input into TORTOISE. The output DMC structural is from TORTOISE.

When I overlay the DWIs that were corrected using the FLIRT corrected T2 (T2 -> SUMA T1) they are in the same space as the original Freesurfer T1 (Freesurfer T1 before being input to SUMA):

I think that I was able to find a workaround. Originally I was just using the @SUMA_Make_Spec_FS -sid subjct1, but when I am using @SUMA_Make_Spec_FS -NIFTI -sid subject1, the SUMA outputs are in the same Freesurfer space (no shift)


I'm curious what version of AFNI you are using?

Also, in the FATCAT+TORTOISE tutorial, there are scripts for helping with a lot of these things, like aligning FreeSurfer T1w output to a T2w/b=0 reference, while also bringing along a lot of ROIs, using the fat_proc_* programs in AFNI. For example, see here for that step in processing.


Thanks! I will check that for my next questions. I am using an older version of AFNI : AFNI_19.3.16 'Nero'

OK. I would strongly recommend updating your AFNI, since there have been so many updates/tweaks and improvements in the past 5 years, across the code base.

For example, this doc page mentions how @SUMA_Make_Spec_FS creates more useful datasets, automatic QC images, and stats files. I don't think in modern usage you would need to add -NIFTI either, as that behavior has become the default.