bvec rotation

Submitted by Tahereh on

Hi TORTOISE experts. Thanks for your nice and handy software.

My question is about rotating bvec. Basically, I have data which has been acquired 90 degrees rotated from stereotaxic orientation, on a Siemens Magnetom Prisma 3T scanner. Both anatomical (T1,T2) and DWI (AP-PA) data have this 90 degree rotation. So when I rotate them back to normal (stereotaxic) orientation in AFNI, I think I should compensate for this rotation by re-adjusting the bvec file, right? So, is this something I can do in TORTOISE or should I try to do it in AFNI?

Alternatively, if I provide both DWI and T2 to TORTOISE in their original orientation (i.e. if I don't rotate them 90 degrees), will this affect the functionality of your software for distortion correction?

Thanks a lot for any hints.

Submitted by nayaka on Wed, 08/04/2021 - 10:57



Please go ahead and perform the necessary reorientation of the structural data. Since you are using AFNI, I am assuming you would use the axialization procedure. 

Provide the reoriented structural data to DRBUDDI and the processing will output the final distortion corrected diffusion data into the reoriented structural data space. Rotation of bvecs is  taken care of by the processing. You can check that by computing bvecs from the final.bmtxt file using the command TORTOISEBmatrixTOFSLBVecs. This will output bvecs and bvals from the final.bmtxt.

Please ensure to perform the post processing QC by fitting a tensor to the final.list file and compute the color maps/ glyph maps from the DT.nii. Use the ComputeGlyphMaps and ComputeNSDECMap executables to generate DEC and/or Glyph maps. This qc procedure is to ensure the final reorientation of data is correct. Please let us know if the output is not as expected. 

Thank you.