Hi Peter,
Thanks again for the help before. In looking through Tortoise as well as the other diffusion afni blogs I had a general question. In tortoise they will upsample files to 1.5 mm resolution, then calculate eddy parameters and then apply the transforms for eddy current correction to the bvecs file.

If I performed this same operation within AFNI, and use 3dAllineate for eddy current correction of an upsampled scan (i.e. 1.5mm iso from the original 2mm iso), I would then extract the rotational components (invert to make it source to base instead of base to source) using Vecwarp -backward, and apply it to the original bvecs.

Question: Since in theory rotational components should be independent of voxel size, I assume I do not have to take the different grid size into account when multiplying the rotational eddy component with the original bvec file. I wanted to make sure I am correct in this assumption and not overlooking anything in regards to how 3dAllineate,Vecwarp etc work.

Also, does going from a non-isotropic voxel dimension to iso affect this value?

If there is an issue, could it be handled by importing a rotational matrix (derived on 1.5mm iso data), putting a master dataset (orig dti space) and then output recalculated values (if this is needed based on answers above). Is there an AFNI tool poised to do such as process if needed?

Thanks,
Ajay



Edited 1 time(s). Last edit at 07/28/2017 12:00PM by AjaySK.