3dvolreg -maxdisp exact meaning

AFNI Message Board
1

Dearest AFNItes

I am kinda embarrassed I just realised I wasn't 100% confident in how mm and mm_delt are calculated from the 6 motion parameters in 3dvolreg, and in particular what the actual distance in mm_delt represents. [For enorm, I am on terra firma].

Apologies if I am missing the documentation - I had a good look round but failed.

Thanks!
Fred

2

Hi, Fred-

Sure, a couple places to look are in the afni14_alignment.pdf AFNI Bootcamp handout:

afni_open -aw afni14_alignment.pdf

and slides 41-45, in particular, about motion parameter considerations, including Enorm.

That material is also covered in this AFNI Academy recording:
Alignment (part 2/4): EPI-EPI

And in this paper about using afni_proc.py, there is also a bit of description of it as an L_2 norm of the first differences of the 6 motion parameters and why that is reasonable (even though 3 have units of mm and 3 have units of rads):

  • Reynolds RC, Glen DR, Chen G, Saad ZS, Cox RW, Taylor PA (2024). Processing, evaluating and understanding FMRI data with afni_proc.py. Imaging Neuroscience 2:1-52.

Please let us know about any further questions.

--pt

3

and the 3dvolreg -maxdisp should be the maximum displacement by any voxel at any time point in the brain max when 3dvolreg is doing its EPI-EPI alignment processing. So, like the largest distance a within-brain-mask voxel moves "as the crow flies".

-pt

4

Hi Paul, thanks for the quasi-instantaneous answer! It was the -maxdisp metric that was confusing to me, in particular the difference in the way mm_delt was calculated compared to enorm (which has the formula conveniently stuck into 3dvolreg help).

I think your answer is what I thought mm and mm_delt were: the distance traversed by the 'moving-est' voxel within the brain mask (presumably one at the edge of brain when there's a rotation).

For a dataset where we're evaluating a motion reduction device, I figured I should really make sure I know what we are measuring :-) I realise I am being extremely lazy by not just looking at the source code since it's sitting there!

cheers
Fred