Computation of motion parameters (sct_fmri_moco)

SCT
1

Hi SCT team,

I have a question regarding the computation of volume-wise motion parameters, when using sct_fmri_moco. Indeed, I have noticed that those parameters are computed by taking the mean over the slices for each volume, as done in l426 of moco.py (SCT v6.1).

# Calculating the slice-wise average moco estimate to provide a QC file
moco_param.append([np.mean(np.ravel(im_warp_XYZ[0].data)),np.mean(np.ravel(im_warp_XYZ[1].data))])

I am wondering if using the mean of the absolute motion for all slices might provide a clearer picture. Currently, it seems like negative motion in one slice might be offset by positive motion in another. What do you think about considering the absolute values for a more straightforward representation of motion?

Thanks for your help!
Nawal

2 Likes
2

Hi @Nawal_Kinany,

Great catch! Thank you for reporting it. I agree with your suggestion and have opened an issue about it.

Cheers,
Julien

3

Thank you!

Cheers,
Nawal

4

We need further inputs from the community to decide how we want to implement the feature, please chip in: Averaging of motion parameters should use unsigned displacement · Issue #4344 · spinalcordtoolbox/spinalcordtoolbox · GitHub