Sct_fmri_moco with poly=0

Hi SCT team,

I recently wanted to compare the performance of slice-wise motion correction with and without the polynomial regularization contained in sct_fmri_moco (using sct 4.3 on Ubuntu 18.04).

However, when I set poly=0 (which I understood from the help text to mean that no regularization will be performed), I didn’t see any effect of the motion correction, i.e. the resulting images looked identical to the input. Interestingly, the motion parameter .tsv file also contained only zeros.

I then checked the underlying ANTs routine (using ants 2.3.1 on Ubuntu 18.04) and observed a similar behavior, i.e. no correction being performed.

I am just wondering if I misunderstood the usage of sct_fmri_moco or whether this is a known bug / expected behavior?

All the best,
Falk

Hi Falk,
I can reproduce this bug. Thanks a lot for reporting it. An issue has been opened and will be addressed with top priority.
Cheers,
Julien

Hi Julien,
Thanks a lot for responding so quickly, that’s great!
Cheers,
Falk

Hi Julien,
Thanks once more for looking into this - is there any news on when the fix might be available in a release?
Cheers,
Falk

this PR will likely be merged tomorrow, meaning that the fix will be available on the master branch. If you prefer to have it available on a stable release, you can wait 1-2 weeks (ETA for v4.4), although there shouldn’t be sensitive changes between the master (after merging PR2821) and the future v4.4.

Great - thanks a lot Julien!
Cheers,
Falk

Hi @falkeippert! Sorry it took longer than expected, but SCT 5.0.0 is now out, and the fix is in there :blush:

Thanks Julien, great that the new release is out! And no worries, I used the master branch until now :slight_smile:

1 Like