Dear all,
I would like to use the ArtRepair toolbox to correct for motion artifacts, and I'm a bit unsure about the capabilities/limitations of the toolbox. So, I was hoping you might be able to give me some advice.

For example, I have a participant that has a large jump in rotation (up to -5 degrees), but otherwise the motion parameters are within the acceptable range. I would like to use the toolbox to correct the bad scans using interpolation between the closest non-outlier volumes (art_global function). Is this possible with very strong jumps or severe motion? Is there a threshold (for example, up to 10 mm/degrees) above which scans cannot be corrected? The instructions only state that all subjects for whom more than 20% of the time series had to be corrected should be excluded.
Here some basic information regarding my procedure: I repaired the realigned images using the ArtRepair Toolbox. Afterwards, I did the remaining preprocessing steps (normalization and smoothing) with the repaired images and performed the model estimation (Scans were deweighted during SPM estimation).
Using the art_summary function, I compared the data before and after the repair: Repairs reduce Std of estimation error by 10.5% and Average ResMS error by 14%.

Nevertheless, I wanted to take another look at the motion parameters to see if the motion improved after the repair, since my criteria for exclusion is anything over 3 mm or 3 degrees. To get the new realignment parameters, I repeated the realignment step with the repaired images. For this purpose, I took the repaired v-images that are not yet normalized and smoothed. In fact, the movement is much better between +-0.1 degrees. Now I wanted to ask if it would make sense to re-evaluate the motion with the repaired images?
Can I refer to the renewed realignment parameters and, thus, include the subject? Can I do this even if he should be excluded due to the original motion parameters?

I would be very grateful if you could help me out.
Thanks in advance.

Best,

Nadja