Hi @brai
Sorry, we’ve put this analysis on hold, and moved forward with other projects. Hopefully, what @oesteban’s suggested helps you 
Good luck!
@oesteban
Again checking for updated recommendations (MP2RAGE in 7T).
or the best way currently is to add --skull-strip-t1w skip?
Dear all,
I have not tried to use it as input for fMRIPrep yet, but a simple multiplication (e.g. with fslmaths) of inv2 and T1w from MP2RAGE already leads to images that look very similar to MPRAGE. Any thoughts why this could be a bad idea as long as the images are only used for spatial processing of fMRI data with tools not capable of directly using MP2RAGE (and not for tasks like actual morphometry)?
Hi I think this strategy is valid for both morphometry and spatial processing of fMRI data.
I recently came across this project that does exactly that: preparing mp2rage for freesurfer, doing so by multiplying UNI by INV2 image.
Otherwise, the more “traditional” way to deal with MP2RAGE UNI data is to use the Robust background removal method introduced by O’Brien et al (2014), that you can find in different flavors:
- matlab original scripts: https://github.com/JosePMarques/MP2RAGE-related-scripts/blob/master/DemoRemoveBackgroundNoise.m
- SPM flavored: https://github.com/benoitberanger/mp2rage
- python flavored: https://github.com/khanlab/mp2rage_genUniDen/blob/master/mp2rage_genUniDen.py
You can find other options, there is at least another one cited in this thread.
2 Likes
Hi all, I am encountering the same issues with MP2RAGE anatomical images and would like to use them in fMRIprep for surface reconstructions and all further processing. Is there any update on this issue with regard to an implementation of MP2RAGE processing in fMRIprep? Or do people still follow any of the suggestions above (presurfer, 3dMPRAGEise, O’Brien 2014, feeding skull-stripped images to fMRIprep) and see which works for their specific use-case? Thank you for any responses!