Hi Cesare

I presume from your posting that you are using the MELODIC GUI or following their default pre-processing pipeline? If you are not getting the results you expect from the normalization within the FSL-MELODIC pipeline, you will need to do normalization "manually" with the flirt (or perhaps fnirt may give you better results) commands.

I have added a helpful link regarding registration problems. Although I don't see any obvious issues with your data from the screenshot provided, I would double check the results of your BET and also that you do not see a dramatic change in global intensity across the entire scan (if your first few volumes are much brighter than the rest, you may try using a later volume as the reference image). Otherwise, you may have to alter your cost function and/or DOF (helpful info on the same link provided). Also, I might consider trying fnirt or boundary-based registration (BBR), although BBR is a bit more tedious and may not be of interest to you.

If these simple fixes do not work, then feel free to come back.

http://web.mit.edu/fsl_v5.0.8/fsl/doc/wiki/FLIRT(2f)FAQ.html

Hello Erik,

thanks for your feedback. Let me give it a try and I'll report back you as soon as I'll have some updates.

Hey Erik,

I have come across the same problem. Can you remember how you solved it and help me fix my registration?

Warm regards

Renée