<maskunique_overlay.png>Dear fsl experts,I am running a group analysis in MELODIC and I have a problem with registration of my functional images.The sum of all masks after registration to MNI (the mask image in the .gica folder) appears larger than the standard MNI. I attached a picture of maskunique- we could see that the dashed-colored line does not overlap the standard image. I am therefore getting activations outside the brain. But I can't seem to be able to specify a single bad functional image. I used robust BET to extract non brain areas of the functional images. I then preprocessed them in FEAT, denoised them using fsl_regfilt and inputted the denoised images in melodic where I registered them to T1 and MNI and ran group ICA.Is the problem likely to be in the brain extraction procedure? What could explain this 'halo' in maskunique that is larger than the MNI image?Your input is much appreciatedThank you!!Helen