 |
 |
Dear Bedda, > > Thank you for your help and suggestions. Skull stripping was done in MIPAV and they were checked for quality. > > I know that it is recommended to use non skull-stripped for the fnirt step. > > I was going to use non skull-stripped T1 images, but the problem is that the skull-stripped and non skull-stripped images are in two different spaces and I need to apply a mask that it is in the same space as the T1 skull-stripped images. if what you want is a good registration to standard space and you have “good” skull stripped data, then that should be fine. I would even say that if one has “perfect” skull stripped data that would be the best to use for non-linear registration. Our recommendation of using the non-skull-stripped images is a concern that often skull stripping isn’t 100% perfect. So, if you trust your skull stripping, just go for it. Jesper > > Should I do the following > > flirt t1 skull-sripped to MNI > fnirt t1 non skull-stripped to MNI > applywarp > > flirt t1 non skull-stripped to MNI > fnirt t1 non skull-stripped to MNI > applywarp > > so I can have both images in MNI space? > > Thanks again, > Bedda > > > > On Wed, May 21, 2014 at 5:24 AM, Jesper Andersson <[log in to unmask]> wrote: > Hi Bedda, > > has the skull stripping worked well (i.e. has it not removed any brain, or left any non-brain, anywhere)? > > If so do the following changes > > # name of reference image > --ref=MNI152_T1_2mm > change to > —ref=MNI152_T1_2mm_brain > > # name of file with mask in reference space > --refmask=MNI152_T1_2mm_brain_mask_dil > comment out > > # If =1, use implicit masking based on value in --ref image. Default =1 > --imprefm=1 > change to > —imprefm=0 > > # If =1, use implicit masking based on value in --in image, Default =1 > --impinm=1 > change to > —impinm=0 > > > If the brain stripping hasn’t worked perfectly (or close to perfectly) you should only do the first of those changes. But in that case I am afraid that your results won’t be as good as they could. There is a reason we recommend using the non-skull-stripped images for fnirt. > > Good luck Jesper > > > > On 20 May 2014, at 22:37, Bedda Rosario <[log in to unmask]> wrote: > >> Hello, >> >> I am trying to apply fnirt to register my T1 structural image to the MNI template (after flirt). However, my structural images are skull stripped. Is it possible to create another config file changing the reference image in the config file to the skull stripped T1_MNI152_2mm_brain? >> >> I would appreciate any help. >> >> Thank you, >> Bedda > >