Hi,
Is your input image brain extracted or not? FNIRT works best with non-brain extracted input images (unlike FLIRT) and so that might help. We certainly would not consider this result the best that you could get.
All the best,
Mark
> On 23 Mar 2017, at 17:33, Ishtiaq Mawla <[log in to unmask]> wrote:
>
> Hi FSL experts,
>
> I'm having some doubts with FNIRT usage...
>
> non-linear T1 to MNI152 transformation using FNIRT looks good when you go to structures deeper than the cortex, e.g., corpus callosum, thalamus, ventricles, etc.
>
> But I'm not super happy with the outcome in the sulci/gyri of the cortex. See first slide in attached pdf. Top image in T1 to MNI and the second image is the MNI152 image. There seems to be a "dent" in the parietal region, as you can see.
>
> This "dent" propagates into my functional image, when I apply FNIRT matrix on the functional (see second slide) and as a result, it seems that I have missing voxels in that region when I place the MNI image as my underlay.
>
> Here's the FNIRT command I used for T1 to MNI:
> fnirt --ref=$MNI2mm --in=T1-2mm_orientOK.nii.gz --aff=$FreeSurferDir/$subj/mri/reg/T1_lin_MNI152.mat --cout=$FreeSurferDir/$subj/mri/reg/T1_nl_MNI152.reg --config=T1_2_MNI152_2mm.cnf --iout=T1_nl_MNI152.nii.gz
>
> Here's the FNIRT command I used for functional to MNI (after FLIRT'ing functional to T1 first):
> applywarp --ref=$MNI2mm --in=FUNCTIONAL.nii.gz --warp=$FreeSurferDir/$subj/mri/reg/T1_nl_MNI152.reg --out=FUNCTIONAL_nl_MNI152.nii.gz
>
> Any suggestions on what I can tweak to improve FNIRT outcomes? Or is it as good it gets? Suggestions appreciated!
>
> Thanks,
> Ishtiaq
> <FNIRTproblemFSL.pdf>
|