Dear FSL experts,
I am attempting to use the probabilistic cerebellar atlas as described here (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/Atlases) and in more detail here (http://www.diedrichsenlab.org/imaging/propatlas.htm). I am running FSL (version 5.0.6) on an iMac (OSX 10.8.2 16GB 2.7 GHz Intel Core i5).
When I run FLIRT and overlay the cerebellar structures from the “Cerebellar Atlas in MNI152 space after normalization with FLIRT”, they appear as they should. I can see the %-likelihood that the x,y,z coordinates represent a region of interest. I can also save this ROI as a mask, threshold it, and binarize it without issue. While this works with FLIRT, I would prefer to use FNIRT and this is where the issue arises.
I first ran FLIRT, and then I ran FNIRT on the FLIRT output T1. No error messages were received and when I opened the FNIRT image in FSLVIEW, I confirmed that my T1 image was indeed in MNI space. At this point everything appeared to be fine.
Then I went to tools and atlases, and then selected “Cerebellar Atlas in MNI152 space after normalization with FNIRT”. I was able to select the atlas I wanted without issue, and with this atlas selected, I was able to get information on the %-likelihood that I was in a particular GM region (e.g. left dentate). I then went to overlay the left dendate structure onto my FNIRT transformed T1 image and I get a pixilated image over the T1. I tried this for several structures with the same result. This seems to cause FSLVIEW to crash as well. This pixilated image covers more than the cerebellum.
Here are the FLIRT and FNIRT parameters I used on the T1 image that was previously extracted from skull and neck tissue.
FLIRT:
I used default FLIRT settings. Reference image = MNI152_T1_1mm_brain.nii.gz. Model/DOF = Affine (12 parameter model) & number of secondary images = 0. Nothing else has been changed.
FNIRT:
Again, I used the default.
fnirt --in=ID#_brain_FLIRT.nii.gz --ref=MNI152_T1_1mm_brain.nii.gz --iout=ID#_brain_FNIRT.nii.gz
Any ideas why the atlas structures do not correctly overlay over the cerebellum after FNIRT? Is it possible that the atlas is corrupted?
Thank you in advance.
Best,
Victor
|