Hi Brad,
did you run FNIRT on the betted data? Unlike FLIRT, which should be
run on brain extracted images, FNIRT should be used on the original
data with skull and scalp preserved. See the examples at http://www.fmrib.ox.ac.uk/fsl/fnirt/index.html#fnirt_examples
hth
Christian
On 14 Aug 2008, at 20:07, Bradley Buchsbaum wrote:
> Hi,
>
> Call me a little overeager, but I've already been testing out FNIRT.
>
> fnirt --in=anat.nii --ref=MNI152_T1_2mm_brain.nii.gz
> --refmask=${FSLDIR}/data/standard/MNI152_T1_2mm_brain_mask_dil.nii.gz
> --aff=flirt_anat_to_MNI.mat --cout=fnirt_anat_to_MNI_transf
> --config=T1_2_MNI152_2mm
>
> The result looks quite good, except for a noticeable "indentation" in
> the normalized result (see cross hair in attached images). The edge of
> the warped brain in one location in parietal lobe
> is approximately 1cm inside the edge of MNI template. In contrast,
> the edge using 12 dof flirt normalization on the same brain is quite
> close to the MNI edge in this location (although other aspects of the
> alignment, are not as good as with FNIRT).
>
> I tried adjusting "warpres" both up (12,12,12) and down (7,7,7) but
> got similar results. Is there an FNIRT parameter that can be tweaked
> to improve edge alignment in these kind of cases?
>
> thanks,
>
> Brad Buchsbaum
> <slice50_fnirt.png><slice50_MNI.png>
|