Dear Sue,
Your image is like I thought it would be - not like a brain scan
you'd see from a scanner (which is what the MNI image looks like)
so you can't expect to be able to register it. However, the good
news is that your original image is already in standard space!
It just has a different Field Of View. So to change resolution or
FOV you just need to resample the image. To do this you use
the -applyxfm option as well as the -usesqform option (as this
uses the information about how the coordinates relate to
standard space). So you'd use the command:
flirt -in brodmann -ref $FSLDIR/data/standard/MNI152_T1_2mm -applyxfm -usesqform -out brodmann_2mm -interp nearestneighbour
You can replace the -ref image with whatever standard
space image you want.
You can also use this command with any image that is in
the same space as brodmann, provided it has the same
sform matrix (as seen by fslhd).
All the best,
Mark
On 23 Feb 2011, at 19:21, Sudhin A. Shah wrote:
> Hi,
>
> the ref number is 709547
> -brodmann.nii.gz (original template from MRIcron)
> -brodmann.reg2mni.nii.gz (registered using flirt)
> -bas.nii.gz (seeds extracted from original)
> -bas2.nii.gz (same seed extracted from registered template)
>
> I am using a modified (multiple BA's, both L and R) version of this batch script -
http://laconic.aecom.yu.edu/wiki/index.php5/SITP_001:_Pull_individual_Brodmann_areas_from_the_MRIcron_Brodmann_dataset>
> The wiki owner suggested I use flirt (exact same command he used) to get the Brodmann MNI template that is referred to in the bottom of the page.
>
> Thanks,
> S
>
> On 02/23/11, Mark Jenkinson <
[log in to unmask]> wrote:
>> Dear Sue,
>>
>> You can upload your original image and registered result to:
>>
http://www.fmrib.ox.ac.uk/cgi-bin/upload.cgi>> and send me the reference number.
>>
>> I'm not sure that I will agree that it "looks like a brain" though,
>> as a normal brain image does not have identifiable brodmann
>> areas given by numbers! Is there an image that this is already
>> registered with that which is more like a brain? If you did have
>> something like this then also send it to the upload above.
>>
>> All the best,
>> Mark
>>
>>
>> On 23 Feb 2011, at 17:51, Sudhin A. Shah wrote:
>>
>> > Hi, Yes it looks like a brain, with ~37 different areas, all with a unique number. i.e. Area 1 has voxel values of 1 etc.. When I extract a seed from the original template, I can get just one chunk of like valued voxels. After registration to MNI space, it still looks like a brain, but the boundaries are all weird. When I then try to extract an area, I get a weird chunk and tiny voxels outside the area (almost like an outline of the brain). Is there an email address I can send attachments to? I could send you the template and seeds..
>> >
>> > Thanks,
>> > S
>> >
>> > On 02/23/11, Mark Jenkinson <
[log in to unmask]> wrote:
>> >> Dear Sue,
>> >>
>> >> I don't know this image - does it look like a brain or is it an
>> >> image with just constant label values in it (a collection of
>> >> ROIs)? If it is the latter then you cannot expect it to register.
>> >> You can only register two things together than contain
>> >> similar information (i.e. tissue boundaries, etc.)
>> >>
>> >> If it does look like a brain then it should be possible to
>> >> register and you'll need to give me more information about
>> >> in what ways it is failing. I couldn't understand your sentence
>> >> about blurry seeds.
>> >>
>> >> The option -odt short is for fslmaths, not flirt, so I think
>> >> you are getting confused there.
>> >>
>> >> All the best,
>> >> Mark
>> >>
>> >>
>> >>
>> >> On 23 Feb 2011, at 17:14, Sudhin A. Shah wrote:
>> >>
>> >> > Hello,
>> >> >
>> >> > I have a problem using flirt to register a template (brodmann.nii.gz from MRICRON) to MNI space.
>> >> >
>> >> > This is the command I used:
>> >> > flirt –in brodmann.nii.gz \
>> >> >
>> >> > –ref $FSLDIR/data/standard/MNI152_T1_2mm_brain \
>> >> >
>> >> > –out brodmann.reg2mni \
>> >> >
>> >> > –omat brodmann.reg2mni.mat \
>> >> >
>> >> > –bins 256 –cost corratio \
>> >> >
>> >> > –searchrx -90 90 \
>> >> >
>> >> > –searchry -90 90 \
>> >> >
>> >> > –searchrz -90 90 \
>> >> >
>> >> > –dof 12 \
>> >> >
>> >> > –interp trilinear
>> >> >
>> >> >
>> >> > Seeds generated with the second file are blurry and contain voxels outside the target areas.
>> >> >
>> >> >
>> >> > I tried adding -odt short to the command line but got an error Unrecognised option. I am using FSL in Ubuntu/Linux.
>> >> >
>> >> >
>> >> > Can you please help/
>> >> >
>> >> >
>> >> > Thanks,
>> >> >
>> >> > Sue
>> >> >
>> >> >
>> >> >
>> >
>> > --
>
