Hi,
These commands look OK to me.
What does the image /kki/g1/jgalea1/eike_brain_reg.nii.gz look like
when it is overlayed onto the MNI152_T1_2mm_brain image?
Does it do this alignment correctly?
Also, you might need to use "-datatype float" in the final flirt call
in case it is rounding your mask values, although that would normally
look like a shrinking mask, not an incorrect displacement. You also
need to rethreshold a transformed mask in order to get binary 0/1
values again (rather than values between 0 and 1).
If you cannot work out what is going wrong then feel free to upload
your data to:
http://www.fmrib.ox.ac.uk/cgi-bin/upload.cgi
and send us the reference number.
All the best,
Mark
On 1 Mar 2009, at 19:26, JOSEPH GALEA wrote:
> Hi Mark,
>
> It seems that the transformation is making very little difference to
> the positions
> of the ROI. If you overlay both the transformed and untransformed
> ROI, they are pretty
> much identical. Also if I overlay the ROI onto the subject's brain,
> it is not
> within the brain, let alone on the anatomical areas I am looking for.
>
>
> Here are the exact commands I am using (within the GUI):
>
> I initially register the BET subject's structural to the MNI brain,
> with 12 parameter model.
>
> /usr/local/fsl/bin/flirt -in /kki/g1/jgalea1/eike_brain.nii.gz -ref /
> usr/local/fsl/data/standard/MNI152_T1_2mm_brain -out /kki/g1/jgalea1/
> eike_brain_reg.nii.gz -omat /kki/g1/jgalea1/eike_brain_reg.mat -bins
> 256 -cost corratio -searchrx -90 90 -searchry -90 90 -searchrz -90
> 90 -dof 12 -interp trilinear
>
> Then invert the transform
>
> /usr/local/fsl/bin/convert_xfm -omat /kki/g1/jgalea1/
> eike_brain_reg_inv.mat -inverse /kki/g1/jgalea1/eike_brain_reg.mat
>
> Apply inverted transform onto my ROI (ppc, this was created using
> the MNI brain), using the subjects head (with skull) as my reference
> volume
>
> /usr/local/fsl/bin/flirt -in /kki/g1/jgalea1/PPC.nii -applyxfm -
> init /kki/g1/jgalea1/eike_brain_reg_inv.mat -out /kki/g1/jgalea1/
> eike_PPC.nii -paddingsize 0.0 -interp trilinear -ref /kki/g1/jgalea1/
> eike_head.nii.gz
>
>
> thanks for any help you may be able to give me,
> Joe
>
>
>
>
>
>
>
>
>>>> Mark Jenkinson <[log in to unmask]> 03/01/09 3:43 AM >>>
> Hi,
>
> All of these sound fine in principle.
> I cannot help without more information.
> What is going wrong?
> How are you assessing the results?
> What are the exact commands that you are using?
>
> All the best,
> Mark
>
>
>
> On 28 Feb 2009, at 20:42, Joe Galea wrote:
>
>> On the basis of previous studies I have created 4 regions of
>> interest using
>> the MNI brain151_T1_2mm. I want to TMS these areas and so am
>> attempting to
>> transform these positions so that they are relative to each
>> individual
>> subject's high-res structural scan (with skull). For these subjects
>> I only
>> have a high-res structural.
>>
>> I have attempted several things:
>>
>> 1) Use FLIRT to register the MNI brain to the structural brain (with
>> skull).
>> Apply this transform to the ROI (with FLIRT transform)
>> 2) Use FLIRT to register the MNI brain to the structural brain
>> (without
>> skull). Apply this transform to the ROI, using the subject's
>> structural with
>> skull as the reference volume.
>> 3) Use FLIRT to register the structural (without skull) to the MNI.
>> Invert
>> this transform, and apply this to the ROI, using the structural with
>> skull
>> as the reference volume.
>> 4) Use FLIRT to register the structural (with skull) to the MNI.
>> Invert this
>> transform, and apply this to the ROI, using the structural with
>> skull as the
>> reference volume.
>>
>> None of these seem to make any difference to the position of my
>> ROIs, and
>> get me nowhere near the anatomical positions I'm looking for (within
>> the
>> parietal cortex).
>>
>> Any help would be much appreciated.
>>
>> Joe
>>
>
|