Another possibility is that you coregistered the anatomical scans to
match the functional data (rather than the functional to match the
structural). If this was done after segmenting, when the the imported
rc1 and rc2 images were created, then things will mess up. This is
because during coregistration, the alignment between any imported
scans and anatomical scans would be lost.
Best regards,
-John
On 24 July 2012 22:35, soha saleh <[log in to unmask]> wrote:
> Dear John and all:
>
> I used the latest spm8 release and repeated the same steps. Unfortunately I
> got the same problem (see the figure attached). In the figure I labeled the
> template, flow field, and a functional scan before and after normalizing.
>
> Do you have any suggestion of what would be the problem?
>
> Mostly I used default values, for voxel sizes I used [3 3 3]. In
> segmentation, I used medium regularization.
>
> Thanks,
>
> Soha
>
>
>
> This time I used the latest SPM8 patches.
>
>> Date: Tue, 24 Jul 2012 16:25:52 +0100
>> From: [log in to unmask]
>
>> Subject: Re: [SPM] DARTEL - Normalize to MNI space problem
>> To: [log in to unmask]
>
>>
>> I assume that the registration was run using the GM and WM segmented
>> from anatomical scans. If you want to apply transformations estimated
>> from anatomical scans to fMRI data of the same subjects, then (for
>> each subject) the fMRI data should be coregistered so that they align
>> with the anatomical scan. You can assess this using Check reg.
>>
>> Also, ensure that you use the latest patches for SPM8. The original
>> version of normalise to MNI space had a few problems - although these
>> were soon fixed.
>>
>> Best regards,
>> -John
>>
>>
>> On 24 July 2012 16:12, soha saleh <[log in to unmask]> wrote:
>> > Hello:
>> >
>> > Recently I started using DARTEL for preprocessing scans of subjects with
>> > lesions due to stroke. So I do not have much experience in it.
>> > I used the segment tool and a map of the lesion created in mricro as a
>> > mask.
>> >
>> > The templates were created with no problem but I have a problem with
>> > normalizing functional scans to the template.
>> > The sw.* images are getting warped abnormally. The figures might give
>> > you
>> > an idea what I mean.
>> >
>> > I tried to coregister and align the functional scans to the template
>> > before
>> > normalizing them but that did not help (see the figure in the second
>> > file).
>> >
>> > I appreciate if anybody has a suggestion or a solution to such problem.
>> >
>> > Thank you!
>> >
>> > Regards,
>> >
>> > Soha
>> >
|