Dear Peter,
> I just tried to coregister a 10slice T1-image (object image)
> to either a mean image of fMRI-time-course T2* (target image)
> or the first image of the fMRI-time-course T2* (target image).
> Unfortunately I only get results that look extremely unlikely,
> despite the fact that both images have the same number of
> slices and were acquired with the same field of view.
>
> Please help, do I need to preprocess my data ?
>
> Peter Erhard
>
> -----------------------------------------
>
> Peter Erhard, Ph.D.
> Nuklearmedizinische Klinik und Poliklinik
> Klinikum rechts der Isar
> Technische Universitaet Muenchen
> Ismaninger Strasse 22
> 81675 Muenchen
> Germany
>
> Tel: 49-89-4140 4570
> Fax: 49-89-4140 4938
> [log in to unmask]
> [log in to unmask]
>
generally, the coregistration of few slices of different modalities
seems to be quite difficult, since there is only few information about
how to perform the coregistration. Therefore I would try to find out,
whether it's possible for future scans to acquire the structural and
functional images such that you can assume that they are in the same
anatomical space (after correction of the voxel sizes). The assumption
is here of course that the subject didn't move between the acquistion of
structural and functional scans or better that the pause between the
sequences wasn't too long.
Given that you cannot assume that the data is in the same space after
scanning, I would guess that you can maybe improve your results of the
SPM96-coregistration by checking, whether the ORIGIN field of the images
are set correctly. If you find e.g. only zeros there, try to define the
(rough) anatomical location of the commissura anterior relative to your
images and specify the origin.
The reason for doing this is that the coregistration step includes a
rough affine normalization of both images to its templates so that your
specification of the ORIGINs gives additional information about the data
to the coregistration routine and you may prevent that your registration
algoritm results in a local minima solution.
Best wishes, Christian Gaser
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|