Hello,

and many thanks for taking time to check this registration problem.

I tried registration straight to highres as you suggested. It seemed
really good but when I tested by drawing some spots to anatomical landmarks and
running these spots through the same registration (with the same matrices)
the spots were misplaced in craniocaudal direction. The displacement was not
great (about 1 cm) but still so much that it is a problem since I am
trying to disclose a small anatomical area in my analysis.

I guess this happens easily when you try to register an axial epi T2 with
some distortion to sagittal T1 and have to use many degrees of freedom.
I would like to be able to control the craniocaudal displacement and slice
tilting by using the T1 FSE axial volume taken with the same slices as the
epi-volume as initial highres image and giving only 3 DOF in
epi2initial_highres phase but I cant get a good registration between
these two.

You mentioned that the voxel sizes (22 FOV, 64x64 matrix in epi, 22 FOV,
256x256 matrix in initial highres) do not match between epi and initial
highres. To get around this, I changed the voxel size in initial highres
to be the same as in epi. Now both volumes have 22cm FOV, 64x64 matrix and
28 slices.

When I tried to register these two with 3DOF, no search there was
only 26 slices in the output and the slices are misplaced so, that slice
13 in the output looks the same as slice 15 in initial highres etc.

I am puzzled about what happens here - to my understanding this should be
a really simple registration? I have put example images on this
(example_func.hdr&img, T1_64_mat.hdr&img and the registration output.hdr&img)
in

amiftp.hut.fi
lce/example_data/epi_T1_64mat

in case you would like to see how they look.

Best regards,
Johanna Pekkola
lce/TKK
Espoo, Finland


On Thu, 9 Oct 2003, Stephen Smith wrote:

> Hi, thanks for sending the data. The problem is that, with the 3DOF
> initial registration, your voxel sizes must be slightly out in one image
> or the other; the 3DOF translation-only registration cannot fix that.
>
> However, you actually have quite a good field-of-view in the fmri data, so
> you can just turn off the registration to "initial structural image"; I
> tried this and the remaining two registrations worked fine.
>
> Thanks, Steve.
>
>
>
> On Wed, 8 Oct 2003, Johanna Pekkola wrote:
>
> > Hello,
> >
> > and thanks for your help last time.
> >
> > Now, I would like to ask for your advice in regisration-related problems.
> > I have tried many things but cannot seem to get around this.
> >
> > For each subject (10) I have EPI-images (28 axial oblique slices with 3.4
> > mm slice thickness, T1 FSE images with exactly the same slice prescription
> > but better in-plane resolution and a sagittal 3D SPGR volume.
> >
> > In FEAT registration I have used the axial T1 images as initial highres
> > and 3D SPGR as highres. I run into problems in the first registration step
> > (epi to initial highres).
> >
> > With all DOF (3,6,7,9,12) the epi-images seem to get
> > tilted in respect to T1 so, that the frontal pole rises and occipital
> > areas are lowered. Also, even with 3DOF and no search the stack of
> > epi-images is misplaced in respect to T1 so, that there are 2-3 epi-slices
> > below the bottommost T1 slice.
> >
> > Using 7 DOF of more the cranial and caudal limits of epi- and T1 slice
> > stacks do correspond to each other quite well but the epi-images are magnified
> > so that most of the cortical surface ends up outside the T1 brain area.
> > Also, recognizable landmarks such as sylvian fissure are visibly
> > displaced between the two volumes.
> >
> > Am I correct to think that when the same slice presciption has been
> > used for epi and T1 the slices should not get misplaced in respect to
> > each other like this?
> >
> > Is there something you could recommend?
> >
> > With Best Regards,
> >
> > Johanna Pekkola
> > LCE/TKK
> > Espoo
> > Finland
> >
>
> Stephen M. Smith DPhil
> Associate Director, FMRIB and Analysis Research Coordinator
>
> Oxford University Centre for Functional MRI of the Brain
> John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
> +44 (0) 1865 222726 (fax 222717)
>
> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
>