Hi mark
i just looked at the sizes of the structural and transformed image and they are exactly the same.
dim1 512
dim2 512
dim3 120
dim4  1
pixdim1 0.468800
pixdim2 0.468800
pixdim3 1.299999
pixdim4 1.00000

thanks
rajani


On Tue, Mar 24, 2009 at 10:51 AM, Mark Jenkinson <[log in to unmask]> wrote:
Hi,

If you can't overlay the results then the transformation has not worked.
Look at the sizes of the images with fslsize for the structural image
and the transformed statistical image.  They should be the same.
If the structural image was used as the reference image then the
output of the applyxfm will definitely be that size, so I'm guessing
that this hasn't been done correctly.

All the best,
  Mark


rajani sebastian wrote:
The transformation worked fine but now i have trouble viewing the image on FSL view. I want to overlay the transformed statistics image on the subjects high resolution image, but when i do that i get this error message " unable to load incompatible overlay. all overlays must have same dimensions as the base image".
I am not sure why this happens.
thanks
rajani

On Tue, Mar 24, 2009 at 8:34 AM, rajani sebastian <[log in to unmask] <mailto:[log in to unmask]>> wrote:

   thanks!!!


   On Tue, Mar 24, 2009 at 4:00 AM, Mark Jenkinson
   <[log in to unmask] <mailto:[log in to unmask]>> wrote:

       Hi,

       If you are using the same highres image in each case then
       there should
       be no difference between the standard2highres.mat files.

       If you want to transform your statistics, then the statistic
       images are the
       input images.  If you've done cluster correction it is
       probably best to
       transform the thresholded statistics rather than the
       unthresholded ones.
       Do not transform the rendered ones though - stick to the
       thresholded
       stats.

       As for the reference volume - it needs to be the highres image
       if that
       is the space that you want the statistics to end up in.

       All the best,
              Mark




       On 23 Mar 2009, at 19:18, rajani sebastian wrote:

           Hello all

           My experiment consists of three runs of the same task in
           patients with brain lesion. Initially, i analyzed data
           from all the three runs separately and then used higher
           level analysis using fixed effects to combine the three
           runs. Now i want to register each patient's higher level
           statistic map back into the patients native space (high
           resolution) from standard space using Apply XFM in fsl.
           I have three questions regarding this:

           1. The input is the transformation matrix that is found
           inside the feat director . My question is which
           transformation matrix "standard2highres.mat' should i use.
           Does it really matter if i used "standard2highres.mat'
           from run 1 rather than ran 2.

           2. What should the input volume be?  I wanted to be able
           to register the higher level statistics map back to the
           subjects native space. So should this be the
           'threshold.nii.gz' file from the gfeat.directory.

           3. What should the reference volume be? should this be the
           high resolution structural image ??.

           Any help would be appreciated.

           thanks
           rajani