I'd suggest taking a look at this for some suggestions:
http://www.fil.ion.ucl.ac.uk/~john/misc/VBMclass10.pdf
I got a bit lost with your first step, which should have produced
rc1*.nii, rc2*.nii, c1*.nii and maybe other data depending on what you
intend to do.
The next step aligns all the rc1*.nii together while also aligning all
the rc2*.nii together, generating u_rc1*.nii files (each encoding the
shapes relative to the group template) and a series of Template_*.nii
files.
The third (and usually final) step is to use the u_rc1*.nii to
spatially normalise the c1*.nii files from the first step. In order
to ensure that the whole dataset is aligned with MNI space, an
additional affine registration is included that aligns the
Template_6.nii file with MNI space. Images would be usually be
written out Jacobian scaled and smoothed. These should be smwc1*.nii
files, which may be used in subsequent VBM analyses.
Best regards,
-John
On 27 October 2010 11:05, Lars Meyer <[log in to unmask]> wrote:
> dear all,
>
> following the suggestion in a post from jonathan peelle on tuesday, 27 jan 2009 20:37:27 +0000, i have set up my vbm pipeline like this:
>
> (i. get minimal bounding box of images if one is bigger than the others; two different 3t scanners were used)
>
> 1. new segment, writing out the deformation fields
> input: t1 images
> output: 'y_' files, 'c1' to 'cn' files
>
> 2. run dartel on grey matter volumes
> input: 'c1' files = grey matter volumen
> output: 'temp_' files = templates, 'u_c1*_temp_1' files = flow fields
>
> 3. normalize to mni space
> input: 'u_c*temp_1' files = flow fields, 'c1' files = grey matter volumes, 'temp_1_6' file = final dartel template
> output: 'smwc1' files (not used any further?)
>
> 4. combine mni-normalized dartel template and individual flow maps to achieve mni-registered warp
> input: 'c1' files = data, 'y_S' files = deformation fields, 'u_c*temp_1' files = flow fields
> output: 'wc1' files
>
> 5. smooth
> input: 'wc1' files
> output: 'swc1' files
>
> my problem is that now the resulting images are not aligned to mni space, i.e. the transformation to mni space seems not to be performed. what is the matter with the pipeline?
>
>
> thanks a lot in advance,
> lars
>
|