If your data are not already spatially normalised, then
the first thing that the segmentation in SPM does is to
attempt a 12-parameter affine registration to one of the
template images. This is so that it can overlay the
prior probability images of GM, WM and CSF. If this stage
fails, then the segmentation can go horribly wrong.
You could test this by spatially normalising the problematic
images (after disabling the nonlinear warping). If this is
the problem, then try first reorienting the images via the
Display button, so that they are transverse, with the
origin near the AC. Once the spatial normalisation is
working, then there should be fewer problems with the
segmentation.
Good luck,
-John
| I use SPM'99 to segment 3-d gradient-echo datasets.(TR
| = 19ms, TE = 4,4 ms, flip = 8, FOV = 23 cm, thickness
| 3mm, gap = 0, slices = 60). I get good results, when I
| use the PD-weigthed template. Usually the segmentation
| is almost perfect. Sometimes (on some heads) the
| segmented areas are much too small, especialy in upper
| regions. I tried quite a lot to find the reason for
| that but I can't find a systematic. I have a couple of
| ideas: Does the ratio between FOV and head size have
| any influence on the segmentation ? How does the
| position of the head (i.e. rigth-left) influence the
| quality of the segmentation ? The results are better
| if the first 5 slices(head first) of the dataset are
| above the skull, but this is difficult to implement
| into clinical routine.
| Does anybody have an idea ?
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|