Motion correction and spatial normalization are two different parts of
the pre-processing. I assume you wish to skip the spatial normalisation.
If you do this, then the results you obtain do not superimpose well on the
"glass brain", and if the position of the images is not well specified, then
many of your blobs will not be visible there at all. If you wish to omit
spatial normalisation for a single subject analysis, then you would first
reposition the imagesusing the <Display> button. By specifying various
translations, and possibly some rotations, you can roughly get the image
in the same orientation and position as the glas brain. Once you have the
correct position, then click the <Reorient images..." button, and a set of
.mat files will be created that include information on the image position.
The <Check Reg> button will allow you to compare the relative positions of
the images with the various template images supplied with SPM99.
Although subsequent analysis of the data will not show the blobs in exactly
the correct position in the glass brain, at least the positions will not be
too far wrong. You can more accurately locate your blobs by superimposing
them on orthogonal or transverse sections of your images.
All the best,
-John
| I have one event-related fMRI run that I want to analyze, without doing
| motion correction. What do I need to do to skip the normalization step in
| SPM99? SPM99b's "spm_fmri.man" (which doesn't seem to exist for the
| post-beta release, SPM99) says: "It is not necessary to spatially normalise
| the data ... simply specify where you think the anterior commisure is, with
| the ORIGIN in the header of the first scan, by using the 'Display' facility
| and proceed directly to smoothing." The last bit is a little too vague for
| me to follow.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|