In spm_transverse.m, the z co-ordinate is worked out on the assumption
that the data has been spatially normalised. I have now fixed this:
ftp://ftp.fil.ion.ucl.ac.uk/spm/spm99_updates/spm_transverse.m
Note that the slices are displayed according to the actual orientation
of the functional images (i.e. if the images were coronal, then they
are displayed coronally - even if the .mat file says otherwise). If
the data are not spatially normalised, then the z co-ordinate that
is displayed only refers to the z co-ordinate of at the crosshair. If
the slices are not purely transverse, then a different z co-ordinate
would be expected at different parts of an image.
Best regards,
-John
| I think there is something basic we do not understand. Let us consider a
single
| subject experiment. We have fmri images of an experiment with visual
| stimulation. By using the DISPLAY tool we note that the images are rotated
| up-down so we rotate all the functional and structural t1 images of pitch
angle
| of 3.14. Then we make the realignment and smoothing steps. We build up the
| design matrix, solve the linear system and perform the statistical analysis.
| When we have the activations, we would like to superimpose them on the
| structural t1 image. What we think is that in this situation we do not need
the
| normalization step. We performed the coregistering step between the functional
| images and the t1 image but we have the same inversion problem when we use the
| overlay with slices (the z-values are inverted). Could you please help us in
| finding our mistake? Another question: should the coregistering operation come
| before the realignment one or after?
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|