Sophie,
No you haven't really missed anything.
If the anatomic images are acquired exactly in the same locs as the functional then in general you are fine. As long as the movement is pretty minimal (less than a voxel) between you functional and anatomic scans then don't worry about it. However, some investigators will acquired a high res 3d image (spgr or flash etc) and typically that needs to be coregistered to the mean functional image produced from the realignment phase of the preprocessing. But if you anatomic are not close enough you can actually use the coregister button and choose the target as the mean functional and the object as the anatomic, correctly specifying acquisition type (epi and t1 (or t2)). You can do the coregistration only - no reslice is necessary. SPM will do the calculation to get the images in the same space (usually works 99% of the time). Basically this is what you are doing by hand, though SPM will also rotate volumes as need be, where you method only does translation. Again if you assess (using checkreg) visually that things are lined up then that is most likely fine.
Robert Welsh
-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Robert C. Welsh, PhD
Research Investigator
Department of Radiology
University of Michigan
(734) - 764 - 2412 (fax)
[log in to unmask]
>>> Sophie Lafaille <[log in to unmask]> 10/03/01 01:47PM >>>
coregistration? after realignment, I usually bring up the anatomic
image, find coordinates of AC and the enter the opposite mm into right,
forward and up to shift the anatomic (reorient image) so that the centre of
the picture is on the AC. I then normalize (R is R default) the a.img to
the a.img using the T1 template. I then bring up a raf*.img image. Our
Techs are very good about making sure the functionals are centred the same
as the anatomical (A&S), we acquire saggitally, and we use a homemade
program to calculate shift of the functional according to the image
(A.img&f*.img) parameters (L/R, A/P, S/I) and then include AC coordinates
to give final shifting parameters for the functional images (these are
applied to all 3 experiments) so that the
functional picture is shifted the same. I then apply the parameters
determined from the a.img to the f*.img.
Have I been missing a step all this time?
thanks sophie
|