Hi,

Sounds like you have a reasonable amount of coverage so you

should probably be OK.  Is your SPGR taken with the same slice

planes as the EPI?  If so, then I recommend treating the SPGR

as the whole_func image and it will hopefully work OK.

Otherwise, you will have to try to register directly from your EPI

example_func to your SPGR, including both translation and

rotation (6 dof).  Whether this works well or not will depend quite

a lot on the quality of the data, the initial alignment and the

exact structures covered or not covered.

The only other thing to do if the above doesn't work well is to

try an initial alignment (using pointflirt) and then refining the

registration from this starting point using the command line.

If you need to do this and get stuck, please email us again.

You might also want to look at  http://www.fmrib.ox.ac.uk/fslfaq/

which contains some information on schedule files and flirt

command line calls.

All the best,

        Mark

On Tuesday, February 17, 2004, at 02:00  am, Matthias Tabert wrote:

Hello,

I read the recommendations for Using FLIRT to register a few FMRI slices (http://www.fmrib.ox.ac.uk/fsl/flirt/ztrans.html):

I have EPI data with 17 (5mm) slices that cover most of the brain (cutting off some parietal cortex and the occipital lobe).  The idea was to maximize frontal and temporal cortex with the slice angle perpendicular to the long axis of the hippocampus.  I also have a whole brain SPGR image for each subject.  My problem is that I do not have the "whole_func" image (whole brain coverage).

I would appreciate any advice on how best to co-register these partial EPI images given the missing intermediate "whole_func" images.

Thanks, Matt

--

Matthias H. Tabert, Ph.D.

Assistant Professor of Clinical Psychology in Psychiatry

Department of Biological Psychiatry

Columbia University/NYSPI

1051 Riverside Drive, Unit 126

New York, NY 10032

Phone: 212-543-5046

Fax: 212-543-5854

--