Hi Amit,
See inline replies below.
Donna
On 08/05/2008 01:20 AM, Amit Etkin wrote:
Hi all,
I'm trying to map my activation map (spm_T's written out as .img and
.hdr pairs) from SPM5 onto a surface in Caret, and getting more
confused the more I do. I've gone through the tutorial, can map
activation files and can even get my activation file mapped, but have
a number of questions about how I should go about doing that below:
As far as I know, Caret still can't read .img/.hdr NIfTI pairs, so write
your spm_T's out in .nii form for feeding to Caret.
1. Which spec file should I use? If I want to do the MFM approach,
then it seems I need to select the Human.PALS_B12.LR.MULTI-
FIDUCIAL_SPM2_fMRI-MAPPER.B1-12.LEFT.73730 file, for example, but it
doesn't seem to have surface with the 12 subjects averaged. If I want
to get the averaged surface, in which case I can only do AFM
registration, I can chose
Human.PALS_B12.LR.AVERAGE-FIDUCIAL_SPM96.LEFT.73730, but then can't do
MFM. I'd like to be able to do MFM but display it on the averaged
surface...is that possible? If so, how?
When I select the SPM2 space and select
Human.PALS_B12.LR.MULTI-FIDUCIAL_SPM2_fMRI-MAPPER.B1-12.LEFT.73730, I
get a window like the attached capture. Note that there is a
distinction between the surface used for mapping and the surface used
for display. Using the default selections in the attached capture,
Caret will map not only to the average SPM2 left surface, but also to
each of the contributing PALS left hems in SPM2 space, and then average
the resulting 12 individual maps.
Either of these maps (your_vol_on_AFM or your_vol_on_MFM) can be
displayed on the SPM2 average fiducial surface. But we usually don't
display fMRI results on average fiducial surfaces. We usually use the
inflated, sometimes very inflated (for opercular results) or flat
surfaces. We do use average fiducial surfaces generated for
study-specific purposes (e.g., to illustrate morphometric differences).
But it can be a little misleading to display results the PALS average
fiducial, because the audience might be tempted to make the implicit
assumption that your subjects' anatomical features exactly match PALS.
While on a broad level this will be true, on a micro level it won't be.
2. When I open a spec file it asks for the space...should I pick what
it defaults to or change it to SPM5?
This reminds me: Tim Coalson has generated SPM5 surfaces which we need
to package and release. The only loose end is the cerebellar surface.
I must coordinate with David (and possibly Joern Diedricksen) on that.
Yes, Jay Hegde, I did not forget you; SPM5 surfaces are on the way VERY
SOON.
For now, select SPM2 (the closest space).
3. When I map functional volumes to surfaces, I'm only given the
option of selecting between .hdr not .img files, and then get the
warning: "You have selected a ".hdr" file. If this ".hdr" file does
not contain an SPM originator, it is unlikely that it will get mapped
to the surface correctly". What does this error mean, and how do I
know if the hdr contains an SPM originator?
Use SPM5 to write your spm_T's as .nii to resolve this problem.
4. Do I need to flip along the X axis or change the location of the
origin of my functional images? If so, how?
If you write as .nii, Caret will flip as needed.
5. Where do I get flat maps, inflated surfaces and brodman area maps
like there are in the tutorial?
Here is the Sept 2006 tutorial dataset:
http://sumsdb.wustl.edu/sums/archivelist.do?archive_id=6595030
Here is the Sept 2006 tutorial document:
http://sumsdb.wustl.edu/sums/archivelist.do?archive_id=6602379
Click on the diskette shaped icon to download them.
Don't be intimidated by the size of the tutorial; skip to Part 5,
section 5.1.
that's all the questions I have for now, and would GREATLY appreciate
all the help I can get.
many thanks,
Amit
--
Amit Etkin, MD, PhD
Department of Psychiatry
Stanford University School of Medicine
401 Quarry Road
Stanford, CA 94305
office: 650-723-2516
------------------------------------------------------------------------
_______________________________________________
caret-users mailing list
[email protected]
http://brainvis.wustl.edu/mailman/listinfo/caret-users
