Hi Dr. Ramon,
Your volume type is correct provided your SPM2-normalized functional
volume was is in neurological orientation (left-on-left); see
http://brainmap.wustl.edu/help/mapper.html#spm2flip for details.
Otherwise, you should choose SPM2 or MEDx in RPI orientation. You chose
the correct target surface. If your volume is left-on-left, then please
upload it here, so I can have a look:
http://pulvinar.wustl.edu/cgi-bin/upload.cgi
If you have a screen capture showing where SPM2 shows the loci, then
that would also be helpful. Otherwise, a description of where they ought
to be might suffice.
Regarding the threshold, I somewhat dread these questions, because I
know my answers are inadequate. I don't understand how popular analysis
tools like SPM and AFNI transform values stored in functional volumes
based on the type of statistic they contain, to make life easier for
users. I know AFNI does this, and perhaps SPM does as well. Caret
thresholds strictly on the value in the currently selected column of the
metric file. If the enclosing voxel or average voxel with neighbors=0
was used as the mapping algorithm, then the metric value reflects the
intensity of the voxel in which the surface node is located. (Caret's
thresholding features were largely driven by wustl.edu users, who tend
to feed Caret pre-thresholded volumes of group results.)
If patient and articulate SPM and/or AFNI users are willing to help us
understand how we can make caret work better for their needs, then John
Harwell probably could enhance caret's threshold handling.
Donna
On 11/22/2004 08:03 AM, Meike Ramon wrote:
Hello,
I have a question considering the threshold settings while mapping
functional data on a fiducial atlas. The data I've imported from SPM2 have a
fwe corrected p-value of 0.05. What is the corresponding threshold in caret?
Furthermore the loci (e.g. global maxima in SPM2)are not analogous (to those
displaying maximum activation) in Caret. While mapping fmri to surface I
chose SPM99, SPM2 or MEDx in LPI Orientation (.hdr) as the volume type and
selected the Human.cloin in SPM2/MEDx Template Space Atlas. Was this
incorrect? If not, what is the differnce in activation location due to?
Thanks in advance for any information!
Meike Ramon
Hi Dr. Ramon,
First, I should mention that we soon plan to release a new version of
Caret that not only revises the entire mapping interface, but also adds
support for the PALS_B12 (population-average landmark and surface based,
Buckner 12 individuals) atlas. The software and data are ready, but
David is sorting out some scientific issues.
Meanwhile, I think the easiest way to map group data to the atlas is for
you to use the map_fmri_to_surface program. It might be available as an
option on your Caret Attributes: Metric menu. But for some Linux
platforms, the mapper hangs when it is launched this way. You can
launch the mapper by entering map_fmri_to_surface at the command line,
and it shouldn't hang.
Then, the steps outlined in the Caret4.6 User's Guide and Tutorial Part
I
(http://brainmap.wustl.edu/caret/pdf/CARET_UsersGuide.03-06.Part-I.pdf),
page 36, should guide you through the process. Note that there are two
SPM volume types from which you can choose: SPM99, SPM2 in LPI
orientation or SPM2 in RPI orientation. If your SPM2-normalized output
is in radiological convention (i.e., analyze.flip=1 in spm_defaults.m),
then choose SPM2 in RPI orientation; otherwise, choose SPM99 or SPM2 in
LPI orientation. See http://brainmap.wustl.edu/help/mapper.html#spm for
other SPM mapping details.
We will announce the release of the new Caret version on caret-announce,
once David has sorted out the remaining issues.
Donna
On 11/18/2004 07:13 AM, Meike Ramon wrote:
Hello Caret users,
I am a complete novice with caret and would be more than greatful if
anybody
could reply to my request.
I am trying to map functional SPM2 data from a group study to a
structural
atlas. I have tried via attributes:metric:map functional volume to metric
and somehow I obviously have managed to import some data (hopefully
correctly) - but when trying to repeat the process I failed. This was
probably due to the fact that I'm not quite sure which individual steps
exactly I must take and the necessary order.
Thanks in advance for any information,
Meike Ramon
_______________________________________________
caret-users mailing list
caret-users@brainvis.wustl.edu
http://pulvinar.wustl.edu/mailman/listinfo/caret-users
