[caret-users] exporting foci coordinates to another stereotaxic space

[Donna Dierker]

Hi Jeff,

You were only one subscription away from becoming our 500th caret-user!

David has been doing a lot of work on foci, so he'll probably have 
better answers than these, but here are my thoughts:

You could load the SPM99 foci on the SPM99 fiducial surface; project 
them; and then "unproject" them on the 711-2B fiducial surface.  Every 
Caret distribution has files like this in 
/usr/local/caret/data_files/fmri_mapping_files:

Human.PALS_B12.LEFT_AVG_B1-12.FIDUCIAL_711-2C.clean.align.73730.coord
...
Human.PALS_B12.LEFT_AVG_B1-12.FIDUCIAL_SPM99.clean.73730.coord
Human.PALS_B12.RIGHT_AVG_B1-12.FIDUCIAL_711-2C.clean.align.73730.coord
...
Human.PALS_B12.RIGHT_AVG_B1-12.FIDUCIAL_SPM99.clean.73730.coord

(On Windows, this might be c:\caret\data_files\fmri_mapping_files.)
*
*So start with SPM99 fiducial loaded in the main window; load foci; 
project; load 711-2C fiducial; and then do:

File: Save Data File: Foci File and make sure 711-2C fiducial is the 
reference surface.

One problem with this method is that the quality of the projection can 
be iffy where the original 3D coordinates are far from the surface.  It 
certainly isn't ideal for subcortical foci, for example.

Applying an affine transform to the coordinates is the most common 
solution.  There are plenty of tools for applying affine transforms to 
lists of coordinates, and Avi Snyder already has computed a avg152T1 
to/from 711-2B matrix, which I can dig out.  This doesn't depend on the 
surface, and works well for cortical and subcortical regions.

As for nonlinear solutions, one could use SPM, FNIRT, AIR or other tool 
to spatially normalize MeanBuckner12_711-2P (the volume) to 
MeanBuckner12_SPM99, and then apply the resulting deformation map to the 
SPM99 3D coordinates.  This mapping doesn't depend on the surface, 
either, but would be more work than is done in most studies to transform 
coordinates across spaces.  The links at 
http://brainvis.wustl.edu/help/pals_volume_normalization give you some 
idea of what is involved.  I could provide those mean volumes.

Donna

On 04/30/2009 02:50 PM, Jeff Zacks wrote:
> Hi all,
>
> I've entered a study in SumsDB and projected it to PALS in caret. The 
> study coordinates are reported in SPM99 space and I would like to 
> export them to 711-Y space for further analysis. Does caret support 
> any scheme for mapping coordinates from one stereo space to another?
>
> Thanks!
>
> Jeff
>
>
> Jeffrey M. Zacks, Ph.D.
> Washington University
> Department of Psychology
> Campus Box 1125
> St. Louis, MO 63130-4899
> 314-935-8454
> 314-935-7588 (fax)
> jzacks at artsci.wustl.edu <mailto:jzacks at artsci.wustl.edu>
> http://dcl.wustl.edu/~jzacks <http://dcl.wustl.edu/%7Ejzacks>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> caret-users mailing list
> caret-users at brainvis.wustl.edu
> http://brainvis.wustl.edu/mailman/listinfo/caret-users
>