Hi caret-users,
Three things:
1. *h.white -- not *h.smoothwm: The initial version of this script
averaged the Freesurfer smoothwm and pial to generate a mid-thickness
fiducial. On Feb 23, 2006, Bruce Fischl wrote to David Van Essen:
I also kept meaning to mention to you that the ?h.smoothwm surface is
not the one we use for the gray/white boundary. It's an initial
estimate that gets refined to create the ?h.white surface.
So I have changed the script to use the white surface. Users of this
script should do the same.
2. Origin setting: The settings in this script produce a segmentation
volume whose AFNI .HEAD produces this 3dinfo output (AFNI command line
utility):
R-to-L extent: -128.000 [R] -to- 127.000 [L] -step- 1.000 mm [256 voxels]
A-to-P extent: -127.000 [A] -to- 128.000 [P] -step- 1.000 mm [256 voxels]
I-to-S extent: -127.000 [I] -to- 128.000 [S] -step- 1.000 mm [256 voxels]
This was based on my interpretation of
http://www.wideman-one.com/gw/brain/fs/coords/fscoords.htm. Per Ziad
Saad, 2/14/2006
(http://afni.nimh.nih.gov/afni/community/board/read.php?f=1&i=9147&t=9128),
"What you have posted is correct to our knowledge." This doesn't
eliminate the disclaimer, but at least two of us will be wrong if it
it's not accurate.;-)
3. This has nothing to do with Freesurfer, but I have changed my name.
Stay tuned for a forthcoming announcement of Caret 5.33...
Donna L. Dierker
(Formerly Donna Hanlon; no change in marital status -- see
http://home.att.net/~donna.hanlon for details.)
On 02/16/2006 08:46 AM, Donna Hanlon wrote:
Hi caret-users,
This script may be helpful to those who import Freesurfer surfaces
into Caret for various purposes.
http://brainmap.wustl.edu/pub/donna/SCRIPTS/FREESURFER/freesurfer2caret.sh
login pub
password download
It does the following:
* calls mris_convert to convert the Freesurfer binary surfaces to ASCII
* converts the Freesurfer ASCII surfaces to Caret coord/topo files
* averages the pial and smoothwm surfaces to generate a midthickness
coord
* creates a spec file with the hem_flag as left or right
* generates the sulcal depth and geography paint file from the
midthickness surface
Run this script in a 'subjects' directory where there is a
subdirectory for each case/subject. This script assumes there is a
surf subdirectory under each case subdirectory, e.g.:
./ED/surf
./EF/surf
./EG/surf
./FH/surf
./FK/surf
./FL/surf
./FM/surf
./FN/surf
Then, change this line in the script to reflect your actual
case/subject subdirectory names:
CASES="ED EF EG FH FK FL FM FN"
If you have additional layers between the case subdirectory and your
Freesurfer surf subdirectory, then you'll need to tweak these lines
accordingly:
cd $CASE/surf
...
cd ../..
Disclaimers:
1. The origin setting in the caret_command -volume-create step
reflects my best current understanding of where the Freesurfer origin
actually is. Help in clarifying this issue is most welcome. The
bottom line is that the depth measures will be accurate regardless of
the actual origin input here, because the segmentation and derivative
hull volume will be generated from the same origin, so the resulting
distance between the fiducial and hull surfaces will not be affected
by the actual origin setting; however, someone might try to align
these derivative volumes with an actual anatomical generated from the
COR-* slices and midthickness fiducial surface. There may be slight
offset issues, although the alignment looks good to me. More
importantly, it would be nice to add a Freesurfer atlas space for the
-volume-create and paint/metric-to-volume features.
2. I confess I haven't tested it as thoroughly as I'd like to.
Apologies: Sorry I didn't get this up sooner; my daughter was sick
yesterday.
Donna
