Dear Donna,
Thanks for your reply.
The command line options will be really useful for scripting stat
tests. I'm going to download and install the update asap. My lab
uses mac osx (intel) and linux platforms. I hope the new versions
are available for these platforms. Thanks John!
Also, just a quick check with regard to the permutation analysis when
doing a one-sample t-test. Obviously in this case you cannot permute
the experimental conditions (because you are testing that t > 0). I
am assuming the spatial positions of the nodes of the surface are
randomly reordered to compute the permuted distribution. Right? Or
is there a random sign flipping of the values of the nodes? These
are two common methods used in permutation tests that I know about.
Thank again!
Clayton Curtis
On Jan 10, 2007, at 4:23 PM, Donna Dierker wrote:
Hi Clayton,
1. A reasonable question, but I'm afraid this document is both
incomplete and obsolete:
http://brainvis.wustl.edu/courses/Caret_5_332_Tutorial_Analysis.doc
For example, Find Significant Clusters is now called Two-Sample T-
test. But maybe it will still prove helpful. This submenu is one of
the fastest changing in Caret right now, so it's very much a moving
target. The two sample t-test is spelled out in the methods of our
Williams Syndrome (WS) paper:
Van Essen, D.C., Dierker, D., Snyder, A., Raichle, M.E., Reiss, A.,
and Korenberg, J. (2006) Symmetry of cortical folding abnormalities
in Williams syndrome revealed by surface-based analyses. J.
Neuroscience 26:5470-5483.
http://brainvis.wustl.edu/resources/WmsJNsci.pdf
It's basically the surface equivalent of the suprathreshold cluster
test described in Nichols & Holmes primer paper. The
interhemispheric test also is described in our WS paper.
If you have questions about other tests, let us know.
2. Thanks to John Harwell, the following tests are available via
command line:
vp.wustl.edu 22% caret_command -help | grep metric-or-shape-stat
INFO: To see command line options, Run with "-help"
METRIC/SHAPE STATISTICS COORDINATE DIFFERENCE -metric-or-shape-stat-
coord-diff
METRIC/SHAPE STATISTICS INTERHEMISPHERIC CLUSTERS -metric-or-shape-
stat-interhemispheric-clusters
METRIC/SHAPE STATISTICS ONE SAMPLE T-TEST -metric-or-shape-stat-one-
sample-t-test
METRIC/SHAPE STATISTICS PAIRED T-TEST -metric-or-shape-stat-paired-
t-test
METRIC/SHAPE STATISTICS TWO SAMPLE T-TEST and WILCOXON -metric-or-
shape-stat-two-sample-t-test
But these weren't added until 12/15, so you probably will need to
download a new version; let us know which platform you use.
On 01/10/2007 12:34 PM, Clayton E Curtis wrote:
Hi,
I have two quick questions.
1. What's the best place to get documentation on the new metric/
surface
shape statistical operations? We've looked at the documentation at
http://brainmap.wustl.edu/caret/caret5.5_help/statistics/ but
would like to have some more details, especially about the
permuation analyses.
2. Are there command line options (e.g., caret_command) for
performing
these statistical tests on metric files?
Thanks!
Clayton Curtis
___
caret-users mailing list
caret-users@brainvis.wustl.edu
http://pulvinar.wustl.edu/mailman/listinfo/caret-users
--
Donna L. Dierker
(Formerly Donna Hanlon; no change in marital status -- see http://
home.att.net/~donna.hanlon for details.)
___
caret-users mailing list
caret-users@brainvis.wustl.edu
http://pulvinar.wustl.edu/mailman/listinfo/caret-users