Dear all,
Adam Thomas reported some differences in the results between the non-stationary
correction of VBM data using the NS-toolbox of Satoru Hayasaka
(http://www.fmri.wfubmc.edu/HTML_nS_Cluster) and my VBM toolboxes. While I use
some functions of Satorus toolbox the main differences of his toolbox is the use of the
function stat_threshold.m from Keith Worsley to compute thresholds and p-values
according to the Random Field Theory. I tested the differences in the past and found only
minor differences. However, Tom Nichols pointed out that stat_threshold.m is accounting
for the extra noise from estimating the smoothness locally, while spm_P.m assumes an
almost perfect RPV image. This might lead to differing p-values for the cluster extent esp.
for VBM data.
Thus, I have now replaced the threshold and p-value estimation by the function
stat_threshold.m from Keith Worsley, which should be more reliable and have updated
the VBM toolboxes at:
http://dbm.neuro.uni-jena.de/vbm/download/
Some more information about the non-stationary correction can be found at:
http://www.fmri.wfubmc.edu/cms/NS-General
http://dbm.neuro.uni-jena.de/vbm/non-stationary-cluster-extent-correction/
Furthermore, I have now renamed the original spm functions, which were used by my
toolbox to prevent interfering with other toolboxes.
The non-stationary correction can be now only used via:
Toolbox|VBM5.1|Results with non-stationary correction
Sorry for the long explanation, which might be only interesting for the VBM users and
thanks to Satoru, Tom, and Adam for their help.
Best regards,
Christian
--
____________________________________________________________________________
Christian Gaser, Ph.D.
Assistant Professor of Computational Neuroscience
Department of Psychiatry
Friedrich-Schiller-University of Jena
Jahnstrasse 3, D-07743 Jena, Germany
Tel: ++49-3641-934752 Fax: ++49-3641-934755
e-mail: [log in to unmask]
http://dbm.neuro.uni-jena.de
|