Dear SPMers,
As some of you know, use of cluster size p-values is highly discouraged
in VBM-type analyses because of non-stationarity (or non-uniform
smoothness) in typical VBM data. However, there is a correction method
which produces valid cluster size p-values under non-stationarity. The
RFT (random field theory) version of this correction has been
implemented in the FMRISTAT package by Keith Worsley. Thanks to Keith, I
was able to port this correction into SPM, so that valid cluster
p-values can be calculated with SPM for VBM-type analyses. In addition,
thanks to a few dozen beta-testers, the program has been tested
extensively on SPM2 and SPM5. I am please to announce that the program
is now publicly available from
http://fmri.wfubmc.edu/cms/NS-General
Features of this non-stationarity extension program include:
-Non-stationarity-adjusted cluster sizes in terms of RESELs
-Non-stationarity-adjusted p-values, both FWE corrected and uncorrected
-Cluster p-values for F-contrasts
The web site above explains how to download and install the program. It
is easy to use, very similar to the process associated with the
"Results" button in SPM.
If you have any questions or comments, please let me know.
Thanks!
-Satoru
Satoru Hayasaka PhD ----------
Assistant Professor, Public Health Sciences & Radiology
Wake Forest University School of Medicine
(ph) +1-336-716-8504 / (fax) +1-336-716-0798
(email) shayasak _at_ wfubmc _dot_ edu
|