Print

Print


Dear Stefan,

> I am trying to perform a correction at the cluster level when using an
> F-test. From checking the archives I understand that this has not yet
been
> implemented in SPM5.
> 
> Does anybody know a temporary fix for this. Fmristat
> (http://www.math.mcgill.ca/keith/fmristat/) has been suggested in an
> earlier
> mail but I could not find the right tools there either.

Well, actually we have written a wrapper program that allows you to
calculate cluster p-values for F-contrasts via fmristat. For more
details, see

http://www.jiscmail.ac.uk/cgi-bin/wa.exe?A2=ind0708&L=SPM&P=R1016&I=-3&m
=22297

This toolbox is designed for cluster p-value calculation in VBM-like
analyses (i.e., under non-uniform smoothness or non-stationarity). But
there is a hidden feature in this NS toolbox that allows you to
calculate cluster p-values (t- or F-contrasts). This can be done by
creating a global variable called ASSUME_STATIONARY and setting it to 1.
Here is an excerpt from its documentation.

% A hidden feature of spm_list_nS allows calculation of cluster p-values
% for t- and F-images under stationarity as well. This is implemented by

% the stat_thresh function using the results by Cao [1]. This feature is
% activated by creating a global variable called ASSUME_STATIONARY and
% set it to 1. Otherwise spm_list_nS calculates cluster p-values under
% non-stationarity.
%
% [1] Cao J, The Size of the Connected Components of Excursion Sets of
%     Chi^2, t, and F Fields. Advances in Applied Probability 31:579-595
%     (1999)

Good luck!
-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