Kenneth,
On Thu, 1 Sep 2005, Sung Lai Yuen wrote:
> I have a naive question regarding the omnibus test. I know in the
> SPM results display we could find the voxel, cluster, and set level
> p-values for a particular study. Nevertheless, if I decided to adopt
> a particular height & extend threshold, e.g. uncorrected p=0.0005,
> extend=16, how could I get the equivalent thresholding p-value at
> the cluster level? Are there any equations/tables available for the
> quick conversion?
To get the corrected cluster P-value for a 16 voxel cluster with an
uncorrected P-value threshold 0.0005, use this snippet
load SPM
k = 16; % cluster size
u = spm_invTcdf(1-0.0005,df); % uncorrected 0.0005 thresh
df = [1 SPM.xX.erdf]; % DF
STAT = 'T'
R = SPM.xVol.R; % Resels
FWHM = SPM.xVol.FWHM;
v2r = 1/prod(FWHM(~isinf(FWHM))); %-voxels to resels factor
n = 1;
P = spm_P_RF(1,k*v2r,u,df,STAT,R,n) % Corrected cluster P
I'm not sure this answers your question though...
-Tom
-- Thomas Nichols -------------------- Department of Biostatistics
http://www.sph.umich.edu/~nichols University of Michigan
[log in to unmask] 1420 Washington Heights
-------------------------------------- Ann Arbor, MI 48109-2029
|