Hi,
Thanks to Hiroki Tanabe and Russ Poldrack, I have found and fixed
a small problem with the FDR threshold function, spm_uc_FDR.m.
For very small q, the threshold was returned as an empty matrix,
instead of Inf. Infinity is a reasonable threshold; it says
that there is no threshold such that the FDR can be controlled at q.
You can get the fixed version at
http://www.sph.umich.edu/~nichols/FDR/spm_uc_FDR.m
or simply replace the last line
u = Ts(I);
with
if isempty(I)
u = Inf;
else
u = Ts(I);
end
-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
|