Hi,
FDR returns a zero threshold if it believes that nothing is
significant (by its definition). Because (voxelwise) FDR doesn't take
advantage of spatial extent of signal, it is possible that it will
therefore not find anything when cluster-based thresolding does.
Cheers.
On 14 Jul 2008, at 14:40, George He wrote:
> Hi,
>
> I have a p-map (from ttologp and fslmaths) that looked reasonable
> (min=0.00027).
> The statistic images "thresholded using clusters determined by Z>2.0
> and a (corrected) cluster significance threshold of P=0.05" showed
> some clusters of activation.
> But fdr returns probability threshold 0 even with -q 0.8. Further
> more, fdr returns nothing between 0 and 0.557279, as you can see
> from my screenshot.
>
> $ fdr -i p1 -m ../mask -q 0.82280782
> Probability Threshold is:
> 0.557279
> $ fdr -i p1 -m ../mask -q 0.82280781
> Probability Threshold is:
> 0
>
> Can this be right? (The mask image seems all right, too). I would
> think the threshold cannot be exactly 0, even though it can be very
> small, say 10^-6.
>
> Thanks,
> George
>
---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------
|