Hello SPM-users,
I would like to re-post a question on FDR multiple test correction and VBM.
We have processed n=80 subjects acc. to an optimized VBM protocol including
Jacobian modulation and smoothing at FWHM 12 mm. Six multiple regression
models have been set up, the main regressor of interest (of each) being a
behavioural result (psychological test battery), along with covariates age
and gender.
(For simpl. I use the term 'activation' for voxels that are significant after
applying a directed t-test on the main regressor).
At uncorrected p < 0.001 correlation maps show activations well balanced
between focal/regional (not too excessive), also reasonable variation between
the tests is found. Correction at FDR 0.05 tends to lead to rather excessive
activations in most contrasts.
Using a more stringent q < 0.005 brings produces areas about comparable with p
uncorr 0.001. However, with the FDR being adaptive the effect of the
correction differs between contrasts and it seems odd to adapt the FDR
threshold for each contrast.
So I used the script FDRill to check the T-distribution: for most contrast it
is clearly shifted (up to about +2), so a number of questions arise:
1. Is FDR appropriate at all and how could I deal with the adaptiveness of
this type of multiple test correction, i. e. is it appropriate to use
different q values to different contrasts - which obviously contain results
at different 'strength' end extent'?
2. Can results - as earlier suggested - be p resented in the way that
uncorrected voxels and clusters are listed at uncorrected p=0.001 but in
addition it is checked if clusters survive FDR, and only such clusters are
discussed as significant result?
3. The 'ill' T-distribution: I have not modelled total GM volume as covariate
because we were also looking for regional (e. g. a whole lobe) effects and
thought this would be overcorrected by such a covariate.
What possibilites are there to improve this?
Any help is appreciated!
Thanks very much in advance!
Philipp Saemann
Max-Planck-Institute of Psychiatry
Kraepelinstr. 2-10
80804 Munich
Germany
|