Hi Joelle,
> Was wondering if anybody can help with this?
We'll see :)
> I'm looking at the results table of a second-level analysis of fMRI
> data. Specifically, I find that all peak level and cluster level are
> significant on the pFWE corrected. However not all are significant in
> the q FDR corrected. Is this a problem for making a conclusion that
> these voxels are sig active?
Basically, the idea is that you really should correct for multiple
comparisons. You can do this in several ways, of which voxelwise FWE and
FDR or cluster-corrections are just different flavors, each with
advantages and disadvantages. If you have selected one approach based on
its merits and you are informed about its shortcomings, it is perfectly
legitimate to simply use this approach and be done with it. You do not
have to compare approaches (although I would suspect that most of us did
for a given dataset, with more or less legitimate motivations :)
> However, I am already setting a FWE correction, with p 0.05, and an
> extent threshold of 20 voxels. Of course there is no one way that is
> agreed upon, but I wanted to know if this is enough to show significance
> here.
A voxelwise FWE correction is among the stricter ways to correct for
multiple comparisons, and you would not even have to set an additional
extent threshold as each voxel already passed the test. I think it may
be neuobiologically legitimate to do so but others may disagree. It will
only make it more stringent, though, so no worries w.r.t. being too lenient.
Cheers,
Marko
--
____________________________________________________
PD Dr. med. Marko Wilke
Facharzt für Kinder- und Jugendmedizin
Leiter, Experimentelle Pädiatrische Neurobildgebung
Universitäts-Kinderklinik
Abt. III (Neuropädiatrie)
Marko Wilke, MD, PhD
Pediatrician
Head, Experimental Pediatric Neuroimaging
University Children's Hospital
Dept. III (Pediatric Neurology)
Hoppe-Seyler-Str. 1
D - 72076 Tübingen, Germany
Tel. +49 7071 29-83416
Fax +49 7071 29-5473
[log in to unmask]
http://www.medizin.uni-tuebingen.de/kinder/epn/
____________________________________________________
|