 |
 |
Thanks, Helmut. This is supremely helpful. It’s clear that one of our reviewers was bewildered by why we would not perform FDR on both the voxel AND cluster levels. Warm regards, Jeff On Sep 22, 2015, at 1:51 PM, H. Nebl <[log in to unmask]> wrote: > Dear Jeff, > > Yes, but you should explicitely state that it's topological FDR on cluster level then, as there are two different topological FDRs. You should also explicitely state the initial uncorrected voxel threshold. For citation go with Chumbley & Friston (2009, Neuroimage, http://dx.doi.org/10.1016/j.neuroimage.2008.05.021 ). > > For uncorrected voxel threshold and FWE-R correction on cluster level go with e.g. Friston et al. (1994, Hum Brain Mapp, http://dx.doi.org/10.1002/hbm.460010306 - there might be other papers). > > And for uncorrected voxel threshold and cluster correction in general, go with Woo et al. (2014, Neuroimage, http://dx.doi.org/10.1016/j.neuroimage.2013.12.058 ) for a recent reference. It's certainly not outdated / bad (and honestly speaking, also more trustworthy as you just have to look at the SPM output, so there's not much to do wrong - in principle at least, seems like the many corrected / uncorrected p values in the SPM output are frequently misunderstood). > > Best > > Helmut >