Print

Print
Dear Jie Yu,

It's not strange. TFCE, either corrected or uncorrected, considers height and spatial extent of signals, whereas the direct result of the t-test, as you are using with FDR, considers only height. Therefore,no surprise to find significant results with TFCE and not with FDR-corrected voxelwise.

All the best,

Anderson



2013/2/3 Jie Yu <[log in to unmask]>
Hi
I performed randomise TFCE and FDR after VBM analysis.
Although I got significant results with TFCE, I "never" get any result with FDR...
That's kind of strange, so I just want to make sure that I did things in the right way.
Thanks!

" randomise -i GM_mod_merg_s3 -m GM_mask -o FDR -d design.mat -t design.con -n 5000 -D -x "


and

" fdr -i FDR_vox_p_tstat1 --oneminusp -m mask -q 0.05 --othresh=thresh1 "

Is it okay?

Thanks a lot!