Dear FSL Experts, Dear Anderson,
We have used PALM to, successfully, explore some DTI data in a large sample of children/adolescents. One comment we have received (from a reviewer) concerns the number of permutations we use in our PALM analysis (N=5000). Based on previous discussions here on the forum our analysis seems to be in-line with lots of research.
I include the comment and our current command below. The GLM design includes 4 contrasts, so 5000 permutations ran on each contrast. If anyone could shed light/recommend evidence I would appreciate it. We are obviously hesitant to re-run the analysis with 10000 permutations unless absolutely necessary. We report all results at TFCE FWE corrected (log(p) = 1.3).
# "....at 5000 permutations the uncertainty in the p-value is approximately 0.0138, meaning that the results presented at 0.04 significance levels could actually be borderline. In such cases it would be advisable to at least mention this caveat or rerun with 10000 permutations or more."
palm -i all_FA_skeletonised.nii -m mean_FA_skeleton_mask.nii -d design_298_ANOVA_IQ_Age.mat -t design_298_ANOVA_IQ_Age.con -eb block.csv -vg auto -n 5000 -T -C 3.1 -fdr -corrcon -o anova_298_fa_iq_age –logp
Many thanks
Jack Rogers
########################################################################
To unsubscribe from the FSL list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=FSL&A=1
|