On 6/10/2020 9:54 AM, Avi Gharehgazlou wrote:
External Email - Use Caution
Hi experts,
I am a bit confused with visualizing my significant clusters after having corrected my results for multiple comparisons. after running mri_glmfit, I corrected with monte carlo and permutation as explained below, will you please confirm my understanding below is correct ?
- I ran monte carlo simulation (1.3 sig level). My understanding is
that "cache.th13.abs.sig.cluster.summary" shows the clusters that are significant after correction for multiple comparisons. I have 1 cluster listed in this file. Thus, I am interpreting this as having 1 significant cluster after correction at p<0.05, is this correct?
Yes
When I want to visualize my corrected results I use the command line below. I am assuming this command line displays only the significant clusters that survived multiple comparison correction because only the cluster that is listed in my "cache.th13.abs.sig.cluster.summary" file appears on the brain here, and because I have put " cache.th13.abs.sig.cluster.mgh" and as overlay " cache.th13.abs.sig.ocn.annot" and set the threshold to 1.3 in my command line. Is it correct that the cluster that appears when running the below command line is my significant cluster surviving multiple comparisons correction at p<0.05?
Yes
vglrun freeview -f $SUBJECTS_DIR/fsaverage/surf/lh.inflated:overlay=lh.lgi.TD_PTSD_age_SA_IQ_FWHM0.glmdir/group.diff/cache.th13.abs.sig.cluster.mgh:overlay_threshold=1.3,5:annot=lh.lgi.TD_PTSD_age_SA_IQ_FWHM0.glmdir/group.diff/cache.th13.abs.sig.ocn.annot -viewport 3d
also, how can I get the clusters to look their appropriate colours when I am vieweing them with the command line above? all my clusters appear gray (regardless of whether they have positive or negative max values, in which case they should be yellow and blue, respectively)
fscalc cache.th13.abs.sig.cluster.mgh mul -1 -o cache.th13.abs.sig.cluster.neg.mgh
- I then ran permutation testing. My
"perm.th13.abs.sig.cluster.summary" file is blank. I interpret this as no clusters having survived multiple comparison correction with permutation testing. Am I correct?
Yes
When I use the below command line to visualize my result I see a blank brain (with no clusters coloured). Which makes sense to me because my "perm.th13.abs.sig.cluster.summary" also showed no significant clusters. is this interpretation and the command line I am using below correct?
vglrun freeview -f $SUBJECTS_DIR/fsaverage/surf/lh.inflated:overlay=lh.lgi.TD_PTSD_age_SA_IQ_FWHM0.glmdir/group.diff/perm.th13.abs.sig.cluster.mgh:overlay_threshold=1.3,5:annot=lh.lgi.TD_PTSD_age_SA_IQ_FWHM0.glmdir/group.diff/perm.th13.abs.sig.ocn.annot -viewport 3d -layout 1
I think my understanding is correct, but I am confused because the FreeSurfer instructions state that the above command lines will display "all clusters, regardless of significance"... but using the above command lines seems to only display my significant clusters. I am assuming this is because I set my threshold to 1.3 (the same threshold I used to correct for multiple comparisons in 1 and 2 above). But wanted to make sure I am correct before proceeding.
If you want all clusters regardless of sig, you have to do two things: 1. When you run mri_glmfit-sim set the -cwp 1 2. Set the overlay_threshold=0.0001,5
Thank you so much for your support,
Avi
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer