Dear Freesurfers,
I have performed a permutation test with the command:
mri_glmfit-sim --glmdir rh.fsgd_subj_gauss005 --sim mc-z 10000 THRESHOLD csd --sim-sign pos --seed 1297708255
And after that I have obtained the clusters with:
mri_surfcluster --subject CNTRL_MCI_RAFA_average –mask rh.fsgd_subj_gauss005/mask.mgh --sum rh.fsgd_subj_gauss005/contrast/clustsum.txt --in rh.fsgd_subj_gauss005/contrast/sig.mgh --cwsig rh.fsgd_subj_gauss005/contrast/cwsig.mgh --vwsig rh.fsgd_subj_gauss005/contrast/vwsig.mgh --sign pos --csd rh.fsgd_subj_gauss005/csd/csd.j001-contrast.csd
I’ve did it over the same dataset but with two different thresholds, THRESHOLD: 2.301 (pvalue=0.005) and 1.301 (pvalue=0.05). As far I know the cwsig.mgh resulting using the 1.3 threshold should be similar than the cwsig.mgh from the second thr, but with a more restrictive significative clusters (some coloured areas should disappear, in other words, clusters returned in the first case should be included in clusters returned by the 1.3 threshold). My results didn’t show that behaviour, some of the cwsig clusters can be contained in the other, but others don’t (see attached Figure). It is my reasoning correct? If so, why the images mismatching?
Thank you all in advance. Rafa.
