Hi Adam, your simulation is based on a whole-brain search space and so it is not appropriate to use a label constraint when running mri_surfcluster. If you want to constrain the correction of multiple comparisons to a label, you have to do it when you do the simulation. The best approach is actually to constrain when running mri_glmfit, then use mri_glmfit-sim rather than running mri_glmfit with the simulation options.
doug
Adam Nitenson wrote:
Hi Freesurfers
I'm a bit confused as to some recent Monte Carlo findings I've been looking at. After generating 10,000 estimations with mri_glmfit. I used mri_surfcluster to generate nii, ocn.nii, and .sum files to analyze the clusters. However I found that, when looking at the dACC, the cluster within an ROI (label) specific analysis, which almost filled the entire label, was not significant, yet the dACC region cluster in the label-free, whole brain analysis was significant.
I figured that, since the whole brain (no label) monte carlo analysis compares the clusters to the entire brain and the label analysis confines the cluster to just the label, that the latter would result in a more signifcant cluster (since it fills up almost all of the space under consideration). However, I have found the opposite, and don't understand why this is.
Any ideas would be greatly appreciated. I've written my commands below (Stable 4).
Thanks,
Adam
mri_glmfit --y /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/error_vs_correct/lh.ces.005.nii --fsgd /autofs/space/ventzl_036/users/RespMon12C/fsgd_and_contrast_files/Adam_25MatchedControls_genotyped.fsgd --C /autofs/space/ventzl_036/users/RespMon12C/fsgd_and_contrast_files/CHomozyg_Minus_TCarrier.mtx --surf fsaverage lh --sim mc-full 1000 1.3 /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/Monte_Carlo_LH_error_vs_correct_6sec/LH_MCSIM1 --fwhm 16.351771
mri_surfcluster --in /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/LH_error_vs_correct_6sec/CHomozyg_Minus_TCarrier/sig.mgh --csd /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/Monte_Carlo_LH_error_vs_correct_6sec/LH_MCSIM1-CHomozyg_Minus_TCarrier.csd ....(rest of monte carlo files).... --srcsubj fsaverage --thmax infinity --annot aparc --clabel /autofs/space/ventzl_036/users/RespMon12C/subjects/fsaverage/label/lh.dACC.All56Grouplabel --sum /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/MC_Findings/LH_56Group_dACC.sum --o /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/MC_Findings/LH_56Group_dACC.nii --ocn /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/MC_Findings/LH_56Group_dACC.ocn.nii
mri_surfcluster --in /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/LH_error_vs_correct_6sec/CHomozyg_Minus_TCarrier/sig.mgh --csd /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/Monte_Carlo_LH_error_vs_correct_6sec/LH_MCSIM1-CHomozyg_Minus_TCarrier.csd ........ --srcsubj fsaverage --thmax infinity --annot aparc --sum /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/MC_Findings/LH_WholeBrain.sum --o /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/MC_Findings/LH_WholeBrain.nii --ocn /autofs/space/ventzl_036/users/RespMon12C/group-analysis/Adam_25MatchedControls_genotyped/RM_correct_error/CC_Minus_T/MC_Findings/LH_WholeBrain.ocn.nii
Adam Nitenson, B.S. Brain Genomics Lab Massachusetts General Hospital _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer