Hi FreeSurfers,
I was just reading this post to the archive (below) regarding using mri_glmfit-sim on volume data, and it suggests that using --grf rather than doing the simulation can be "inaccurate at low smoothness and low thresholds." At what smoothness values and threshold values should we opt to use the simulation over GRF for volume-based cluster corrections?
Thanks,
~Sue Ruiz
Hi Yuko, it should be performing Gaussian Random Field (GRF) cluster-wise correction. Based on the smoothness, the size of the search space, and the voxel-wise threshold, it attempts to compute the p-value of resulting clusters. The differences between GRF and doing the simulation are that GRF runs almost instantly but it can be inaccurate at low smoothnesses and low thresholds. It looks like the GRF was not implemented until 5.1
doug
yukoyy wrote:
- Dear all,
*>>* I'm conducting volume based group analysis and corrections for multiple comparisons. *>* My environment :Mac OS 10.6.8, Freesurfer 5.0.0 *>>* I followed the instruction I found here: *>* http://surfer.nmr.mgh.harvard.edu/fswiki/FsFastTutorialV5.1/FsFastGroupLevel http://surfer.nmr.mgh.harvard.edu/fswiki/FsFastTutorialV5.1/FsFastGroupLevel *>>* When I first successfully ran: *>* mri_glmfit --y ces.005.nii --wls cesvar.005.nii --osgm --glmdir F2glm.wls --nii.gz *>* tkmedit fsaverage orig.mgz -aparc+aseg -overlay F2glm.wls/osgm/sig.nii.gz -fminmax 3 7 *>>* Then, I wanted to do corrections for multiple comparison. *>* mri_glmfit-sim --glmdir F2glm.wls --grf 3.0 pos --cwpvalthresh .0166 *>>* However, FreeSurfer does not seem to recognize --grf, and I don't understand what does grf do. *>* %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% *>* ERROR: Flag --grf unrecognized. *>* --glmdir F2glm.wls --grf 3.0 pos --cwpvalthresh .0166 *>* %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% *>>* I chose to use --grf because it was suggested in the tutorial, and in FreeSurfer mailing list archive. *>* http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg20573.html http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg20573.html *>>* I am guessing that I cannot use --cache for volume analysis because I don't have *>* /Applications/freesurfer/average/mult-comp-cor/fsaverage/volume/cortex/fwhm12/pos/th30/mc-z.csd *>>* I also tried monte-carlo simulation, and it worked okay. *>>* So, here is my question. *>* What does --grf do? and Is there any documentation? *>>* Thank you! *>* Yuko Yotsumoto *>* yukoyy at a7.keio.jp https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer *>* _______________________________________________ *>* Freesurfer mailing list *>* Freesurfer at nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer *>* https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer *>>>
If you want to use GRF, you should use smooth by twice the native voxel size. I would not go higher than .01 on the threshold doug
On 01/26/2014 12:52 PM, Susan Ruiz wrote:
Hi FreeSurfers, I was just reading this post to the archive (below) regarding using mri_glmfit-sim
on volume data, and it suggests that using --grf rather than doing the simulation
can be "inaccurate at low smoothness and low thresholds."
At what smoothness values and threshold values should we opt to use the
simulation over GRF for volume-based cluster corrections? Thanks, ~Sue Ruiz
Hi Yuko, it should be performing Gaussian Random Field (GRF) cluster-wise correction. Based on the smoothness, the size of the search space, and the voxel-wise threshold, it attempts to compute the p-value of resulting clusters. The differences between GRF and doing the simulation are that GRF runs almost instantly but it can be inaccurate at low smoothnesses and low thresholds. It looks like the GRF was not implemented until 5.1
doug
yukoyy wrote:
/ Dear all,
/>/ />/ I'm conducting volume based group analysis and corrections for multiple comparisons. />/ My environment :Mac OS 10.6.8, Freesurfer 5.0.0 />/ />/ I followed the instruction I found here: />/ http://surfer.nmr.mgh.harvard.edu/fswiki/FsFastTutorialV5.1/FsFastGroupLevel />/ />/ When I first successfully ran: />/ mri_glmfit --y ces.005.nii --wls cesvar.005.nii --osgm --glmdir F2glm.wls --nii.gz />/ tkmedit fsaverage orig.mgz -aparc+aseg -overlay F2glm.wls/osgm/sig.nii.gz -fminmax 3 7 />/ />/ Then, I wanted to do corrections for multiple comparison. />/ mri_glmfit-sim --glmdir F2glm.wls --grf 3.0 pos --cwpvalthresh .0166 />/ />/ However, FreeSurfer does not seem to recognize --grf, and I don't understand what does grf do. />/ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% />/ ERROR: Flag --grf unrecognized. />/ --glmdir F2glm.wls --grf 3.0 pos --cwpvalthresh .0166 />/ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% />/ />/ I chose to use --grf because it was suggested in the tutorial, and in FreeSurfer mailing list archive. />/ http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg20573.html />/ />/ I am guessing that I cannot use --cache for volume analysis because I don't have />/ /Applications/freesurfer/average/mult-comp-cor/fsaverage/volume/cortex/fwhm12/pos/th30/mc-z.csd />/ />/ I also tried monte-carlo simulation, and it worked okay. />/ />/ So, here is my question. />/ What does --grf do? and Is there any documentation? />/ />/ Thank you! />/ Yuko Yotsumoto />/ yukoyy at a7.keio.jp https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer />/ _______________________________________________ />/ Freesurfer mailing list />/ Freesurfer at nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer />/ https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer />/ />/ />/ / -- Douglas N. Greve, Ph.D. MGH-NMR Center greve at nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer Phone Number: 617-724-2358 Fax: 617-726-7422 --
Susan Mosher Ruiz, Ph.D. Research Scientist Boston University School of Medicine Laboratory of Neuropsychology VA Boston Healthcare System smosher@bu.edu mailto:smosher@bu.edu
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
freesurfer@nmr.mgh.harvard.edu
-
Douglas N Greve -
Susan Ruiz