Frida has helped find another bug (thanks Frida!). When running mri_surfcluster with average7 inside your csd file, make sure to specify --subject average7 --really-use-average7 or else it will underestimate the clustersize and cause the significances to be too low. I'll mod the code when I get back from CA.
doug
On Tue, 14 Mar 2006, Frida Polli wrote:
yes, that indeed fixes it - you might want to put some of this stuff in the documentation or send out an email or something b/c i spent two weeks running another analysis b/c i thought my clusters weren't meeting significance, and only discovered this discrepancy when i did a regular mri_surfcluster :). so i can imagine since many people are still using average7, this would be really importnat for them to know. i am cc'ing nurit from moshe;s lab, as i know she's been using it - there are probably others. that having been said, thanks for the tool - its great. thanks, frida On Monday, March 13, 2006, at 04:44 PM, Doug Greve wrote:
Oh, I know what happened. It only fixes the vertex area when you pass average7 in explicitly. When you use a csd, it get's the subject name from the csd. The ratio of the areas is 1.5. I think it will work if you specify --subject average7 with the csd (and --really-use-average7). Can you try that and let me know?
doug
On Mon, 13 Mar 2006, Frida Polli wrote:
hi doug i am very confused. i used mri_surfcluster on one file in two different ways, and for one cluster, got two different sizes for all the clusters (see example of one cluster below). the first way is just the standard way, the second way is using csd files and having it output CWPs.
for one cluster, the first way gave me the following result: 4 3.230 71230 1065.25 8.0 12.9 25.4 1788
the second way gave me the following result: 4 3.230 71230 710.17 8.0 12.9 25.4 0.36277 0.35092 0.37500 1788
as you can see, this is the same cluster, but with two wildly discrepant values for surface area. what is going on, and which is correct?
the first way, i ran: mri_surfcluster --really-use-average7 --src BG3b_glmfit/rh.ASevASc_6s_BwGroup3b.fsfast.glmdir/ BwGroupError3b_NvSZ_contrast/sig.mgh --hemi rh --srcsubj average7 --thmin 1.3 --minarea 100 --sum rh.BG3_6s.sum
the second way i ran: mri_surfcluster --src BG3b_glmfit/rh.ASevASc_6s_BwGroup3b.fsfast.glmdir/ BwGroupError3b_NvSZ_contrast/sig.mgh --csd csdfiles/BG3b_ASevASc_6s_rh_MCSIMz-BwGroupError3b_NvSZ_contrast.csd --sum BG3b_glmfit/rh.ASevASc_6s_BwGroup3b.fsfast.glmdir/ BwGroupError3b_NvSZ_contrast/sig.cluster.sum --ocp BG3b_glmfit/rh.ASevASc_6s_BwGroup3b.fsfast.glmdir/ BwGroupError3b_NvSZ_contrast/sig.cluster.mgh --thmax infinity
in /space/ventzl/1/users/SUBJECTS_DIR
thanks, frida
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
In order to help us help you, please follow the steps in: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
Frida Polli PhD Candidate, Suffolk University Research Fellow in Psychiatric Neuroimaging Massachusetts General Hospital CNY 149-2608, 13th Street Charlestown, MA 02129 phone: (617) 724 4604 fax: (617) 726 4078 email: fpolli@nmr.mgh.harvard.edu