1) Sorry, I don't understand what you are trying to do 2) Depends on what you are testing. If you are looking at the slopes of the covariates, then demeaning is not important. If you are looking at group differences, you should first look for a difference in the slopes. If there is no significant differences, then reanalyze using DOSS instead of DODS in which case demeaning will not matter.
3) Correct On 09/29/2016 09:09 AM, Damien MARIE wrote: > Hi, > > Many thanks for your prompt reply Pr. Greve. > > I did : > > mri_glmfit --y > /home/damien/bin/freesurfer/subjects/qdec/RH_Dys-CT_CSA_AgeGenderWRCSA/y.mgh > --fsgd > /home/damien/bin/freesurfer/subjects/qdec/RH_Dys-CT_CSA_AgeGenderWRCSA/qdec.fsgd > > dods --glmdir > /home/damien/bin/freesurfer/subjects/qdec/RH_Dys-CT_CSA_AgeGenderWRCSA > --surf fsaverage rh --label > /home/damien/bin/freesurfer/subjects/fsaverage/label/rh.aparc.label > --C > /home/damien/bin/freesurfer/subjects/qdec/RH_Dys-CT_CSA_AgeGenderWRCSA/contrasts/rh-Avg-Intercept-area.mat > > --C > /home/damien/bin/freesurfer/subjects/qdec/RH_Dys-CT_CSA_AgeGenderWRCSA/contrasts/rh-Diff-dyslexic-control-Intercept-area.mat > > --C > /home/damien/bin/freesurfer/subjects/qdec/RH_Dys-CT_CSA_AgeGenderWRCSA/contrasts/rh-Diff-male-female-Intercept-area.mat > > --C > /home/damien/bin/freesurfer/subjects/qdec/RH_Dys-CT_CSA_AgeGenderWRCSA/contrasts/rh-X-group-gender-Intercept-area.mat > > --eres-save > > mri_segstats --in > /home/damien/bin/freesurfer/subjects/qdec/RH_Dys-CT_CSA_AgeGenderWRCSA/eres.mgh > > --sum toto --annot fsaverage rh mc-z.abs.th23.sig.ocn --avgwf > RH_Dys-CT_CSA_AgeGenderWRCSA.dat > > > 1) I don’t get it though. Probably I made a bad description of my > goal, perhaps using this terminology of “residuals” also is not very > accurate. > > I exported the txt file (output of mri_segstats computed on eres.mgh) > to a statistical software and ran a mancova per cluster on cluster > values with covariatess : age, right hemispheric surface area + gender > factor. I find significant effects of the covariates on 7/9 clusters. > > I was not expecting such effects. I wanted to extract residuals of > surface area in the significant clusters obtain using the Qdec Monte > Carlo simulation on the contrast of interest (Dys - Ct) after removing > all covariate effects. > > My aim was to look at the differences between the two levels of the > factor of interest to get a sense of what is happening in this Qdec > analysis and this particular contrast of interest. These results are > in an opposite direction in comparison with other Qdec analysis > without covariates / ROI analysis on surface area residuals after > removing effect of gender, age and right hemispheric surface area > (Destrieux's labels) / Whole-brain VBM analyses with covariates (not > on surface area then...) > > Basically I have the feeling there is something wrong with the > contrast display in this particular qdec analysis. I don’t know how I > could get what I called the CSA residuals in the framework of this > analysis. > > > 2) Another question: should I standardize or center my covariates > before proceeding to the analysis? > > > 3) A silly question probably but to be 100% sure… If in the > qdec.table.dat, order of subjects is first group “ dys “ then second > group “ ct “ and that the contrast is called: > rh-Diff-dys-dyslexic-control-intercept-area , negative p value means > decrease of surface area in the first group (dys) as compared to the > second group (ct), right ? > > > Thank you and best, > *Damien MARIE* > * > * > * > * > * > * > * > * > * > * > *Douglas Greve > <http://www.mail-archive.com/search?l=freesurfer@nmr.mgh.harvard.edu&q=from:%22Douglas+Greve%22>Wed, > > 28 Sep 2016 12:40:51 -0700 > <http://www.mail-archive.com/search?l=freesurfer@nmr.mgh.harvard.edu&q=date:20160928>* > * > * > *Those are not residuals. If you want residuals, you will need to run > mri_glmfit from the command line adding --eres-save. You can get the > command line from the mri_glmfit.log file in the glmdir* > > > > > * > On 9/28/16 10:58 AM, Damien MARIE wrote: > > Dear FreeSurfers, > > Sorry for the repost. I’m still trying to extract the residuals of > cortical surface area from a qdec analysis with two factors (2 > levels) and 2 covariates. > > I tried this command line: > > mri_segstats --in > > /home/damien/bin/freesurfer/subjects/qdec/RH_Dys-CT_CSA_AgeGenderWRCSA/y.mgh > --sum toto --annot fsaverage rh mc-z.abs.th23.sig.ocn --avgwf > RH_Dys-CT_CSA_AgeGenderWRCSA.dat I’ve got a text file with 9 > columns (number of clusters) and 69 lines (number of subjects). > Now I am not 100% sure it’s really the residuals (y.mgh) so I > would be glad if somebody could tell me please. > > Best, > Damien > Original post: > Dear experts, > > I would like to know what is the best and simplest way to extract cortical > surface area and thickness from clusters obtained after a Monte Carlo > correction. This is a group comparison I did with qdec including gender as > discrete variable + age and mean thickness or whole hemispheric surface > area as > nuisance factors, so I would like to get the feature residuals after the > removal of the covariate effects. > I am sorry, I saw several posts about that but for me it’s not completely > clear > what command line I should use to get tables for this… > > Thank you and best, > Damien > > > > > *Damien MARIE* > > > * > > > _______________________________________________ > Freesurfer mailing list > Freesurfer@nmr.mgh.harvard.edu > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Douglas N. Greve, Ph.D. MGH-NMR Center gr...@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422 Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 www.nmr.mgh.harvard.edu/facility/filedrop/index.html Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
