Hi Alex,

I don't know what "var2" means. Spearman cor is nonparametric based on rank. It's more closely related to Pearson, and would be the same as a Pearson if you only had one regressor. The sign indicates the direction of the contrast -- red for positive and blue for neg. You're already accounting for age in the FSGD (which is just a spec for a design matrix), then just set the age-related regressor cells in the contrast matrix to 0.

Did that mris_glm command-line work? It looks like you have the wrong number of components in the gcv

doug

Alexandre DaSilva wrote:

Hello everybody,

I am using the mris_glm tool to the test for the effect of a var2 on thickness (see file and command bellow), and I would like to understand the outcome. First, what is the difference in interpretation of the resulting p-value, and the p-value you'd get from a Pearson's or Spearman's correlation test between thickness and frequency. Second, what does the sign of the resulting significance map mean (red vs blue)? Third, is there a way to account for the effect of a regressor of no interest, for example age?

Thanks,

Alex

X.fsgdf

GroupDescriptorFile 1 Class MyTitle Variables Dur Freq Age Input subjid1 Class1 10 15 1000 Input subjid2 Class2 20 20 2000 Input subjid3 Class2 20 20 2000

mris_glm –fsgd X.fsdgf --surfmeas thickness --trgsubj brain –nsmooth 250 –gcv 0010 –sight ./thicknessCORfreq-sight-rh.w paint –hemi rh

---

---

Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer