can you send the design matrix Xg.mtx and the full terminal output?

On 08/08/2017 09:13 AM, Sullivan, Danielle Renee wrote:

Hi FreeSurfer Experts,

I am trying to run a dods cortical thickness analysis examining the interaction of group on a continuous variable while controlling for other variables.

Specifically, I have 6 classes (gender by three groups [zero, one, two]). I am interested in the group interaction with a continuous variable while controlling for gender (among 4 other continuous variables). I set up my fsgd file (see attached) and contrast file (36 dimensions- F test):

*0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 [1 -1 0 1 -1 0]/2*

*0 0 0 0 0 0 **0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 [1 -0 -1 1 0 -1]/2*

and ran this command:

*mri_glmfit --y **lh.age.PC1_forFSGD.PC2_forFSGD.PC3_forFSGD.GGT_forFSGD.rs8042149_A_bysex_forfsgd.thickness.10.mgh --fsgd fsgd.age.PC1_forFSGD.PC2_forFSGD.PC3_forFSGD.GGT_forFSGD.rs8042149_A_bysex_forFSGD.152.thickness.08082017.txt dods --C /GLM_CONTRAST.MAT/Contrast.Male.Female.Age.PC1.PC2.PC3.GGT.RORAbySex.mtx --surf fsaverage lh --cortex --glmdir lh.age.PC1_forFSGD.PC2_forFSGD.PC3_forFSGD.GGT_forFSGD.rs8042149_A_bysex_forfsgd.thickness.10.glmdir

But keep getting this error:

*ERROR: matrix is ill-conditioned or badly scaled, condno = 1e+08*

*--------------------------------*

*Possible problem with experimental design:*

*Check for duplicate entries and/or lack of range of*

*continuous variables within a class.*

*If you seek help with this problem, make sure to send:*

• 1. Your command line:*

• mri_glmfit --y

lh.age.PC1_forFSGD.PC2_forFSGD.PC3_forFSGD.GGT_forFSGD.rs8042149_A_bysex_forfsgd.thickness.10.mgh --fsgd fsgd.age.PC1_forFSGD.PC2_forFSGD.PC3_forFSGD.GGT_forFSGD.rs8042149_A_bysex_forFSGD.152.thickness.08082017.txt dods --C /Volumes/VA_Imaging/Projects/salat/2389_TRT/dmiller/GGT_CRP/GLM_CONTRAST.MAT/Contrast.Male.Female.Age.PC1.PC2.PC3.GGT.RORAbySex.mtx --surf fsaverage lh --cortex --glmdir lh.age.PC1_forFSGD.PC2_forFSGD.PC3_forFSGD.GGT_forFSGD.rs8042149_A_bysex_forfsgd.thickness.10.glmdir

• 2. The FSGD file (if using one)*
• 3. And the design matrix above*

Any idea on what I am doing wrong?

Danielle R. Miller, Ph.D.

National Center for PTSD

VA Boston Healthcare System Jamaica Plain

Boston University School of Medicine

OFFICE: (857) 364-4022

---

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