I would normalize all the variables. Eg, column 2 is nearly identical to column 1

On 1/22/2020 1:03 AM, vittal korann wrote:

External Email - Use Caution

Dear Douglas

I did normalize ICV values. This is something else I guess.

With regards Vittal

On Tue, Jan 21, 2020 at 6:09 PM vittal korann <vittalkorann@gmail.commailto:vittalkorann@gmail.com> wrote: Dear FreeSurfer experts

I ran QDEC and got the following error.

Design matrix ------------------ 1.00000 1.04000 -1.38574 2.00000 29.00000 19.00000; 1.00000 1.04000 -0.55066 2.00000 35.00000 19.00000; 1.00000 1.03000 2.31721 2.00000 23.00000 14.00000; 1.00000 1.03000 -0.22919 2.00000 23.00000 16.00000; 1.00000 1.03000 -2.69688 -2.00000 28.00000 19.00000; 1.00000 1.01000 0.84981 -2.00000 22.00000 19.00000; 1.00000 1.01000 0.43845 2.00000 36.00000 21.00000; 1.00000 1.06000 1.32888 2.00000 28.00000 17.00000; 1.00000 1.02000 0.52292 2.00000 34.00000 22.00000; 1.00000 1.02000 0.18104 2.00000 32.00000 17.00000; 1.00000 1.03000 -0.33783 -2.00000 32.00000 22.00000; 1.00000 1.04000 0.65805 2.00000 30.00000 17.00000; 1.00000 1.03000 1.21888 2.00000 37.00000 10.00000; 1.00000 1.02000 2.11598 2.00000 29.00000 10.00000; 1.00000 1.02000 -0.02853 2.00000 27.00000 16.00000; 1.00000 1.04000 0.21777 2.00000 36.00000 16.00000; 1.00000 1.01000 0.30821 2.00000 29.00000 15.00000; 1.00000 1.03000 -0.69389 2.00000 30.00000 12.00000; 1.00000 1.04000 0.44792 2.00000 23.00000 15.00000; 1.00000 1.03000 0.15876 2.00000 28.00000 15.00000; 1.00000 1.02000 -1.05796 2.00000 36.00000 10.00000; 1.00000 1.05000 -1.51878 2.00000 32.00000 15.00000; 1.00000 1.03000 -0.91870 -2.00000 30.00000 15.00000; -------------------------------- ERROR: matrix is ill-conditioned or badly scaled, condno = 68200 -------------------------------- 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.bin --y /usr/local/freesurfer/subjects/qdec/Untitled/y.mgh --fsgd /usr/local/freesurfer/subjects/qdec/Untitled/qdec.fsgd dods --glmdir /usr/local/freesurfer/subjects/qdec/Untitled --surf fsaverage lh --label /usr/local/freesurfer/subjects/fsaverage/label/lh.aparc.label --C /usr/local/freesurfer/subjects/qdec/Untitled/contrasts/lh-Avg-Intercept-thickness.mat --C /usr/local/freesurfer/subjects/qdec/Untitled/contrasts/lh-Avg-thickness-Vein_Avg_tort-Cor.mat 2. The FSGD file (if using one) 3. And the design matrix above Attempting to diagnose further SumSq: Min=4.795832 (col 1), Max=145.206757 (col 5) The scale is much different between columns 1 and 5, you may want to normalize by subtracting the mean and dividing by the standard deviation.

Any help would be appreciated!

With regards Korann

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