This has nothing to do with the contrast. Try the suggestion that was printed out at the end (ie, remove the mean from the age and divide by the stddev)
On 06/04/2018 09:28 AM, C.P.E. Rollins wrote:
External Email - Use Caution Dear Freesurfer developpers,
I posted a while back asking how to control for scanner site in a multi-centre study. I think I now understand the formatting to do so for a DOSS model, but was wondering how this would look for a DODS model. There are 6 scanner sites and I also want to control for age and sex. However, there are no females in 3 of the sites. I tried to create the contrast for this but got the following 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.bin --y rh.5.lgi.centre.mgh --fsgd FSGD_centre.txt dods --glmdir rh.5.lgi.centre.glmdir --surf fsaverage rh --C H-NH_centre.dods.txt 2. The FSGD file (if using one) 3. And the design matrix above Attempting to diagnose further SumSq: Min=1.000000 (col 4), Max=183.117447 (col 31) The scale is much different between columns 4 and 31, you may want to normalize by subtracting the mean and dividing by the standard deviation.
My contrast was 1 -1 -1 -1 -1 -1 -1 -1 -1 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0.75 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 and FSGD attached (does not include full list of subjects).
Thanks very much, Colleen
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer