External Email - Use Caution
Hi again,
What about for DODS? 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 again, Colleen
On 2018-05-08 16:56, C.P.E. Rollins wrote:
Ah, I think I understand now, thanks. I've attached my contrast and FSGD file to confirm. In the FSGD file, the format for the classes are <group>_<sex>_<centre>. Please let me know.
Thanks again, Colleen
On 2018-05-04 09:49, C.P.E. Rollins wrote:
Hi,
Sorry for coming back to this, but I'm still confused about the issue of having unbalanced classes. I thought that I could not remove the class as you suggested since this would mean that the contrast positive and negative values would no longer add up to 0. For instance if I remove only class "patient_centre2_female", then my contrast would have 11 (-1) and 12 (1) as opposed to 12 and 12. Could you please clarify?
Thanks again, Colleen
On 2018-04-30 14:43, C.P.E. Rollins wrote:
I thought that I could not remove the class since this would mean that the contrast positive and negative values would no longer add up to 0. For instance if I remove only class "patient_centre2_female", then my contrast would have 11 (-1) and 12 (1) as opposed to 12 and 12. Could you please clarify?
Thanks again, Colleen
-------- Original Message -------- Subject: Re: [Freesurfer] FSGD design for multi-centre study Date: 2018-04-20 11:19 From: "C.P.E. Rollins" cper2@cam.ac.uk To: Freesurfer freesurfer@nmr.mgh.harvard.edu
Thanks a lot for the explanation. The issue is that I don't think Freesurfer will run a design for which there are no subjects for a given class. So if I keep the classes (24 classes since 6 centres x 2 gender x 2 groups (patient vs. control), and have the contrast with 12 (-1) and 12 (1), -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1 1 1 1 1 1 1 1 1 0 I get the 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.
I'm assuming this is because there are no participants in, for example, patient_centre2_female Is there any way to get around this issue, or should I remove those centres from my analysis (since I can't only remove "patient_centre2_female", as this would make the contrast unbalanced (positives and negatives would not add to the same number). I hope this was clear but please let me know if it was not.
Thanks again, Colleen