Hi Doug,
This is the complete terminal output:
cwd /SA
cmdline mri_glmfit --y /FREESURFER/lh.thickness.15.mgh --fsgd SA_base.fsgd --C /FREESURFER/con4.mtx --surf fsaverage lh --cortex --glmdir /FREESURFER/lh.pair-thick-T1-00.glmdir --pvr /FREESURFER/lh.paired-cope1.psc.00.mgh
sysname Linux
hostname einstein
machine x86_64
user pgautam
FixVertexAreaFlag = 1
UseMaskWithSmoothing 1
OneSampleGroupMean 0
y /FREESURFER/lh.thickness.15.mgh
logyflag 0
usedti 0
FSGD SA_base.fsgd
labelmask /fsaverage/label/lh.cortex.label
maskinv 0
glmdir /FREESURFER/lh.pair-thick-T1-00.glmdir
IllCondOK 0
ReScaleX 1
DoFFx 0
Creating output directory /FREESURFER/lh.pair-thick-T1-00.glmdir
Loading y from /FREESURFER/lh.thickness.15.mgh
INFO: gd2mtx_method is dods
Saving design matrix to /FREESURFER/lh.pair-thick-T1-00.glmdir/Xg.dat
Normalized matrix condition is 1.30805
Matrix condition is 6.34773
Found 149955 points in label.
Pruning voxels by thr: 0.000000
Found 149953 voxels in mask
Saving mask to /FREESURFER/lh.pair-thick-T1-00.glmdir/mask.mgh
Reshaping mriglm->mask...
search space = 74612.446501
DOF = 54
Starting fit and test
Matrix is Ill-conditioned
1.000 0.000 1.029 0.000 0.000;
1.000 0.000 0.838 0.000 0.000;
1.000 0.000 0.318 0.000 0.000;
1.000 0.000 0.561 0.000 0.000;
1.000 0.000 0.622 0.000 0.000;
1.000 0.000 0.882 0.000 0.000;
1.000 0.000 2.059 0.000 0.000;
1.000 0.000 0.945 0.000 0.000;
1.000 0.000 1.090 0.000 0.000;
1.000 0.000 1.245 0.000 0.000;
1.000 0.000 0.571 0.000 0.000;
1.000 0.000 0.337 0.000 0.000;
1.000 0.000 1.105 0.000 0.000;
1.000 0.000 0.990 0.000 0.000;
1.000 0.000 1.074 0.000 0.000;
1.000 0.000 1.552 0.000 0.000;
1.000 0.000 1.020 0.000 0.000;
1.000 0.000 0.829 0.000 0.000;
1.000 0.000 0.522 0.000 0.000;
1.000 0.000 1.152 0.000 0.000;
1.000 0.000 0.429 0.000 0.000;
1.000 0.000 1.003 0.000 0.000;
1.000 0.000 0.515 0.000 0.000;
1.000 0.000 1.500 0.000 0.000;
1.000 0.000 -4.639 0.000 0.000;
1.000 0.000 -5.227 0.000 0.000;
1.000 0.000 -5.051 0.000 0.000;
1.000 0.000 -4.918 0.000 0.000;
1.000 0.000 -5.316 0.000 0.000;
1.000 0.000 -5.311 0.000 0.000;
0.000 1.000 0.000 2.172 0.000;
0.000 1.000 0.000 1.115 0.000;
0.000 1.000 0.000 1.056 0.000;
0.000 1.000 0.000 1.372 0.000;
0.000 1.000 0.000 1.267 0.000;
0.000 1.000 0.000 0.847 0.000;
0.000 1.000 0.000 0.682 0.000;
0.000 1.000 0.000 0.524 0.000;
0.000 1.000 0.000 0.682 0.000;
0.000 1.000 0.000 0.930 0.000;
0.000 1.000 0.000 0.126 0.000;
0.000 1.000 0.000 1.113 0.000;
0.000 1.000 0.000 1.066 0.000;
0.000 1.000 0.000 0.517 0.000;
0.000 1.000 0.000 0.598 0.000;
0.000 1.000 0.000 2.061 0.000;
0.000 1.000 0.000 2.003 0.000;
0.000 1.000 0.000 1.204 0.000;
0.000 1.000 0.000 0.975 0.000;
0.000 1.000 0.000 1.445 0.000;
0.000 1.000 0.000 1.327 0.000;
0.000 1.000 0.000 1.731 0.000;
0.000 1.000 0.000 1.610 0.000;
0.000 1.000 0.000 1.145 0.000;
0.000 1.000 0.000 0.102 0.000;
0.000 1.000 0.000 -4.364 0.000;
0.000 1.000 0.000 -5.073 0.000;
0.000 1.000 0.000 -4.607 0.000;
0.000 1.000 0.000 -5.355 0.000;
Matrix is Ill-conditioned
..
------------------------------
Message: 8
Date: Thu, 13 Jun 2013 13:21:00 -0400
Subject: Re: [Freesurfer] pvr help
Content-Type: text/plain; charset="iso-8859-1"
Hi Prapti, can you send the full terminal output? The ill-conditioned
error is probably not related to the contrast. It is usually related to
the design matrix, which makes it difficult to track down because you
will have a different one at each vertex with --pvr.
doug
On 6/13/13 1:14 PM, Gautam, Prapti wrote:
Dear FreeSurfers
I am trying to control for thickness in another surface using the pvr
option in mri_glmfit. However, I'm confused as to how I can include
the pvr option correctly in the contrast files.
I have two groups, and would like to utilize DODS for this analyses as
I think that my groups both have different intercepts and slopes.
This is my fsgd set up :
GroupDescriptorFile1
TitleMyTitle
ClassControls
CLASS FA
VariablesAge
InputPMS1Controls14.442163
InputPMS2Controls14.354552
..
InputPMS7FA15.726215
InputPMS28FA14.45859
This is the command I run:
mri_glmfit --y lh.surface.mgh --fsgd SA_age.fsgd --C con.mtx --surf
fsaverage lh --cortex --glmdir test.glmdir --pvr lh.thickness.psc.mgh
--pvr lh.thickness2.psc.mgh
The contrast I would like to run is: 0 0 0 0 1 -1 so as to look at
group interactions within thickness after controlling for another
cortical surface. However, I can't get this to run as there is a
"Matrix ill-conditioned" error.
The contrast that does run is 0 0 0 0 1 (after passing only one pvr
file). But this option seems to be neither dods nor doss.
Is there a correct way to do a DODS analyses looking for interactions
using the pvr option?
Thank you,
Prapti