External Email - Use Caution
Hi experts,
I am running following command:
mri_glmfit.bin --y SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol.8.mgh --fsgd PAI_SUI_ONLY_MILD_DEP2.fsgd dods --C Corr-PAI-SUI_ONLY_DEP.mtx --surf fsaverage lh --cortex --glmdir SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol8.glmdir
Both FSGD and Contrast files are also attached.
After I run above command, I get following error (highlighted in red):
mri_glmfit --y SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol.8.mgh --fsgd PAI_SUI_ONLY_MILD_DEP2.fsgd dods --C Corr-PAI-SUI_ONLY_DEP.mtx --surf fsaverage lh --cortex --glmdir SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol8.glmdir
gdfRead(): reading PAI_SUI_ONLY_MILD_DEP2.fsgd
INFO: demeaning continuous variables
Continuous Variable Means (all subjects)
0 SUI 52.4444 2.11403
1 Age 32.1111 7.23332
2 DEP 55.1111 7.83786
Class Means of each Continuous Variable
1 Male 52.0000 32.0000 54.0000
2 Female 52.6667 32.1667 55.6667
INFO: gd2mtx_method is dods
Reading source surface /data/emot/Freesurfer/FreeSurferSegmentation/SB_Behavioral/fsaverage/surf/lh.white
Number of vertices 163842
Number of faces 327680
Total area 65416.984375
AvgVtxArea 0.399269
AvgVtxDist 0.721953
StdVtxDist 0.195470
$Id: mri_glmfit.c,v 1.241.2.4 2016/12/08 22:02:40 zkaufman Exp $
cwd /data/emot/Freesurfer/FreeSurferSegmentation/SB_Behavioral
cmdline mri_glmfit.bin --y SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol.8.mgh --fsgd PAI_SUI_ONLY_MILD_DEP2.fsgd dods --C Corr-PAI-SUI_ONLY_DEP.mtx --surf fsaverage lh --cortex --glmdir SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol8.glmdir
sysname Linux
hostname hyperion
machine x86_64
user scanlab
FixVertexAreaFlag = 1
UseMaskWithSmoothing 1
OneSampleGroupMean 0
y /data/emot/Freesurfer/FreeSurferSegmentation/SB_Behavioral/SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol.8.mgh
logyflag 0
usedti 0
FSGD PAI_SUI_ONLY_MILD_DEP2.fsgd
labelmask /data/emot/Freesurfer/FreeSurferSegmentation/SB_Behavioral/fsaverage/label/lh.cortex.label
maskinv 0
glmdir SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol8.glmdir
IllCondOK 0
ReScaleX 1
DoFFx 0
Creating output directory SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol8.glmdir
Loading y from /data/emot/Freesurfer/FreeSurferSegmentation/SB_Behavioral/SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol.8.mgh
... done reading.
INFO: gd2mtx_method is dods
Saving design matrix to SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol8.glmdir/Xg.dat
Computing normalized matrix
Normalized matrix condition is 1e+08
Design matrix ------------------
0.00000 1.00000 0.00000 1.68188 0.00000 -0.01536 0.00000 -1.67279;
1.00000 0.00000 -0.68326 0.00000 1.09063 0.00000 -0.77969 0.00000;
0.00000 1.00000 0.00000 -0.68326 0.00000 -1.39785 0.00000 2.02720;
0.00000 1.00000 0.00000 -0.68326 0.00000 1.22888 0.00000 -0.14176;
1.00000 0.00000 0.73582 0.00000 -0.29186 0.00000 0.62375 0.00000;
1.00000 0.00000 -0.68326 0.00000 -0.84486 0.00000 -0.26935 0.00000;
0.00000 1.00000 0.00000 1.68188 0.00000 -1.12135 0.00000 0.11341;
0.00000 1.00000 0.00000 -0.68326 0.00000 1.50538 0.00000 -0.65211;
0.00000 1.00000 0.00000 -0.68326 0.00000 -0.15361 0.00000 0.75134;
--------------------------------
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 SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol.8.mgh --fsgd PAI_SUI_ONLY_MILD_DEP2.fsgd dods --C Corr-PAI-SUI_ONLY_DEP.mtx --surf fsaverage lh --cortex --glmdir SB_SUI_ONLY_MILD_DEP2/lh.SUI_vol8.glmdir
2. The FSGD file (if using one)
3. And the design matrix above
Attempting to diagnose further
SumSq: Min=1.034183 (col 7), Max=2.816108 (col 8)
... could not determine the cause of the problem
Both fsgd and contrast files are created in LINUX, and I also normalized my data using DeMeanFlag 1
ReScaleFlag 1 as suggested here: https://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg49123.html
Could you please help in resolving this issue?
Thanks !
Sahil