It could be a couple of things: 1. Your covariates have a large scale. You can try demeaning and dividing by the stddev to bring them into the range of 1.0 2. It looks like your covariates are highly correlated. You may need to reduce them or do a component analysis.
On 5/30/19 12:53 PM, Martin Juneja wrote:
External Email - Use Caution
Dear Douglas, Bruce (and FS experts),
I was just wondering if you got a chance to look at the Xg.dat file attached in my previous email. I would really appreciate any help in resolving the following error I get after I run mri_glmfit command:
ERROR: matrix is ill-conditioned or badly scaled, condno = 385372
Possible problem with experimental design:
Check for duplicate entries and/or lack of range of
continuous variables within a class.
Thanks.
On Fri, May 24, 2019 at 9:42 AM Martin Juneja <mj70481@gmail.com mailto:mj70481@gmail.com> wrote:
Hi Dough, Please find the Xg.dat file attached here. Thanks. On Fri, May 24, 2019 at 7:30 AM Greve, Douglas N.,Ph.D. <DGREVE@mgh.harvard.edu <mailto:DGREVE@mgh.harvard.edu>> wrote: Can you send the Xg.dat file? On 5/23/2019 8:43 PM, Martin Juneja wrote:External Email - Use Caution Hi FS experts, I want to find clusters showing an association between cortical measures and behavior x6 (sex, age and variables x1-x5 as covariates). My FSGD file looks like: GroupDescriptorFile 1 Class Male Class Female Variables x1 x2 x3 x4 x5 x6 Age InputS001Male1201031091241139411336 InputS003Male9110186103109959629 InputS004Male1041161109611510811239 InputS005Female11611410312711510611832 InputS006Female9310090100117909723 InputS008Male7080806678837140 InputS009Female12511411512511810412445 InputS010Female96919310191929442 InputS011Male1049297111841039845 InputS012Female7981907482807822 InputS014Male120941181141058710744 InputS015Female94869210080938829 InputS016Male98921078895909533 InputS017Female9610011283951049921 InputS018Male989396103104879642 InputS019Female1051209911011311411426 InputS020Female10510710410311310010927 InputS022Female116981151111118910925 InputS023Female98919410586969421 InputS024Male1161011191059011211143 InputS025Female1431021341291109512021 InputS028Female7280777182787141 InputS029Male104100116929410510425 InputS030Female120961151161039210918 InputS031Female1141031111119910511229 InputS033Male121991081271049611120 InputS035Male11210410411510310311020 InputS036Female1111031031131069810928 InputS037Male1089110810392919930 InputS038Male103991021021178910318 InputS039Male1061141169411210711221 InputS040Male999897102941009941 InputS041Male1031099710911010310926 My contrast file looks like: 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.5 0.5 0 0 After I run mri_glmfit command, I get following error: ERROR: matrix is ill-conditioned or badly scaled, condno = 385372 -------------------------------- Possible problem with experimental design: Check for duplicate entries and/or lack of range of continuous variables within a class. I would really appreciate any help in figuring out this issue. Thanks, MJ _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer