Hi Doug, Thanks! I actually thought I only had to demean variables that were not of interest. But this was indeed the problem, it now works!
Thank you very much! Best Heidi ================================================ Dr. Heidi Jacobs Postdoc researcher Faculty of Health, Medicine and Life Sciences School for Mental Health and Neurosciences Division Cognitive Neuropsychiatry and Clinical Neurosciences Alzheimer Center Limburg h.jac...@maastrichtuniversity.nl www.maastrichtuniversity.nl www.heidijacobs.nl Dr. Tanslaan 12, 6229 ET Maastricht P.O. Box 616, 6200 MD Maastricht, The Netherlands T +31 43 38 84 090 F +31 43 38 84 092 ================================================ ________________________________________ Van: Douglas N Greve [gr...@nmr.mgh.harvard.edu] Verzonden: dinsdag 27 augustus 2013 18:27 Aan: Jacobs H (NP) CC: freesurfer@nmr.mgh.harvard.edu Onderwerp: Re: matrix ill-conditioned.... Hi Heidi, the problem is being caused by the interaction column. Try demeaning the ed_log and MMSE before multiplying them together to get the interaction. Do the demeaning on the entire data set, not on a group by group basis. doug On 08/27/2013 02:04 AM, Jacobs H (NP) wrote: > Sure, here it is. > Thanks > Heidi > ================================================ > Dr. Heidi Jacobs > Postdoc researcher > Faculty of Health, Medicine and Life Sciences > School for Mental Health and Neurosciences > Division Cognitive Neuropsychiatry and Clinical Neurosciences > Alzheimer Center Limburg > h.jac...@maastrichtuniversity.nl > www.maastrichtuniversity.nl > www.heidijacobs.nl > > Dr. Tanslaan 12, 6229 ET Maastricht > P.O. Box 616, 6200 MD Maastricht, The Netherlands > T +31 43 38 84 090 F +31 43 38 84 092 > ================================================ > ________________________________________ > Van: Douglas N Greve [gr...@nmr.mgh.harvard.edu] > Verzonden: maandag 26 augustus 2013 23:53 > Aan: Jacobs H (NP) > CC: freesurfer@nmr.mgh.harvard.edu > Onderwerp: Re: matrix ill-conditioned.... > > Can you send me the design matrix (Xg.dat) and fsgd file? > > > On 08/26/2013 04:35 PM, Jacobs H (NP) wrote: >> Dear Doug, >> >> Thank you. I tried to run it with this version of mri_glmfit (saved as >> mri_glmfit2), but it still gives me the same error message. >> I don't see the problem. I already divided the column with the interaction >> by 10. >> Fsgd files without MMSE and interaction factors, did not gave a problem... >> Could you have a look at it? I will attach the fsgd and contrast file. >> >> Thanks! I really appreciate your help! >> Heidi >> >> >> -------------------------------- >> ERROR: matrix is ill-conditioned or badly scaled, condno = 101281 >> -------------------------------- >> 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_glmfit2 --y >> rh.Alls_educlog_MMSE_inter_agecent_ICVcent_fsgd.sm20.mgh --fsgd >> /media/Data_disk/Nieuwe_map/FreeSurferOutputAD/FreeSurferOutputv450/FSGD/fsgd.Alls_educlog_MMSE_inter_agecent_ICVcent_fsgd.txt >> dods --glmdir >> rh.Alls_educlog_MMSE_inter_agecent_ICVcent_fsgd.dods.20.glmdir --surf >> fsaverage rh --C >> /media/Data_disk/Nieuwe_map/FreeSurferOutputAD/FreeSurferOutputv450/GLM_CONTRAST.MAT/contrast_3groups_Finteraction.mat >> 2. The FSGD file (if using one) >> 3. And the design matrix above >> >> >> ================================================ >> Dr. Heidi Jacobs >> Postdoc researcher >> Faculty of Health, Medicine and Life Sciences >> School for Mental Health and Neurosciences >> Division Cognitive Neuropsychiatry and Clinical Neurosciences >> Alzheimer Center Limburg >> h.jac...@maastrichtuniversity.nl >> www.maastrichtuniversity.nl >> www.heidijacobs.nl >> >> Dr. Tanslaan 12, 6229 ET Maastricht >> P.O. Box 616, 6200 MD Maastricht, The Netherlands >> T +31 43 38 84 090 F +31 43 38 84 092 >> ================================================ >> ________________________________________ >> Van: Douglas N Greve [gr...@nmr.mgh.harvard.edu] >> Verzonden: maandag 26 augustus 2013 21:47 >> Aan: Jacobs H (NP) >> CC: freesurfer@nmr.mgh.harvard.edu >> Onderwerp: Re: matrix ill-conditioned.... >> >> You can try with this version of mri_glmfit which will do the rescaling >> for you >> ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/mri_glmfit >> >> If the problem persists then it may be an issue with your design >> >> doug >> >> >> On 08/26/2013 01:42 PM, Jacobs H (NP) wrote: >>> Hi Doug, >>> >>> I divided everything by 10, but the problem remains. >>> I have been searching a bit more. Could it be that the problem is that the >>> range in cognitive scores in the control group is very small? >>> When I run the groups separately, the patients are not a problem (the range >>> in cognition is very large), but the controls are still the problem. >>> >>> Thanks! >>> Heidi >>> ================================================ >>> Dr. Heidi Jacobs >>> Postdoc researcher >>> Faculty of Health, Medicine and Life Sciences >>> School for Mental Health and Neurosciences >>> Division Cognitive Neuropsychiatry and Clinical Neurosciences >>> Alzheimer Center Limburg >>> h.jac...@maastrichtuniversity.nl >>> www.maastrichtuniversity.nl >>> www.heidijacobs.nl >>> >>> Dr. Tanslaan 12, 6229 ET Maastricht >>> P.O. Box 616, 6200 MD Maastricht, The Netherlands >>> T +31 43 38 84 090 F +31 43 38 84 092 >>> ================================================ >>> ________________________________________ >>> Van: Douglas N Greve [gr...@nmr.mgh.harvard.edu] >>> Verzonden: vrijdag 23 augustus 2013 17:31 >>> Aan: Jacobs H (NP) >>> CC: freesurfer@nmr.mgh.harvard.edu >>> Onderwerp: Re: [Freesurfer] FW: FW: Contrast and F-test >>> >>> The values in the interaction column are probably much larger than the >>> values in the other columns (thus they are "badly scaled"). You can >>> divide them by a number to bring their scale down to that of the other >>> columns without changing the p-values. Version 5.x does this automatically. >>> >>> >>> >>> >>> >>> On 08/23/2013 11:25 AM, Jacobs H (NP) wrote: >>>> I am using version 4.5, as all the old analyses were also done with this >>>> version. >>>> What do you mean with rescale? >>>> THanks >>>> Heidi >>>> ________________________________________ >>>> Van: freesurfer-boun...@nmr.mgh.harvard.edu >>>> [freesurfer-boun...@nmr.mgh.harvard.edu] namens Douglas N Greve >>>> [gr...@nmr.mgh.harvard.edu] >>>> Verzonden: vrijdag 23 augustus 2013 17:23 >>>> Aan: freesurfer@nmr.mgh.harvard.edu >>>> Onderwerp: Re: [Freesurfer] FW: FW: Contrast and F-test >>>> >>>> what version of FS are you using? It may be that you just need to >>>> rescale that column (if the scale is much more than the other columns) >>>> >>>> On 08/23/2013 11:12 AM, Jacobs H (NP) wrote: >>>>> Hi again! >>>>> >>>>> I am trying to run these contrasts but FreeSurfer tells me that "ERROR: >>>>> matrix is ill-conditioned or badly scaled, condno = 101380 >>>>> Possible problem with experimental design: >>>>> Check for duplicate entries and/or lack of range of >>>>> continuous variables within a class." >>>>> >>>>> Could this be due to the fact that there is a high correlation between >>>>> cognition, education and cognition x education? (the product is >>>>> correlated with the factors of that product) >>>>> Or is something else going on? >>>>> How can I solve this? >>>>> >>>>> Thanks >>>>> Heidi >>>>> >>>>> ________________________________________ >>>>> Van: freesurfer-boun...@nmr.mgh.harvard.edu >>>>> [freesurfer-boun...@nmr.mgh.harvard.edu] namens Douglas N Greve >>>>> [gr...@nmr.mgh.harvard.edu] >>>>> Verzonden: donderdag 22 augustus 2013 17:52 >>>>> Aan: freesurfer@nmr.mgh.harvard.edu >>>>> Onderwerp: Re: [Freesurfer] FW: FW: Contrast and F-test >>>>> >>>>> both of those contrasts are correct >>>>> doug >>>>> >>>>> On 08/22/2013 10:51 AM, Jacobs H (NP) wrote: >>>>>> Arrghh, again wrong email, sorry.... >>>>>> >>>>>> Thanks. I think I understand the rational behind it. >>>>>> Just to be sure, would this be correct: >>>>>> >>>>>> For each group separately >>>>>> 0 0 0 1 0 0 (group - cognition - education - cognition x education - >>>>>> age - ICV) >>>>>> >>>>>> For the F-test (same order) >>>>>> 0 0 0 0 0 0 0 0 0 1 -1 0 0 0 0 0 0 0 >>>>>> 0 0 0 0 0 0 0 0 0 1 0 -1 0 0 0 0 0 0 >>>>>> >>>>>> Heidi >>>>>> >>>>>> >>>>>> ________________________________________ >>>>>> Van: freesurfer-boun...@nmr.mgh.harvard.edu >>>>>> [freesurfer-boun...@nmr.mgh.harvard.edu] namens Douglas N Greve >>>>>> [gr...@nmr.mgh.harvard.edu] >>>>>> Verzonden: donderdag 22 augustus 2013 16:46 >>>>>> Aan: freesurfer@nmr.mgh.harvard.edu >>>>>> Onderwerp: Re: [Freesurfer] FW: Contrast and F-test >>>>>> >>>>>> you need to have all 3. If you want to examine the interaction, then you >>>>>> need a new contrast to test the interaction. Set all to 0 except for the >>>>>> interaction term >>>>>> doug >>>>>> >>>>>> >>>>>> On 08/22/2013 10:42 AM, Jacobs H (NP) wrote: >>>>>>> Dear Doug, >>>>>>> Thank you for the information. Sorry for not sending this to the list. >>>>>>> Do you mean that I then have to include the three variables (covariate >>>>>>> 1, covariate 2, product of these two) or only the product in the fsgd >>>>>>> file? >>>>>>> If I need to include all three of them, can you advise me on the >>>>>>> contrasts? >>>>>>> If it is only the product, I think I can use the same contrasts as >>>>>>> before, right? >>>>>>> >>>>>>> Thanks >>>>>>> Heidi >>>>>>> ================================================ >>>>>>> Dr. Heidi Jacobs >>>>>>> Postdoc researcher >>>>>>> Faculty of Health, Medicine and Life Sciences >>>>>>> School for Mental Health and Neurosciences >>>>>>> Division Cognitive Neuropsychiatry and Clinical Neurosciences >>>>>>> Alzheimer Center Limburg >>>>>>> h.jac...@maastrichtuniversity.nl >>>>>>> www.maastrichtuniversity.nl >>>>>>> www.heidijacobs.nl >>>>>>> >>>>>>> Dr. Tanslaan 12, 6229 ET Maastricht >>>>>>> P.O. Box 616, 6200 MD Maastricht, The Netherlands >>>>>>> T +31 43 38 84 090 F +31 43 38 84 092 >>>>>>> ================================================ >>>>>>> ________________________________________ >>>>>>> Van: freesurfer-boun...@nmr.mgh.harvard.edu >>>>>>> [freesurfer-boun...@nmr.mgh.harvard.edu] namens Douglas N Greve >>>>>>> [gr...@nmr.mgh.harvard.edu] >>>>>>> Verzonden: donderdag 22 augustus 2013 16:34 >>>>>>> Aan: freesurfer@nmr.mgh.harvard.edu >>>>>>> Onderwerp: Re: [Freesurfer] Contrast and F-test >>>>>>> >>>>>>> If you want to test for an interaction between continuous variables, >>>>>>> then you should create a new regressor by multiplying them together then >>>>>>> test that regressor >>>>>>> doug >>>>>>> >>>>>>> >>>>>>> On 08/22/2013 08:09 AM, Jacobs H (NP) wrote: >>>>>>>> Hi all, >>>>>>>> >>>>>>>> I my manuscript, I examined the association between cognition and >>>>>>>> cortical thickness. I first did this in three groups separately and >>>>>>>> then ran a F-test. >>>>>>>> One of the reviewers, now wants me to examine the interaction between >>>>>>>> cognition and education (all continuous). >>>>>>>> >>>>>>>> Would this be the correct contrast for each group separately (last two >>>>>>>> zeros are for covariates, also continuous) >>>>>>>> >>>>>>>> 0 1 -1 0 0 >>>>>>>> >>>>>>>> Would this be the correct way for the F-test: >>>>>>>> >>>>>>>> 0 0 0 0 0 1 -1 0 -1 1 0 0 0 0 0 0 0 >>>>>>>> 0 0 0 1 0 -1 -1 0 1 0 0 0 0 0 0 >>>>>>>> >>>>>>>> Thanks! >>>>>>>> >>>>>>>> Heidi >>>>>>>> >>>>>>>> ================================================ >>>>>>>> Dr. Heidi Jacobs >>>>>>>> Postdoc researcher >>>>>>>> Faculty of Health, Medicine and Life Sciences >>>>>>>> School for Mental Health and Neurosciences >>>>>>>> Division Cognitive Neuropsychiatry and Clinical Neurosciences >>>>>>>> Alzheimer Center Limburg >>>>>>>> h.jac...@maastrichtuniversity.nl >>>>>>>> www.maastrichtuniversity.nl >>>>>>>> www.heidijacobs.nl >>>>>>>> >>>>>>>> Dr. Tanslaan 12, 6229 ET Maastricht >>>>>>>> P.O. Box 616, 6200 MD Maastricht, The Netherlands >>>>>>>> T +31 43 38 84 090 F +31 43 38 84 092 >>>>>>>> ================================================ >>>>>>>> _______________________________________________ >>>>>>>> Freesurfer mailing list >>>>>>>> Freesurfer@nmr.mgh.harvard.edu >>>>>>>> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >>>>>>>> >>>>>>>> >>>>>>> -- >>>>>>> Douglas N. Greve, Ph.D. >>>>>>> MGH-NMR Center >>>>>>> gr...@nmr.mgh.harvard.edu >>>>>>> Phone Number: 617-724-2358 >>>>>>> Fax: 617-726-7422 >>>>>>> >>>>>>> Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting >>>>>>> FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 >>>>>>> www.nmr.mgh.harvard.edu/facility/filedrop/index.html >>>>>>> Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Freesurfer mailing list >>>>>>> Freesurfer@nmr.mgh.harvard.edu >>>>>>> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >>>>>>> >>>>>>> >>>>>>> The information in this e-mail is intended only for the person to whom >>>>>>> it is >>>>>>> addressed. If you believe this e-mail was sent to you in error and the >>>>>>> e-mail >>>>>>> contains patient information, please contact the Partners Compliance >>>>>>> HelpLine at >>>>>>> http://www.partners.org/complianceline . If the e-mail was sent to you >>>>>>> in error >>>>>>> but does not contain patient information, please contact the sender and >>>>>>> properly >>>>>>> dispose of the e-mail. >>>>>>> >>>>>>> >>>>>>> _______________________________________________ >>>>>>> Freesurfer mailing list >>>>>>> Freesurfer@nmr.mgh.harvard.edu >>>>>>> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >>>>>>> >>>>>>> >>>>>> -- >>>>>> Douglas N. Greve, Ph.D. >>>>>> MGH-NMR Center >>>>>> gr...@nmr.mgh.harvard.edu >>>>>> Phone Number: 617-724-2358 >>>>>> Fax: 617-726-7422 >>>>>> >>>>>> Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting >>>>>> FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 >>>>>> www.nmr.mgh.harvard.edu/facility/filedrop/index.html >>>>>> Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ >>>>>> >>>>>> _______________________________________________ >>>>>> Freesurfer mailing list >>>>>> 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 >>>>>> >>>>>> >>>>> -- >>>>> Douglas N. Greve, Ph.D. >>>>> MGH-NMR Center >>>>> gr...@nmr.mgh.harvard.edu >>>>> Phone Number: 617-724-2358 >>>>> Fax: 617-726-7422 >>>>> >>>>> Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting >>>>> FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 >>>>> www.nmr.mgh.harvard.edu/facility/filedrop/index.html >>>>> Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ >>>>> >>>>> _______________________________________________ >>>>> Freesurfer mailing list >>>>> 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 >>>>> >>>>> >>>> -- >>>> Douglas N. Greve, Ph.D. >>>> MGH-NMR Center >>>> gr...@nmr.mgh.harvard.edu >>>> Phone Number: 617-724-2358 >>>> Fax: 617-726-7422 >>>> >>>> Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting >>>> FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 >>>> www.nmr.mgh.harvard.edu/facility/filedrop/index.html >>>> Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ >>>> >>>> _______________________________________________ >>>> Freesurfer mailing list >>>> Freesurfer@nmr.mgh.harvard.edu >>>> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >>>> >>>> >>> -- >>> Douglas N. Greve, Ph.D. >>> MGH-NMR Center >>> gr...@nmr.mgh.harvard.edu >>> Phone Number: 617-724-2358 >>> Fax: 617-726-7422 >>> >>> Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting >>> FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 >>> www.nmr.mgh.harvard.edu/facility/filedrop/index.html >>> Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ >>> >>> >>> >> -- >> Douglas N. Greve, Ph.D. >> MGH-NMR Center >> gr...@nmr.mgh.harvard.edu >> Phone Number: 617-724-2358 >> Fax: 617-726-7422 >> >> Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting >> FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 >> www.nmr.mgh.harvard.edu/facility/filedrop/index.html >> Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ >> > -- > Douglas N. Greve, Ph.D. > MGH-NMR Center > gr...@nmr.mgh.harvard.edu > Phone Number: 617-724-2358 > Fax: 617-726-7422 > > Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting > FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 > www.nmr.mgh.harvard.edu/facility/filedrop/index.html > Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ -- Douglas N. Greve, Ph.D. MGH-NMR Center gr...@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422 Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 www.nmr.mgh.harvard.edu/facility/filedrop/index.html Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
