Hi all, I try to run mri_glmfit, but it export "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 upload the data in the attachment, and the glmfit_output is the standard output of the "mri_glmfit". Thanks. All the best.
2013-12-19
Rujing Zha
Could be several things. You only have two IA_female and and only 3 con_female. Also, the age and edu are highly correlated. I would try creating a single IA group with males and females and a single con group doug
On 12/18/2013 11:21 AM, Rujing Zha wrote:
Hi all, I try to run mri_glmfit, but it export "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 upload the data in the attachment, and the glmfit_output is the standard output of the "mri_glmfit". Thanks. All the best. 2013-12-19
/Rujing Zha/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Hi Doug and others, Thanks Doug. I modifiled my number of class in the fsgd which can be got in the attachment. And I am confused with 3 problems: 1st: Is what I modified proper? 2nd: How can I deal with my high correlation between age and edu? 3rd: Is it necessary to tell mri_glmfit to distinguish which subject is female or male in each class if dividing all subjects into 2 classes (i.e con and IA )? Thanks in advance. All the best.
2013-12-19
Rujing Zha
发件人：Douglas N Greve greve@nmr.mgh.harvard.edu 发送时间：2013-12-19 04:59 主题：Re: [Freesurfer] mri_glmfit error 收件人："freesurfer"freesurfer@nmr.mgh.harvard.edu 抄送：
Could be several things. You only have two IA_female and and only 3 con_female. Also, the age and edu are highly correlated. I would try creating a single IA group with males and females and a single con group doug
On 12/18/2013 11:21 AM, Rujing Zha wrote:
Hi all, I try to run mri_glmfit, but it export "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 upload the data in the attachment, and the glmfit_output is the standard output of the "mri_glmfit". Thanks. All the best. 2013-12-19
/Rujing Zha/
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 greve@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.
On 12/18/13 10:26 PM, Rujing Zha wrote:
Hi Doug and others, Thanks Doug. I modifiled my number of class in the fsgd which can be got in the attachment. And I am confused with 3 problems: 1st: Is what I modified proper?
Yes
2nd: How can I deal with my high correlation between age and edu?
If they are nuisance variables, then you could orthogonalize one with respect to the other. If you really want to test for a difference in groups for the age slope or the edu slope, then there is nothing you can do except get more subjects.
3rd: Is it necessary to tell mri_glmfit to distinguish which subject is female or male in each class if dividing all subjects into 2 classes (i.e con and IA )?
If you only have con and IA, you cannot have a gender factor. If you think that a gender factor is really necessary, then you need to get more females. doug
Thanks in advance. All the best. 2013-12-19
/Rujing Zha/
*发件人：*Douglas N Greve greve@nmr.mgh.harvard.edu *发送时间：*2013-12-19 04:59 *主题：*Re: [Freesurfer] mri_glmfit error *收件人：*"freesurfer"freesurfer@nmr.mgh.harvard.edu *抄送：* Could be several things. You only have two IA_female and and only 3 con_female. Also, the age and edu are highly correlated. I would try creating a single IA group with males and females and a single con group doug On 12/18/2013 11:21 AM, Rujing Zha wrote:
Hi all, I try to run mri_glmfit, but it export "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 upload the data in the attachment, and the glmfit_output is the standard output of the "mri_glmfit". Thanks. All the best. 2013-12-19
/Rujing Zha/
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 greve@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.
Hi Doug and others, Thanks Doug. For the 2nd problem: I understand orthogonalization between nuisance variables. I just want to test the difference between two group mean pial thickness controling the age and edu (if it is necessary to control the age and sex). So I guess I have two option: one: I can select one of them discarding the other one in the fsgd. the other one: I select both of them, and ignoring the high correlation. For the 3rd problem: As I just want to know the difference between 2 group pial thickness controling some other factors which may explain my result, so I guess I just donot consider the sex as a nuisance factor. Is that appropriate? There is a new problem for me: I also want to do a partial analysis between pial thickness analysis and behavior score. The numbers of class in fsgd file are the same as this one except adding another colume for the score. Is this correct? Thanks. All the best.
2013-12-19
Rujing Zha
发件人：Douglas Greve greve@nmr.mgh.harvard.edu 发送时间：2013-12-19 11:34 主题：Re: [Freesurfer] mri_glmfit error 收件人："Rujing Zha"charujing123@163.com 抄送："freesurfer"freesurfer@nmr.mgh.harvard.edu
On 12/18/13 10:26 PM, Rujing Zha wrote:
Hi Doug and others, Thanks Doug. I modifiled my number of class in the fsgd which can be got in the attachment. And I am confused with 3 problems: 1st: Is what I modified proper? Yes
2nd: How can I deal with my high correlation between age and edu? If they are nuisance variables, then you could orthogonalize one with respect to the other. If you really want to test for a difference in groups for the age slope or the edu slope, then there is nothing you can do except get more subjects.
3rd: Is it necessary to tell mri_glmfit to distinguish which subject is female or male in each class if dividing all subjects into 2 classes (i.e con and IA )? If you only have con and IA, you cannot have a gender factor. If you think that a gender factor is really necessary, then you need to get more females. doug
Thanks in advance. All the best.
2013-12-19
Rujing Zha
发件人：Douglas N Greve greve@nmr.mgh.harvard.edu 发送时间：2013-12-19 04:59 主题：Re: [Freesurfer] mri_glmfit error 收件人："freesurfer"freesurfer@nmr.mgh.harvard.edu 抄送：
Could be several things. You only have two IA_female and and only 3 con_female. Also, the age and edu are highly correlated. I would try creating a single IA group with males and females and a single con group doug
On 12/18/2013 11:21 AM, Rujing Zha wrote:
Hi all, I try to run mri_glmfit, but it export "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 upload the data in the attachment, and the glmfit_output is the standard output of the "mri_glmfit". Thanks. All the best. 2013-12-19
/Rujing Zha/
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 greve@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.
On 12/18/2013 11:14 PM, Rujing Zha wrote:
Hi Doug and others, Thanks Doug. For the 2nd problem: I understand orthogonalization between nuisance variables. I just want to test the difference between two group mean pial thickness controling the age and edu (if it is necessary to control the age and sex). So I guess I have two option: one: I can select one of them discarding the other one in the fsgd. the other one: I select both of them, and ignoring the high correlation. For the 3rd problem: As I just want to know the difference between 2 group pial thickness controling some other factors which may explain my result, so I guess I just donot consider the sex as a nuisance factor. Is that appropriate?
I don't know. It is not a question for me to answer since it depends on the data.
There is a new problem for me: I also want to do a partial analysis between pial thickness analysis and behavior score. The numbers of class in fsgd file are the same as this one except adding another colume for the score. Is this correct?
Yes
Thanks. All the best. 2013-12-19
/Rujing Zha/
*发件人：*Douglas Greve greve@nmr.mgh.harvard.edu *发送时间：*2013-12-19 11:34 *主题：*Re: [Freesurfer] mri_glmfit error *收件人：*"Rujing Zha"charujing123@163.com *抄送：*"freesurfer"freesurfer@nmr.mgh.harvard.edu
On 12/18/13 10:26 PM, Rujing Zha wrote:
Hi Doug and others, Thanks Doug. I modifiled my number of class in the fsgd which can be got in the attachment. And I am confused with 3 problems: 1st: Is what I modified proper?
Yes
2nd: How can I deal with my high correlation between age and edu?
If they are nuisance variables, then you could orthogonalize one with respect to the other. If you really want to test for a difference in groups for the age slope or the edu slope, then there is nothing you can do except get more subjects.
3rd: Is it necessary to tell mri_glmfit to distinguish which subject is female or male in each class if dividing all subjects into 2 classes (i.e con and IA )?
If you only have con and IA, you cannot have a gender factor. If you think that a gender factor is really necessary, then you need to get more females. doug
Thanks in advance. All the best. 2013-12-19
/Rujing Zha/
*发件人：*Douglas N Greve greve@nmr.mgh.harvard.edu *发送时间：*2013-12-19 04:59 *主题：*Re: [Freesurfer] mri_glmfit error *收件人：*"freesurfer"freesurfer@nmr.mgh.harvard.edu *抄送：* Could be several things. You only have two IA_female and and only 3 con_female. Also, the age and edu are highly correlated. I would try creating a single IA group with males and females and a single con group doug On 12/18/2013 11:21 AM, Rujing Zha wrote:
Hi all, I try to run mri_glmfit, but it export "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 upload the data in the attachment, and the glmfit_output is the standard output of the "mri_glmfit". Thanks. All the best. 2013-12-19
/Rujing Zha/
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 greve@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.
Hi Doug and others, Thanks Doug. All the best.
2013-12-20
Rujing Zha
发件人：Douglas N Greve greve@nmr.mgh.harvard.edu 发送时间：2013-12-20 00:11 主题：Re: [Freesurfer] mri_glmfit error 收件人："Rujing Zha"charujing123@163.com 抄送："freesurfer"freesurfer@nmr.mgh.harvard.edu
On 12/18/2013 11:14 PM, Rujing Zha wrote:
Hi Doug and others, Thanks Doug. For the 2nd problem: I understand orthogonalization between nuisance variables. I just want to test the difference between two group mean pial thickness controling the age and edu (if it is necessary to control the age and sex). So I guess I have two option: one: I can select one of them discarding the other one in the fsgd. the other one: I select both of them, and ignoring the high correlation. For the 3rd problem: As I just want to know the difference between 2 group pial thickness controling some other factors which may explain my result, so I guess I just donot consider the sex as a nuisance factor. Is that appropriate?
I don't know. It is not a question for me to answer since it depends on the data.
There is a new problem for me: I also want to do a partial analysis between pial thickness analysis and behavior score. The numbers of class in fsgd file are the same as this one except adding another colume for the score. Is this correct?
Yes
Thanks. All the best. 2013-12-19
/Rujing Zha/
*发件人：*Douglas Greve greve@nmr.mgh.harvard.edu *发送时间：*2013-12-19 11:34 *主题：*Re: [Freesurfer] mri_glmfit error *收件人：*"Rujing Zha"charujing123@163.com *抄送：*"freesurfer"freesurfer@nmr.mgh.harvard.edu
On 12/18/13 10:26 PM, Rujing Zha wrote:
Hi Doug and others, Thanks Doug. I modifiled my number of class in the fsgd which can be got in the attachment. And I am confused with 3 problems: 1st: Is what I modified proper?
Yes
2nd: How can I deal with my high correlation between age and edu?
If they are nuisance variables, then you could orthogonalize one with respect to the other. If you really want to test for a difference in groups for the age slope or the edu slope, then there is nothing you can do except get more subjects.
3rd: Is it necessary to tell mri_glmfit to distinguish which subject is female or male in each class if dividing all subjects into 2 classes (i.e con and IA )?
If you only have con and IA, you cannot have a gender factor. If you think that a gender factor is really necessary, then you need to get more females. doug
Thanks in advance. All the best. 2013-12-19
/Rujing Zha/
*发件人：*Douglas N Greve greve@nmr.mgh.harvard.edu *发送时间：*2013-12-19 04:59 *主题：*Re: [Freesurfer] mri_glmfit error *收件人：*"freesurfer"freesurfer@nmr.mgh.harvard.edu *抄送：* Could be several things. You only have two IA_female and and only 3 con_female. Also, the age and edu are highly correlated. I would try creating a single IA group with males and females and a single con group doug On 12/18/2013 11:21 AM, Rujing Zha wrote:
Hi all, I try to run mri_glmfit, but it export "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 upload the data in the attachment, and the glmfit_output is the standard output of the "mri_glmfit". Thanks. All the best. 2013-12-19
/Rujing Zha/
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 greve@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.
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@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@nmr.mgh.harvard.edu
-
Douglas Greve -
Douglas N Greve
-
Rujing Zha