Dear FS experts.I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out.ThanksPablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh'); %---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------% 3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Here I am getting the following problems:%-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);Error using lme_mass_fit (line 108)The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the designmatrix X Error in lme_mass_fit_vw (line 73)[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users,I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families.Thanks in advance!Pablo From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi everybody,
I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem:
1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17].
2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one.
3. I could use a mixed effect model approach and specify site as a random effect.
If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches?
I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows:
Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ...
As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output:
stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex);
Starting matlabpool using the 'local' profile ... connected to 8 workers.
Starting model fitting at each location ...
Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ...
I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence.
Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate?
Thanks, Michael
_______________________________________________ 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
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
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 Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data)
see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------% lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex); Error using lme_mass_fit (line 108) The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design matrix X Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); My matrix is organised in "family", "group", Sex" and "age" columns".4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset Hi Pablo I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data. Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family. When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects. Hope that helps Cheers -Jorge
------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 10 de septiembre de 2015 8:07 *Asunto:* [Freesurfer] A mixed effect model approach in within subject dataset Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) <http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels%29> but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael _______________________________________________ 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 <mailto: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 <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 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
Thank you Martin.I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue?Just in case I am also including a snapshot of my loaded variables at the bottom.Many thanks,Pablo>> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Warning: matlabpool will be removed in a future release.To query the size of an already started parallel pool, query the 'NumWorkers'property of the pool.To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release.To query the size of an already started parallel pool, query the 'NumWorkers'property of the pool.To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release.Use parpool instead. Starting matlabpool using the 'local' profile ... Error using matlabpool (line 148)Failed to start a parallel pool. (For information in addition to the causing error,validate the profile 'local' in the Cluster Profile Manager.) Error in lme_mass_fit (line 128) matlabpool(prs); Error in lme_mass_fit_vw (line 73)[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); Caused by: Error using parallel.internal.pool.InteractiveClient/start (line 330) Failed to start pool. Error using parallel.Job/submit (line 304) You requested a minimum of 8 workers, but the cluster "local" has the NumWorkers property set to allow a maximum of 4 workers. To run a communicating job on more workers than this (up to a maximum of 512 for the Local cluster), increase the value of the NumWorkers property for the cluster. The default value of NumWorkers for a Local cluster is the number of cores on the local machine. To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see
https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.:
lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label');
2-Read the data file eg.:
[lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.:
lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users,I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families.Thanks in advance!Pablo From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi everybody,
I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem:
1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17].
2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one.
3. I could use a mixed effect model approach and specify site as a random effect.
If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches?
I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows:
Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ...
As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output:
stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex);
Starting matlabpool using the 'local' profile ... connected to 8 workers.
Starting model fitting at each location ...
Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ...
I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence.
Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate?
Thanks, Michael
_______________________________________________ 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
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
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
Hi Pablo,
looks like on your system only 4 parallel workers are allowed, while the code asks for 8. Probably there is an easy way to specify that it should only use 4. Maybe Jorge or someone else who knows parallel matlab programming can answer.
Best, Martin
On 10/15/2015 07:18 AM, pablo najt wrote:
Thank you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just in case I am also including a snapshot of my loaded variables at the bottom. Many thanks, Pablo
lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. Use parpool instead. Starting matlabpool using the 'local' profile ... Error using matlabpool (line 148) Failed to start a parallel pool. (For information in addition to the causing error, validate the profile 'local' in the Cluster Profile Manager.) Error in lme_mass_fit (line 128) matlabpool(prs); Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); Caused by: Error using parallel.internal.pool.InteractiveClient/start (line 330) Failed to start pool. Error using parallel.Job/submit (line 304) You requested a minimum of 8 workers, but the cluster "local" has the NumWorkers property set to allow a maximum of 4 workers. To run a communicating job on more workers than this (up to a maximum of 512 for the Local cluster), increase the value of the NumWorkers property for the cluster. The default value of NumWorkers for a Local cluster is the number of cores on the local machine.
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh'); %---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------% 3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------% lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex); Error using lme_mass_fit (line 108) The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design matrix X Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig'); ------------------------------------------------------------------------ Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset Hi Pablo I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data. Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family. When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects. Hope that helps Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 10 de septiembre de 2015 8:07 *Asunto:* [Freesurfer] A mixed effect model approach in within subject dataset Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo From: michaelnotter@hotmail.com <mailto:michaelnotter@hotmail.com> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) <http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels%29> but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto: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 <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 <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu _______________________________________________ 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
Hi Pablo The error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4); Please make sure that sum(ni) and length(X) give the same value before running the above command. Cheers-Jorge
De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 15 de octubre de 2015 7:18 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
#yiv7790603773 #yiv7790603773 --.yiv7790603773hmmessage P{margin:0px;padding:0px;}#yiv7790603773 body.yiv7790603773hmmessage{font-size:12pt;font-family:Calibri;}#yiv7790603773 Thank you Martin.I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue?Just in case I am also including a snapshot of my loaded variables at the bottom.Many thanks,Pablo>> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
Use parpool instead.
Starting matlabpool using the 'local' profile ...
Error using matlabpool (line 148)
Failed to start a parallel pool. (For information in addition to the causing error,
validate the profile 'local' in the Cluster Profile Manager.)
Error in lme_mass_fit (line 128)
matlabpool(prs);
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
Caused by:
Error using parallel.internal.pool.InteractiveClient/start (line 330)
Failed to start pool.
Error using parallel.Job/submit (line 304)
You requested a minimum of 8 workers, but the cluster "local" has the
NumWorkers property set to allow a maximum of 4 workers. To run a
communicating job on more workers than this (up to a maximum of 512 for the
Local cluster), increase the value of the NumWorkers property for the
cluster. The default value of NumWorkers for a Local cluster is the number of
cores on the local machine.
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
#yiv7790603773 #yiv7790603773 --.yiv7790603773ExternalClass .yiv7790603773ecxhmmessage P {padding:0px;}#yiv7790603773 .yiv7790603773ExternalClass body.yiv7790603773ecxhmmessage {font-size:12pt;font-family:Calibri;}#yiv7790603773 Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.:CM.C = [your contrast matrix];F_lhstats = lme_mass_F(lhstats, CM);5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000From: jbernal0019@yahoo.esTo: freesurfer@nmr.mgh.harvard.eduSubject: Re: [Freesurfer] A mixed effect model approach in within subject datasetHi PabloI think you can useLME to analyze your data by ordering the rows of your design matrixappropriately. You can consider all subjects belonging to the samefamily as if they were a single subject in a longitudinal analysis.You can put in your design matrix all subjects belonging to family1first, then all subjects belonging to family 2 and so on. Then the'ni' required by lme_mass_fit_vw is a vector with the number ofsubjects in each family as its entries (ordered according to yourdesign matrix). So the length of the 'ni' vector is equal to thenumber of different families in your data. Now you can gofurther and additionally order the rows of your design matrix withineach family by age. This will allow you to test the effect of agewithin family. When choosing therandom effects for your statistical model remember that a randomeffect can only be the intercept term or any covariate that varieswithin family. For example you can compare a model with a singlerandom effect for the intercept term against the same model butconsidering both the intercept term and age as random effects.Hope that helpsCheers-Jorge De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset #yiv7790603773 #yiv7790603773 --.yiv7790603773ExternalClass #yiv7790603773ecxyiv8323390595 .yiv7790603773ecxyiv8323390595hmmessage P {padding:0px;}#yiv7790603773 .yiv7790603773ExternalClass #yiv7790603773ecxyiv8323390595 body.yiv7790603773ecxyiv8323390595hmmessage {font-size:12pt;font-family:Calibri;}#yiv7790603773 Dear Freesurfer users,I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families.Thanks in advance!PabloFrom: michaelnotter@hotmail.comTo: freesurfer@nmr.mgh.harvard.eduDate: Wed, 19 Feb 2014 13:10:09 +0100Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach#yiv7790603773 #yiv7790603773 --.yiv7790603773ExternalClass #yiv7790603773ecxyiv8323390595 .yiv7790603773ecxyiv8323390595ExternalClass .yiv7790603773ecxyiv8323390595ecxhmmessage P {padding:0px;}#yiv7790603773 .yiv7790603773ExternalClass #yiv7790603773ecxyiv8323390595 .yiv7790603773ecxyiv8323390595ExternalClass body.yiv7790603773ecxyiv8323390595ecxhmmessage {font-size:12pt;font-family:Calibri;}#yiv7790603773 Hi everybody,I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem:1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17].2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one.3. I could use a mixed effect model approach and specify site as a random effect.If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches?I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows:Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site41 1 0 0 11.1 99 0 0 1 01 0 1 0 11.1 101 0 0 1 01 1 0 0 11.4 95 1 0 0 01 0 0 1 12.4 100 1 0 0 0...As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output:>> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex);Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions.Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000.Location 62484: Index exceeds matrix dimensions.Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000....I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence.Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate?Thanks,Michael _______________________________________________Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurferThe information in this e-mail is intended only for the person to whom it isaddressed. If you believe this e-mail was sent to you in error and the e-mailcontains patient information, please contact the Partners Compliance HelpLine athttp://www.partners.org/complianceline . If the e-mail was sent to you in errorbut does not contain patient information, please contact the sender and properlydispose of the e-mail. _______________________________________________Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurferThe information in this e-mail is intended only for the person to whom it isaddressed. If you believe this e-mail was sent to you in error and the e-mailcontains patient information, please contact the Partners Compliance HelpLine athttp://www.partners.org/complianceline . If the e-mail was sent to you in errorbut does not contain patient information, please contact the sender and properlydispose of the e-mail. _______________________________________________Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurferThe information in this e-mail is intended only for the person to whom it isaddressed. If you believe this e-mail was sent to you in error and the e-mailcontains patient information, please contact the Partners Compliance HelpLine athttp://www.partners.org/complianceline . If the e-mail was sent to you in errorbut does not contain patient information, please contact the sender and properlydispose of the e-mail. _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Thanks you for the replies.Jorge and FS experts,I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects.The analysis gave the following 'error'(?) below:I looked up a previous thread coming across this. At that case you recommendedWould you recommend this afain> Aproximate percentage of fitted locations: 100%Warning: matlabpool will be removed in a future release.To query the size of an already started parallel pool, query the 'NumWorkers'property of the pool.To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release.To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead. Parallel pool using the 'local' profile is shutting down. Summary:Algorithm did not converge at 90.0637 percent of the total number of locations.Total elapsed time is 550.1023 minutes. Also almost all the time the screen showed the following message:144114: Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i.Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i.Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i.Location 144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448.Aproximate percentage of fitted locations: 100% So my two questions are:1. Is this problematic?2. Are there any fixes to this issue?Thank you,Pablo Date: Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo The error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4); Please make sure that sum(ni) and length(X) give the same value before running the above command. Cheers-Jorge
De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 15 de octubre de 2015 7:18 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thank you Martin.I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue?Just in case I am also including a snapshot of my loaded variables at the bottom.Many thanks,Pablo>> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);Warning: matlabpool will be removed in a future release.To query the size of an already started parallel pool, query the 'NumWorkers'property of the pool.To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release.To query the size of an already started parallel pool, query the 'NumWorkers'property of the pool.To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release.Use parpool instead. Starting matlabpool using the 'local' profile ... Error using matlabpool (line 148)Failed to start a parallel pool. (For information in addition to the causing error,validate the profile 'local' in the Cluster Profile Manager.)Error in lme_mass_fit (line 128) matlabpool(prs);Error in lme_mass_fit_vw (line 73)[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);Caused by: Error using parallel.internal.pool.InteractiveClient/start (line 330) Failed to start pool. Error using parallel.Job/submit (line 304) You requested a minimum of 8 workers, but the cluster "local" has the NumWorkers property set to allow a maximum of 4 workers. To run a communicating job on more workers than this (up to a maximum of 512 for the Local cluster), increase the value of the NumWorkers property for the cluster. The default value of NumWorkers for a Local cluster is the number of cores on the local machine.
To: freesurfer@nmr.mgh.harvard.eduFrom: mreuter@nmr.mgh.harvard.eduDate: Wed, 14 Oct 2015 10:54:41 -0400Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users,I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families.Thanks in advance!Pablo From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi everybody,
I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem:
1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17].
2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one.
3. I could use a mixed effect model approach and specify site as a random effect.
If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches?
I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows:
Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ...
As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output:
stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex);
Starting matlabpool using the 'local' profile ... connected to 8 workers.
Starting model fitting at each location ...
Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ...
I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence.
Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate?
Thanks, Michael
_______________________________________________ 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
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
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
Hi Pablo Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term: lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4); If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly. Cheers-Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Viernes 16 de octubre de 2015 1:43 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
#yiv7275178360 #yiv7275178360 --.yiv7275178360hmmessage P{margin:0px;padding:0px;}#yiv7275178360 body.yiv7275178360hmmessage{font-size:12pt;font-family:Calibri;}#yiv7275178360 Thanks you for the replies.Jorge and FS experts,I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects.The analysis gave the following 'error'(?) below:I looked up a previous thread coming across this. At that case you recommendedWould you recommend this afain>
Aproximate percentage of fitted locations: 100%Warning: matlabpool will be removed in a future release.To query the size of an already started parallel pool, query the 'NumWorkers'property of the pool.To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release.To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead. Parallel pool using the 'local' profile is shutting down. Summary:Algorithm did not converge at 90.0637 percent of the total number of locations.Total elapsed time is 550.1023 minutes.
Also almost all the time the screen showed the following message: 144114: Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i.Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i.Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i.Location 144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448.Aproximate percentage of fitted locations: 100%
So my two questions are:1. Is this problematic?2. Are there any fixes to this issue?Thank you,Pablo Date: Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo The error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4); Please make sure that sum(ni) and length(X) give the same value before running the above command. Cheers-Jorge
De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 15 de octubre de 2015 7:18 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
#yiv7275178360 #yiv7275178360 --.yiv7275178360ExternalClass #yiv7275178360ecxyiv7790603773 body.yiv7275178360ecxyiv7790603773hmmessage {font-size:12pt;font-family:Calibri;}#yiv7275178360 Thank you Martin.I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue?Just in case I am also including a snapshot of my loaded variables at the bottom.Many thanks,Pablo>> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
Use parpool instead.
Starting matlabpool using the 'local' profile ...
Error using matlabpool (line 148)
Failed to start a parallel pool. (For information in addition to the causing error,
validate the profile 'local' in the Cluster Profile Manager.)
Error in lme_mass_fit (line 128)
matlabpool(prs);
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
Caused by:
Error using parallel.internal.pool.InteractiveClient/start (line 330)
Failed to start pool.
Error using parallel.Job/submit (line 304)
You requested a minimum of 8 workers, but the cluster "local" has the
NumWorkers property set to allow a maximum of 4 workers. To run a
communicating job on more workers than this (up to a maximum of 512 for the
Local cluster), increase the value of the NumWorkers property for the
cluster. The default value of NumWorkers for a Local cluster is the number of
cores on the local machine.
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
#yiv7275178360 #yiv7275178360 --.yiv7275178360ExternalClass #yiv7275178360ecxyiv7790603773 .yiv7275178360ecxyiv7790603773ExternalClass body.yiv7275178360ecxyiv7790603773ecxhmmessage {font-size:12pt;font-family:Calibri;}#yiv7275178360 Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.:CM.C = [your contrast matrix];F_lhstats = lme_mass_F(lhstats, CM);5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000From: jbernal0019@yahoo.esTo: freesurfer@nmr.mgh.harvard.eduSubject: Re: [Freesurfer] A mixed effect model approach in within subject datasetHi PabloI think you can useLME to analyze your data by ordering the rows of your design matrixappropriately. You can consider all subjects belonging to the samefamily as if they were a single subject in a longitudinal analysis.You can put in your design matrix all subjects belonging to family1first, then all subjects belonging to family 2 and so on. Then the'ni' required by lme_mass_fit_vw is a vector with the number ofsubjects in each family as its entries (ordered according to yourdesign matrix). So the length of the 'ni' vector is equal to thenumber of different families in your data. Now you can gofurther and additionally order the rows of your design matrix withineach family by age. This will allow you to test the effect of agewithin family. When choosing therandom effects for your statistical model remember that a randomeffect can only be the intercept term or any covariate that varieswithin family. For example you can compare a model with a singlerandom effect for the intercept term against the same model butconsidering both the intercept term and age as random effects.Hope that helpsCheers-Jorge De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset #yiv7275178360 #yiv7275178360 --.yiv7275178360ExternalClass #yiv7275178360ecxyiv7790603773 .yiv7275178360ecxyiv7790603773ExternalClass #yiv7275178360ecxyiv7790603773ecxyiv8323390595 body.yiv7275178360ecxyiv7790603773ecxyiv8323390595hmmessage {font-size:12pt;font-family:Calibri;}#yiv7275178360 Dear Freesurfer users,I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families.Thanks in advance!PabloFrom: michaelnotter@hotmail.comTo: freesurfer@nmr.mgh.harvard.eduDate: Wed, 19 Feb 2014 13:10:09 +0100Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach#yiv7275178360 #yiv7275178360 --.yiv7275178360ExternalClass #yiv7275178360ecxyiv7790603773 .yiv7275178360ecxyiv7790603773ExternalClass #yiv7275178360ecxyiv7790603773ecxyiv8323390595 .yiv7275178360ecxyiv7790603773ecxyiv8323390595ExternalClass body.yiv7275178360ecxyiv7790603773ecxyiv8323390595ecxhmmessage {font-size:12pt;font-family:Calibri;}#yiv7275178360 Hi everybody,I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem:1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17].2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one.3. I could use a mixed effect model approach and specify site as a random effect.If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches?I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows:Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site41 1 0 0 11.1 99 0 0 1 01 0 1 0 11.1 101 0 0 1 01 1 0 0 11.4 95 1 0 0 01 0 0 1 12.4 100 1 0 0 0...As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output:>> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex);Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions.Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000.Location 62484: Index exceeds matrix dimensions.Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000....I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence.Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate?Thanks,Michael _______________________________________________Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurferThe information in this e-mail is intended only for the person to whom it isaddressed. If you believe this e-mail was sent to you in error and the e-mailcontains patient information, please contact the Partners Compliance HelpLine athttp://www.partners.org/complianceline . If the e-mail was sent to you in errorbut does not contain patient information, please contact the sender and properlydispose of the e-mail. _______________________________________________Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurferThe information in this e-mail is intended only for the person to whom it isaddressed. If you believe this e-mail was sent to you in error and the e-mailcontains patient information, please contact the Partners Compliance HelpLine athttp://www.partners.org/complianceline . If the e-mail was sent to you in errorbut does not contain patient information, please contact the sender and properlydispose of the e-mail. _______________________________________________Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurferThe information in this e-mail is intended only for the person to whom it isaddressed. If you believe this e-mail was sent to you in error and the e-mailcontains patient information, please contact the Partners Compliance HelpLine athttp://www.partners.org/complianceline . If the e-mail was sent to you in errorbut does not contain patient information, please contact the sender and properlydispose of the e-mail. _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Hi Jorge and FS experts,I have run again the analysis and still get the convergence problem.I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design. Instructions on lme in the wiki for preprocessing specify the following: mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mghmri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshapeMy data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: 1) for preprocessing should I follow instructions as it would be longitudinal data?2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ?Thank youPablo ____________________________________________________________________________________________________________________________________
Date: Fri, 16 Oct 2015 16:24:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term: lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4); If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly. Cheers-Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Viernes 16 de octubre de 2015 1:43 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thanks you for the replies.Jorge and FS experts,I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects.The analysis gave the following 'error'(?) below:I looked up a previous thread coming across this. At that case you recommendedWould you recommend this afain>Aproximate percentage of fitted locations: 100%Warning: matlabpool will be removed in a future release.To query the size of an already started parallel pool, query the 'NumWorkers'property of the pool.To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release.To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead. Parallel pool using the 'local' profile is shutting down. Summary:Algorithm did not converge at 90.0637 percent of the total number of locations.Total elapsed time is 550.1023 minutes.Also almost all the time the screen showed the following message:144114: Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i.Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i.Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i.Location 144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448.Aproximate percentage of fitted locations: 100%So my two questions are:1. Is this problematic?2. Are there any fixes to this issue?Thank you,PabloDate: Thu, 15 Oct 2015 13:38:22 +0000From: jbernal0019@yahoo.esTo: freesurfer@nmr.mgh.harvard.eduCC: pablonajt@hotmail.comSubject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi PabloThe error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4);Please make sure that sum(ni) and length(X) give the same value before running the above command.Cheers-Jorge De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 15 de octubre de 2015 7:18 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Thank you Martin.I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue?Just in case I am also including a snapshot of my loaded variables at the bottom.Many thanks,Pablo>> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);Warning: matlabpool will be removed in a future release.To query the size of an already started parallel pool, query the 'NumWorkers'property of the pool.To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release.To query the size of an already started parallel pool, query the 'NumWorkers'property of the pool.To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release.Use parpool instead. Starting matlabpool using the 'local' profile ... Error using matlabpool (line 148)Failed to start a parallel pool. (For information in addition to the causing error,validate the profile 'local' in the Cluster Profile Manager.)Error in lme_mass_fit (line 128) matlabpool(prs);Error in lme_mass_fit_vw (line 73)[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);Caused by: Error using parallel.internal.pool.InteractiveClient/start (line 330) Failed to start pool. Error using parallel.Job/submit (line 304) You requested a minimum of 8 workers, but the cluster "local" has the NumWorkers property set to allow a maximum of 4 workers. To run a communicating job on more workers than this (up to a maximum of 512 for the Local cluster), increase the value of the NumWorkers property for the cluster. The default value of NumWorkers for a Local cluster is the number of cores on the local machine. To: freesurfer@nmr.mgh.harvard.eduFrom: mreuter@nmr.mgh.harvard.eduDate: Wed, 14 Oct 2015 10:54:41 -0400Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users,I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families.Thanks in advance!Pablo From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi everybody,
I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem:
1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17].
2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one.
3. I could use a mixed effect model approach and specify site as a random effect.
If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches?
I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows:
Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ...
As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output:
stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex);
Starting matlabpool using the 'local' profile ... connected to 8 workers.
Starting model fitting at each location ...
Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ...
I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence.
Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate?
Thanks, Michael
_______________________________________________ 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
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
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
Hi Pablo,
since you did not run this through the longitudinal stream, you won't have a base ( or subject-template), also you won't have <tpid>.long.<base> directories, so the first command should fail. For that you simply replace the --qdec-long with --qdec and use the same table. Should work. The second command is fine (it just smoothes the thickness stack).
Best, Martin
On 10/19/2015 06:51 AM, pablo najt wrote:
Hi Jorge and FS experts, I have run again the analysis and still get the convergence problem. I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design.
Instructions on lme in the wiki for preprocessing specify the following:
mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mgh mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape My data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: *1) for preprocessing should I follow instructions as it would be longitudinal data?* *2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ?* Thank you Pablo ____________________________________________________________________________________________________________________________________
Date: Fri, 16 Oct 2015 16:24:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo
Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term:
lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4);
If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly.
Cheers -Jorge
------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <freesurfer@nmr.mgh.harvard.edu> *Enviado:* Viernes 16 de octubre de 2015 1:43 *Asunto:* Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Thanks you for the replies. Jorge and FS experts, I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects. The analysis gave the following 'error'(?) below: I looked up a previous thread coming across this. At that case you recommended Would you recommend this afain> /Aproximate percentage of fitted locations: 100%/ /Warning: matlabpool will be removed in a future release./ /To query the size of an already started parallel pool, query the 'NumWorkers'/ /property of the pool./ /To check if a pool is already started use 'isempty(gcp('nocreate'))'. / /Warning: matlabpool will be removed in a future release./ /To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead. / /Parallel pool using the 'local' profile is shutting down./ // /Summary:/ /Algorithm did not converge at 90.0637 percent of the total number of locations./ /Total elapsed time is 550.1023 minutes./ Also almost all the time the screen showed the following message: /144114: Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i./ /Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i./ /Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i./ /Location 144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448./ /Aproximate percentage of fitted locations: 100%/ / / So my two questions are: 1. Is this problematic? 2. Are there any fixes to this issue? Thank you, Pablo ------------------------------------------------------------------------ Date: Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo The error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4); Please make sure that sum(ni) and length(X) give the same value before running the above command. Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> *Para:* Freesurfer support list <freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 15 de octubre de 2015 7:18 *Asunto:* Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Thank you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just in case I am also including a snapshot of my loaded variables at the bottom. Many thanks, Pablo >> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. Use parpool instead. Starting matlabpool using the 'local' profile ... Error using matlabpool (line 148) Failed to start a parallel pool. (For information in addition to the causing error, validate the profile 'local' in the Cluster Profile Manager.) Error in lme_mass_fit (line 128) matlabpool(prs); Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); Caused by: Error using parallel.internal.pool.InteractiveClient/start (line 330) Failed to start pool. Error using parallel.Job/submit (line 304) You requested a minimum of 8 workers, but the cluster "local" has the NumWorkers property set to allow a maximum of 4 workers. To run a communicating job on more workers than this (up to a maximum of 512 for the Local cluster), increase the value of the NumWorkers property for the cluster. The default value of NumWorkers for a Local cluster is the number of cores on the local machine. ------------------------------------------------------------------------ To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo, you should run something like this to get the ni: [M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels hope that helps, Martin On 10/14/2015 10:43 AM, pablo najt wrote: Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh'); %---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------% 3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------% lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex); Error using lme_mass_fit (line 108) The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design matrix X Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig'); ------------------------------------------------------------------------ Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset Hi Pablo I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data. Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family. When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects. Hope that helps Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 10 de septiembre de 2015 8:07 *Asunto:* [Freesurfer] A mixed effect model approach in within subject dataset Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo From: michaelnotter@hotmail.com <mailto:michaelnotter@hotmail.com> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) <http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels%29> but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto: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 <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu <http://reuter.mit.edu/> _______________________________________________ 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 <mailto: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 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 <mailto: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 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
Thank you
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Mon, 19 Oct 2015 09:00:57 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
since you did not run this through the longitudinal stream, you won't have a base ( or subject-template), also you won't have <tpid>.long.<base> directories, so the first command should fail. For that you simply replace the --qdec-long with --qdec and use the same table. Should work. The second command is fine (it just smoothes the thickness stack).
Best, Martin
On 10/19/2015 06:51 AM, pablo najt wrote:
Hi Jorge and FS experts, I have run again the analysis and still get the convergence problem. I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design.
Instructions on lme in the wiki for preprocessing specify the following:
mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mgh
mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape
My data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: 1) for preprocessing should I follow instructions as it would be longitudinal data? 2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ? Thank you Pablo
____________________________________________________________________________________________________________________________________
Date: Fri, 16 Oct 2015 16:24:36 +0000
From: jbernal0019@yahoo.es
To: freesurfer@nmr.mgh.harvard.edu
CC: pablonajt@hotmail.com
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo
Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term:
lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4);
If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly.
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com
Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu
Enviado: Viernes 16 de octubre de 2015 1:43
Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thanks you for the replies. Jorge and FS experts, I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects. The analysis gave the following 'error'(?) below: I looked up a previous thread coming across this. At that case you recommended Would you recommend this afain>
Aproximate percentage of fitted locations: 100%
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead.
Parallel pool using the 'local' profile is shutting down.
Summary:
Algorithm did not converge at 90.0637 percent of the total number of locations.
Total elapsed time is 550.1023 minutes.
Also almost all the time the screen showed the following message:
144114: Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i.
Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i.
Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i.
Location 144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448.
Aproximate percentage of fitted locations: 100%
So my two questions are: 1. Is this problematic? 2. Are there any fixes to this issue? Thank you, Pablo
Date: Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo
The error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this:
lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4);
Please make sure that sum(ni) and length(X) give the same value before running the above command.
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu
Enviado: Jueves 15 de octubre de 2015 7:18 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thank you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just in case I am also including a snapshot of my loaded variables at the bottom. Many thanks, Pablo >> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
Use parpool instead.
Starting matlabpool using the 'local' profile ...
Error using matlabpool (line 148)
Failed to start a parallel pool. (For information in addition to the causing error,
validate the profile 'local' in the Cluster Profile Manager.)
Error in lme_mass_fit (line 128)
matlabpool(prs);
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
Caused by:
Error using parallel.internal.pool.InteractiveClient/start (line 330)
Failed to start pool.
Error using parallel.Job/submit (line 304)
You requested a minimum of 8 workers, but the cluster "local" has the
NumWorkers property set to allow a maximum of 4 workers. To run a
communicating job on more workers than this (up to a maximum of 512 for the
Local cluster), increase the value of the NumWorkers property for the
cluster. The default value of NumWorkers for a Local cluster is the number of
cores on the local machine.
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo
From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers.
Starting model fitting at each location ...
Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael
_______________________________________________ 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 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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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
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
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
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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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.
Dear FS experts,I am following up on an issue I am having when trying to run linear mixed effects model. I keep getting 'Algorithm did not converge'.So far I have changed the preprocessing steps following literally from the wiki as in message below. Also I tried to use only one random effect for the intercept. Also I have checked that sum of vectors of (ni) is equal to the number of rows in (X). As none of the above worked I started to have concerns about the dimensions of my lh.thickness_sm10.mgh. When I load in matlab in working space says 'too many variables'. Is this ok?
Also, when I open 'lhY' variable it shows a column of Zeros every nine columns as below. Is this alright?
Perhaps this is fine but I would like to check on anything that is creating this problem. For matrix 'M' I created a matlab script that loads a matrix with intercept (all ones), HC (ones to HC otherwise zero), U_PT (relatives of patients ones otherwise zero), PT (patients one otherwise zero), gender, age (see below). The three columns for groups do not match the qdec.table.dat which have only one column for groups (HC, U_PT, PT). However the order of the rows is identical in M and qdec.table.dat
Please let me know if this should be set differently.Thank youPablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Mon, 19 Oct 2015 09:00:57 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
since you did not run this through the longitudinal stream, you won't have a base ( or subject-template), also you won't have <tpid>.long.<base> directories, so the first command should fail. For that you simply replace the --qdec-long with --qdec and use the same table. Should work. The second command is fine (it just smoothes the thickness stack).
Best, Martin
On 10/19/2015 06:51 AM, pablo najt wrote:
Hi Jorge and FS experts, I have run again the analysis and still get the convergence problem. I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design.
Instructions on lme in the wiki for preprocessing specify the following:
mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mgh
mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape
My data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: 1) for preprocessing should I follow instructions as it would be longitudinal data? 2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ? Thank you Pablo
____________________________________________________________________________________________________________________________________
Date: Fri, 16 Oct 2015 16:24:36 +0000
From: jbernal0019@yahoo.es
To: freesurfer@nmr.mgh.harvard.edu
CC: pablonajt@hotmail.com
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo
Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term:
lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4);
If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly.
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com
Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu
Enviado: Viernes 16 de octubre de 2015 1:43
Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thanks you for the replies. Jorge and FS experts, I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects. The analysis gave the following 'error'(?) below: I looked up a previous thread coming across this. At that case you recommended Would you recommend this afain>
Aproximate percentage of fitted locations: 100%
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead.
Parallel pool using the 'local' profile is shutting down.
Summary:
Algorithm did not converge at 90.0637 percent of the total number of locations.
Total elapsed time is 550.1023 minutes.
Also almost all the time the screen showed the following message:
144114: Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i.
Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i.
Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i.
Location 144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448.
Aproximate percentage of fitted locations: 100%
So my two questions are: 1. Is this problematic? 2. Are there any fixes to this issue? Thank you, Pablo
Date: Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo
The error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this:
lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4);
Please make sure that sum(ni) and length(X) give the same value before running the above command.
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu
Enviado: Jueves 15 de octubre de 2015 7:18 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thank you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just in case I am also including a snapshot of my loaded variables at the bottom. Many thanks, Pablo >> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
Use parpool instead.
Starting matlabpool using the 'local' profile ...
Error using matlabpool (line 148)
Failed to start a parallel pool. (For information in addition to the causing error,
validate the profile 'local' in the Cluster Profile Manager.)
Error in lme_mass_fit (line 128)
matlabpool(prs);
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
Caused by:
Error using parallel.internal.pool.InteractiveClient/start (line 330)
Failed to start pool.
Error using parallel.Job/submit (line 304)
You requested a minimum of 8 workers, but the cluster "local" has the
NumWorkers property set to allow a maximum of 4 workers. To run a
communicating job on more workers than this (up to a maximum of 512 for the
Local cluster), increase the value of the NumWorkers property for the
cluster. The default value of NumWorkers for a Local cluster is the number of
cores on the local machine.
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo
From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers.
Starting model fitting at each location ...
Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael
_______________________________________________ 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 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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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
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
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
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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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 Pablo,
the zero columns could be outside the cortex and then this would be fine. Do min(sum(lhY(:,lhcortex))) if you get 0 you have at least one location where the column is all zero (as we should not have any negative values, you can test that with min(lhY(:)) Now even if you still have a column of 0 , then that vertex is problematic, but who cares about a single vertex. If you do a region approach, it will be part of a region anyway. so you should check how often you get 0 columns (inside cortex).
About your M: I would subtract 1 from gender (to have 0 and 1) Also you only need 2 columns for the groups: keep only column 3 and 4. A HC is someone who has 0 in both.
Best, Martin
On 10/20/2015 04:37 AM, pablo najt wrote:
Dear FS experts, I am following up on an issue I am having when trying to run linear mixed effects model. I keep getting 'Algorithm did not converge'. So far I have changed the preprocessing steps following literally from the wiki as in message below. Also I tried to use only one random effect for the intercept. Also I have checked that sum of vectors of (ni) is equal to the number of rows in (X).
As none of the above worked I started to have concerns about the dimensions of my lh.thickness_sm10.mgh. When I load in matlab in working space says 'too many variables'. Is this ok?
Also, when I open 'lhY' variable it shows a column of Zeros every nine columns as below. Is this alright?
Perhaps this is fine but I would like to check on anything that is creating this problem. For matrix 'M' I created a matlab script that loads a matrix with intercept (all ones), HC (ones to HC otherwise zero), U_PT (relatives of patients ones otherwise zero), PT (patients one otherwise zero), gender, age (see below). The three columns for groups do not match the qdec.table.dat which have only one column for groups (HC, U_PT, PT). However the order of the rows is identical in M and qdec.table.dat
Please let me know if this should be set differently. Thank you Pablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Mon, 19 Oct 2015 09:00:57 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
since you did not run this through the longitudinal stream, you won't have a base ( or subject-template), also you won't have <tpid>.long.<base> directories, so the first command should fail. For that you simply replace the --qdec-long with --qdec and use the same table. Should work. The second command is fine (it just smoothes the thickness stack).
Best, Martin
On 10/19/2015 06:51 AM, pablo najt wrote:
Hi Jorge and FS experts, I have run again the analysis and still get the convergence problem. I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design. Instructions on lme in the wiki for preprocessing specify the following: mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mgh mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape My data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: *1) for preprocessing should I follow instructions as it would be longitudinal data?* *2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ?* Thank you Pablo ____________________________________________________________________________________________________________________________________ Date: Fri, 16 Oct 2015 16:24:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> CC: pablonajt@hotmail.com <mailto:pablonajt@hotmail.com> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term: lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4); If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly. Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Viernes 16 de octubre de 2015 1:43 *Asunto:* Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Thanks you for the replies. Jorge and FS experts, I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects. The analysis gave the following 'error'(?) below: I looked up a previous thread coming across this. At that case you recommended Would you recommend this afain> /Aproximate percentage of fitted locations: 100%/ /Warning: matlabpool will be removed in a future release./ /To query the size of an already started parallel pool, query the 'NumWorkers'/ /property of the pool./ /To check if a pool is already started use 'isempty(gcp('nocreate'))'. / /Warning: matlabpool will be removed in a future release./ /To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead. / /Parallel pool using the 'local' profile is shutting down./ // /Summary:/ /Algorithm did not converge at 90.0637 percent of the total number of locations./ /Total elapsed time is 550.1023 minutes./ Also almost all the time the screen showed the following message: /144114: Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i./ /Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i./ /Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i./ /Location 144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448./ /Aproximate percentage of fitted locations: 100%/ / / So my two questions are: 1. Is this problematic? 2. Are there any fixes to this issue? Thank you, Pablo ------------------------------------------------------------------------ Date: Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> CC: pablonajt@hotmail.com <mailto:pablonajt@hotmail.com> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo The error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4); Please make sure that sum(ni) and length(X) give the same value before running the above command. Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* Freesurfer support list <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 15 de octubre de 2015 7:18 *Asunto:* Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Thank you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just in case I am also including a snapshot of my loaded variables at the bottom. Many thanks, Pablo >> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. Use parpool instead. Starting matlabpool using the 'local' profile ... Error using matlabpool (line 148) Failed to start a parallel pool. (For information in addition to the causing error, validate the profile 'local' in the Cluster Profile Manager.) Error in lme_mass_fit (line 128) matlabpool(prs); Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); Caused by: Error using parallel.internal.pool.InteractiveClient/start (line 330) Failed to start pool. Error using parallel.Job/submit (line 304) You requested a minimum of 8 workers, but the cluster "local" has the NumWorkers property set to allow a maximum of 4 workers. To run a communicating job on more workers than this (up to a maximum of 512 for the Local cluster), increase the value of the NumWorkers property for the cluster. The default value of NumWorkers for a Local cluster is the number of cores on the local machine. ------------------------------------------------------------------------ To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> From: mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo, you should run something like this to get the ni: [M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels hope that helps, Martin On 10/14/2015 10:43 AM, pablo najt wrote: Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh'); %---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------% 3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------% lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex); Error using lme_mass_fit (line 108) The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design matrix X Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig'); ------------------------------------------------------------------------ Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset Hi Pablo I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data. Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family. When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects. Hope that helps Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 10 de septiembre de 2015 8:07 *Asunto:* [Freesurfer] A mixed effect model approach in within subject dataset Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo From: michaelnotter@hotmail.com <mailto:michaelnotter@hotmail.com> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto: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 <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu <http://reuter.mit.edu/> _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto: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 <mailto: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 <mailto: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 <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu
_______________________________________________ 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
Thank you Martin for your response.I am not sure if I understood. 1. If I get '0' with this formula, min(sum(lhY(:,lhcortex))) means that there is at least one location within the brain with '0' value at a vertex.I do not get '0' it gives me 170.8210 2.But this formula min(lhY(:)) gives me '0'. What does this mean then if I get NON '0' in formula 1. and '0' in formula 2? Will correct the matrix. Could this be the cause of the 'Algorithm did not converge'?ThanksPablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Tue, 20 Oct 2015 09:06:53 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} {Disarmed}
Hi Pablo,
the zero columns could be outside the cortex and then this would be fine. Do
min(sum(lhY(:,lhcortex)))
if you get 0 you have at least one location where the column is all zero (as we should not have any negative values, you can test that with min(lhY(:))
Now even if you still have a column of 0 , then that vertex is problematic, but who cares about a single vertex. If you do a region approach, it will be part of a region anyway. so you should check how often you get 0 columns (inside cortex).
About your M: I would subtract 1 from gender (to have 0 and 1)
Also you only need 2 columns for the groups: keep only column 3 and 4. A HC is someone who has 0 in both.
Best, Martin
On 10/20/2015 04:37 AM, pablo najt wrote:
Dear FS experts, I am following up on an issue I am having when trying to run linear mixed effects model. I keep getting 'Algorithm did not converge'. So far I have changed the preprocessing steps following literally from the wiki as in message below. Also I tried to use only one random effect for the intercept. Also I have checked that sum of vectors of (ni) is equal to the number of rows in (X).
As none of the above worked I started to have concerns about the dimensions of my lh.thickness_sm10.mgh. When I load in matlab in working space says 'too many variables'. Is this ok?
Also, when I open 'lhY' variable it shows a column of Zeros every nine columns as below. Is this alright?
Perhaps this is fine but I would like to check on anything that is creating this problem. For matrix 'M' I created a matlab script that loads a matrix with intercept (all ones), HC (ones to HC otherwise zero), U_PT (relatives of patients ones otherwise zero), PT (patients one otherwise zero), gender, age (see below). The three columns for groups do not match the qdec.table.dat which have only one column for groups (HC, U_PT, PT). However the order of the rows is identical in M and qdec.table.dat
Please let me know if this should be set differently. Thank you Pablo
To: freesurfer@nmr.mgh.harvard.edu
From: mreuter@nmr.mgh.harvard.edu
Date: Mon, 19 Oct 2015 09:00:57 -0400
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
since you did not run this through the longitudinal stream, you won't have a base ( or subject-template), also you won't have <tpid>.long.<base> directories, so the first command should fail. For that you simply replace the --qdec-long with --qdec and use the same table. Should work. The second command is fine (it just smoothes the thickness stack).
Best, Martin
On 10/19/2015 06:51 AM, pablo najt wrote:
Hi Jorge and FS experts, I have run again the analysis and still get the convergence problem. I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design.
Instructions on lme in the wiki for preprocessing specify the following:
mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mgh
mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape
My data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: 1) for preprocessing should I follow instructions as it would be longitudinal data? 2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ? Thank you Pablo
____________________________________________________________________________________________________________________________________
Date: Fri, 16 Oct 2015 16:24:36 +0000
From: jbernal0019@yahoo.es
To: freesurfer@nmr.mgh.harvard.edu
CC: pablonajt@hotmail.com
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo
Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term:
lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4);
If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly.
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com
Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu
Enviado: Viernes 16 de octubre de 2015 1:43
Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thanks you for the replies. Jorge and FS experts, I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects. The analysis gave the following 'error'(?) below: I looked up a previous thread coming across this. At that case you recommended Would you recommend this afain>
Aproximate
percentage of fitted locations: 100%
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead.
Parallel pool using the 'local' profile is shutting down.
Summary:
Algorithm did not converge at 90.0637 percent of the total number of locations.
Total elapsed time is 550.1023 minutes.
Also
almost all the time the screen showed the following message:
144114:
Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i.
Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i.
Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i.
Location
144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448.
Aproximate percentage of fitted locations: 100%
So my two questions are: 1. Is this problematic? 2. Are there any fixes to this issue? Thank
you, Pablo
Date:
Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo
The
error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this:
lhstats
= lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4);
Please
make sure that sum(ni) and length(X) give the same value before running the above command.
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu
Enviado: Jueves 15 de octubre de 2015 7:18 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thank
you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just
in case I am also including a snapshot of my loaded variables at the bottom. Many
thanks, Pablo >>
lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Warning:
matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property
of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning:
matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property
of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning:
matlabpool will be removed in a future release.
Use parpool instead.
Starting
matlabpool using the 'local' profile ...
Error using matlabpool (line 148)
Failed
to start a parallel pool. (For information in addition to the causing error,
validate
the profile 'local' in the Cluster Profile Manager.)
Error in lme_mass_fit (line 128)
matlabpool(prs);
Error in lme_mass_fit_vw (line 73)
[stats1,st1]
= lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
Caused
by:
Error using parallel.internal.pool.InteractiveClient/start (line 330)
Failed to start pool.
Error using parallel.Job/submit
(line 304)
You requested a minimum of 8 workers, but the cluster "local" has the
NumWorkers property set to allow a maximum of 4 workers. To run a
communicating job on more workers than this (up to a maximum of 512 for the
Local cluster), increase the value of the NumWorkers property for the
cluster. The default value of NumWorkers for a Local cluster is the number of
cores on the local machine.
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On
10/14/2015 10:43 AM, pablo najt wrote:
Dear
FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date:
Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to
control for the non-independence of data contributed by patients and relatives from the same families. Thanks
in advance! Pablo
From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi
everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics
of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael
_______________________________________________ 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 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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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
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 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
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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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 Pablo,
1. Good, getting 170... means you have no vertex with all zeros inside the cortex label. (If there was a column with all zeros inside cortex, the sum would be zero). 2. Good, means that you have no negative values anywhere, which is expected.
3. Yes, that causes the problem. Otherwise M is singular (before you could linearly combine row 1,3 and 4 to get 2: r1 - r3 -r4 = r2)
Cheers, Martin
On 10/20/2015 10:10 AM, pablo najt wrote:
Thank you Martin for your response. I am not sure if I understood.
- If I get '0' with this formula, min(sum(lhY(:,lhcortex))) means
that there is at least one location within the brain with '0' value at a vertex. I do not get '0' it gives me 170.8210
2.But this formula min(lhY(:)) gives me '0'. What does this mean then if I get NON '0' in formula 1. and '0' in formula 2?
Will correct the matrix. Could this be the cause of the 'Algorithm did not converge'? Thanks Pablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Tue, 20 Oct 2015 09:06:53 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} {Disarmed}
Hi Pablo,
the zero columns could be outside the cortex and then this would be fine. Do min(sum(lhY(:,lhcortex))) if you get 0 you have at least one location where the column is all zero (as we should not have any negative values, you can test that with min(lhY(:)) Now even if you still have a column of 0 , then that vertex is problematic, but who cares about a single vertex. If you do a region approach, it will be part of a region anyway. so you should check how often you get 0 columns (inside cortex).
About your M: I would subtract 1 from gender (to have 0 and 1) Also you only need 2 columns for the groups: keep only column 3 and 4. A HC is someone who has 0 in both.
Best, Martin
On 10/20/2015 04:37 AM, pablo najt wrote:
Dear FS experts, I am following up on an issue I am having when trying to run linear mixed effects model. I keep getting 'Algorithm did not converge'. So far I have changed the preprocessing steps following literally from the wiki as in message below. Also I tried to use only one random effect for the intercept. Also I have checked that sum of vectors of (ni) is equal to the number of rows in (X). As none of the above worked I started to have concerns about the dimensions of my lh.thickness_sm10.mgh. When I load in matlab in working space says 'too many variables'. Is this ok? Also, when I open 'lhY' variable it shows a column of Zeros every nine columns as below. Is this alright? Perhaps this is fine but I would like to check on anything that is creating this problem. For matrix 'M' I created a matlab script that loads a matrix with intercept (all ones), HC (ones to HC otherwise zero), U_PT (relatives of patients ones otherwise zero), PT (patients one otherwise zero), gender, age (see below). The three columns for groups do not match the qdec.table.dat which have only one column for groups (HC, U_PT, PT). However the order of the rows is identical in M and qdec.table.dat Please let me know if this should be set differently. Thank you Pablo ------------------------------------------------------------------------ To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> From: mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> Date: Mon, 19 Oct 2015 09:00:57 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo, since you did not run this through the longitudinal stream, you won't have a base ( or subject-template), also you won't have <tpid>.long.<base> directories, so the first command should fail. For that you simply replace the --qdec-long with --qdec and use the same table. Should work. The second command is fine (it just smoothes the thickness stack). Best, Martin On 10/19/2015 06:51 AM, pablo najt wrote: Hi Jorge and FS experts, I have run again the analysis and still get the convergence problem. I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design. Instructions on lme in the wiki for preprocessing specify the following: mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mgh mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape My data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: *1) for preprocessing should I follow instructions as it would be longitudinal data?* *2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ?* Thank you Pablo ____________________________________________________________________________________________________________________________________ Date: Fri, 16 Oct 2015 16:24:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> CC: pablonajt@hotmail.com <mailto:pablonajt@hotmail.com> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term: lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4); If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly. Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Viernes 16 de octubre de 2015 1:43 *Asunto:* Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Thanks you for the replies. Jorge and FS experts, I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects. The analysis gave the following 'error'(?) below: I looked up a previous thread coming across this. At that case you recommended Would you recommend this afain> /Aproximate percentage of fitted locations: 100%/ /Warning: matlabpool will be removed in a future release./ /To query the size of an already started parallel pool, query the 'NumWorkers'/ /property of the pool./ /To check if a pool is already started use 'isempty(gcp('nocreate'))'. / /Warning: matlabpool will be removed in a future release./ /To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead. / /Parallel pool using the 'local' profile is shutting down./ // /Summary:/ /Algorithm did not converge at 90.0637 percent of the total number of locations./ /Total elapsed time is 550.1023 minutes./ Also almost all the time the screen showed the following message: /144114: Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i./ /Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i./ /Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i./ /Location 144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448./ /Aproximate percentage of fitted locations: 100%/ / / So my two questions are: 1. Is this problematic? 2. Are there any fixes to this issue? Thank you, Pablo ------------------------------------------------------------------------ Date: Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> CC: pablonajt@hotmail.com <mailto:pablonajt@hotmail.com> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo The error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4); Please make sure that sum(ni) and length(X) give the same value before running the above command. Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* Freesurfer support list <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 15 de octubre de 2015 7:18 *Asunto:* Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Thank you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just in case I am also including a snapshot of my loaded variables at the bottom. Many thanks, Pablo >> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. Use parpool instead. Starting matlabpool using the 'local' profile ... Error using matlabpool (line 148) Failed to start a parallel pool. (For information in addition to the causing error, validate the profile 'local' in the Cluster Profile Manager.) Error in lme_mass_fit (line 128) matlabpool(prs); Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); Caused by: Error using parallel.internal.pool.InteractiveClient/start (line 330) Failed to start pool. Error using parallel.Job/submit (line 304) You requested a minimum of 8 workers, but the cluster "local" has the NumWorkers property set to allow a maximum of 4 workers. To run a communicating job on more workers than this (up to a maximum of 512 for the Local cluster), increase the value of the NumWorkers property for the cluster. The default value of NumWorkers for a Local cluster is the number of cores on the local machine. ------------------------------------------------------------------------ To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> From: mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo, you should run something like this to get the ni: [M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels hope that helps, Martin On 10/14/2015 10:43 AM, pablo najt wrote: Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh'); %---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------% 3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------% lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex); Error using lme_mass_fit (line 108) The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design matrix X Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig'); ------------------------------------------------------------------------ Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset Hi Pablo I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data. Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family. When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects. Hope that helps Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 10 de septiembre de 2015 8:07 *Asunto:* [Freesurfer] A mixed effect model approach in within subject dataset Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo From: michaelnotter@hotmail.com <mailto:michaelnotter@hotmail.com> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto: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 <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu <http://reuter.mit.edu/> _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto: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 <mailto: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 <mailto: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 <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu
_______________________________________________ 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
Martin,Doing the changes in the matrix allowed me to run the command. The command with just one random effect for the intercept term give me the following output:Aproximate percentage of fitted locations: 100%Summary:Algorithm did not converge at 23.407 percent of the total number of locations.Total elapsed time is 116.2184 minutes. Is this ok? Is there anything I should check?Should I also try two random effects as well?Thank youPablo To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Tue, 20 Oct 2015 09:06:53 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} {Disarmed}
Hi Pablo,
the zero columns could be outside the cortex and then this would be fine. Do
min(sum(lhY(:,lhcortex)))
if you get 0 you have at least one location where the column is all zero (as we should not have any negative values, you can test that with min(lhY(:))
Now even if you still have a column of 0 , then that vertex is problematic, but who cares about a single vertex. If you do a region approach, it will be part of a region anyway. so you should check how often you get 0 columns (inside cortex).
About your M: I would subtract 1 from gender (to have 0 and 1)
Also you only need 2 columns for the groups: keep only column 3 and 4. A HC is someone who has 0 in both.
Best, Martin
On 10/20/2015 04:37 AM, pablo najt wrote:
Dear FS experts, I am following up on an issue I am having when trying to run linear mixed effects model. I keep getting 'Algorithm did not converge'. So far I have changed the preprocessing steps following literally from the wiki as in message below. Also I tried to use only one random effect for the intercept. Also I have checked that sum of vectors of (ni) is equal to the number of rows in (X).
As none of the above worked I started to have concerns about the dimensions of my lh.thickness_sm10.mgh. When I load in matlab in working space says 'too many variables'. Is this ok?
Also, when I open 'lhY' variable it shows a column of Zeros every nine columns as below. Is this alright?
Perhaps this is fine but I would like to check on anything that is creating this problem. For matrix 'M' I created a matlab script that loads a matrix with intercept (all ones), HC (ones to HC otherwise zero), U_PT (relatives of patients ones otherwise zero), PT (patients one otherwise zero), gender, age (see below). The three columns for groups do not match the qdec.table.dat which have only one column for groups (HC, U_PT, PT). However the order of the rows is identical in M and qdec.table.dat
Please let me know if this should be set differently. Thank you Pablo
To: freesurfer@nmr.mgh.harvard.edu
From: mreuter@nmr.mgh.harvard.edu
Date: Mon, 19 Oct 2015 09:00:57 -0400
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
since you did not run this through the longitudinal stream, you won't have a base ( or subject-template), also you won't have <tpid>.long.<base> directories, so the first command should fail. For that you simply replace the --qdec-long with --qdec and use the same table. Should work. The second command is fine (it just smoothes the thickness stack).
Best, Martin
On 10/19/2015 06:51 AM, pablo najt wrote:
Hi Jorge and FS experts, I have run again the analysis and still get the convergence problem. I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design.
Instructions on lme in the wiki for preprocessing specify the following:
mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mgh
mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape
My data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: 1) for preprocessing should I follow instructions as it would be longitudinal data? 2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ? Thank you Pablo
____________________________________________________________________________________________________________________________________
Date: Fri, 16 Oct 2015 16:24:36 +0000
From: jbernal0019@yahoo.es
To: freesurfer@nmr.mgh.harvard.edu
CC: pablonajt@hotmail.com
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo
Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term:
lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4);
If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly.
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com
Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu
Enviado: Viernes 16 de octubre de 2015 1:43
Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thanks you for the replies. Jorge and FS experts, I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects. The analysis gave the following 'error'(?) below: I looked up a previous thread coming across this. At that case you recommended Would you recommend this afain>
Aproximate
percentage of fitted locations: 100%
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead.
Parallel pool using the 'local' profile is shutting down.
Summary:
Algorithm did not converge at 90.0637 percent of the total number of locations.
Total elapsed time is 550.1023 minutes.
Also
almost all the time the screen showed the following message:
144114:
Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i.
Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i.
Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i.
Location
144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448.
Aproximate percentage of fitted locations: 100%
So my two questions are: 1. Is this problematic? 2. Are there any fixes to this issue? Thank
you, Pablo
Date:
Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo
The
error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this:
lhstats
= lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4);
Please
make sure that sum(ni) and length(X) give the same value before running the above command.
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu
Enviado: Jueves 15 de octubre de 2015 7:18 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thank
you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just
in case I am also including a snapshot of my loaded variables at the bottom. Many
thanks, Pablo >>
lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Warning:
matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property
of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning:
matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property
of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning:
matlabpool will be removed in a future release.
Use parpool instead.
Starting
matlabpool using the 'local' profile ...
Error using matlabpool (line 148)
Failed
to start a parallel pool. (For information in addition to the causing error,
validate
the profile 'local' in the Cluster Profile Manager.)
Error in lme_mass_fit (line 128)
matlabpool(prs);
Error in lme_mass_fit_vw (line 73)
[stats1,st1]
= lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
Caused
by:
Error using parallel.internal.pool.InteractiveClient/start (line 330)
Failed to start pool.
Error using parallel.Job/submit
(line 304)
You requested a minimum of 8 workers, but the cluster "local" has the
NumWorkers property set to allow a maximum of 4 workers. To run a
communicating job on more workers than this (up to a maximum of 512 for the
Local cluster), increase the value of the NumWorkers property for the
cluster. The default value of NumWorkers for a Local cluster is the number of
cores on the local machine.
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On
10/14/2015 10:43 AM, pablo najt wrote:
Dear
FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date:
Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to
control for the non-independence of data contributed by patients and relatives from the same families. Thanks
in advance! Pablo
From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi
everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics
of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael
_______________________________________________ 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 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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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
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 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
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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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 Pablo,
23% still seems a lot, but may be OK (maybe Jorge can comment on that)?
If you want to do a model comparison (2 random vs 1) take a look at the wiki. It describes how to do that.
Best, Martin
On 10/21/2015 01:19 AM, pablo najt wrote:
Martin, Doing the changes in the matrix allowed me to run the command. The command with just one random effect for the intercept term give me the following output:
/Aproximate percentage of fitted locations: 100%/ / Summary: / / Algorithm did not converge at 23.407 percent of the total number of locations. / / Total elapsed time is 116.2184 minutes. //
/Is this ok? Is there anything I should check? Should I also try two random effects as well? Thank you Pablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Tue, 20 Oct 2015 09:06:53 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} {Disarmed}
Hi Pablo,
the zero columns could be outside the cortex and then this would be fine. Do min(sum(lhY(:,lhcortex))) if you get 0 you have at least one location where the column is all zero (as we should not have any negative values, you can test that with min(lhY(:)) Now even if you still have a column of 0 , then that vertex is problematic, but who cares about a single vertex. If you do a region approach, it will be part of a region anyway. so you should check how often you get 0 columns (inside cortex).
About your M: I would subtract 1 from gender (to have 0 and 1) Also you only need 2 columns for the groups: keep only column 3 and 4. A HC is someone who has 0 in both.
Best, Martin
On 10/20/2015 04:37 AM, pablo najt wrote:
Dear FS experts, I am following up on an issue I am having when trying to run linear mixed effects model. I keep getting 'Algorithm did not converge'. So far I have changed the preprocessing steps following literally from the wiki as in message below. Also I tried to use only one random effect for the intercept. Also I have checked that sum of vectors of (ni) is equal to the number of rows in (X). As none of the above worked I started to have concerns about the dimensions of my lh.thickness_sm10.mgh. When I load in matlab in working space says 'too many variables'. Is this ok? Also, when I open 'lhY' variable it shows a column of Zeros every nine columns as below. Is this alright? Perhaps this is fine but I would like to check on anything that is creating this problem. For matrix 'M' I created a matlab script that loads a matrix with intercept (all ones), HC (ones to HC otherwise zero), U_PT (relatives of patients ones otherwise zero), PT (patients one otherwise zero), gender, age (see below). The three columns for groups do not match the qdec.table.dat which have only one column for groups (HC, U_PT, PT). However the order of the rows is identical in M and qdec.table.dat Please let me know if this should be set differently. Thank you Pablo ------------------------------------------------------------------------ To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> From: mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> Date: Mon, 19 Oct 2015 09:00:57 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo, since you did not run this through the longitudinal stream, you won't have a base ( or subject-template), also you won't have <tpid>.long.<base> directories, so the first command should fail. For that you simply replace the --qdec-long with --qdec and use the same table. Should work. The second command is fine (it just smoothes the thickness stack). Best, Martin On 10/19/2015 06:51 AM, pablo najt wrote: Hi Jorge and FS experts, I have run again the analysis and still get the convergence problem. I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design. Instructions on lme in the wiki for preprocessing specify the following: mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mgh mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape My data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: *1) for preprocessing should I follow instructions as it would be longitudinal data?* *2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ?* Thank you Pablo ____________________________________________________________________________________________________________________________________ Date: Fri, 16 Oct 2015 16:24:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> CC: pablonajt@hotmail.com <mailto:pablonajt@hotmail.com> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term: lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4); If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly. Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Viernes 16 de octubre de 2015 1:43 *Asunto:* Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Thanks you for the replies. Jorge and FS experts, I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects. The analysis gave the following 'error'(?) below: I looked up a previous thread coming across this. At that case you recommended Would you recommend this afain> /Aproximate percentage of fitted locations: 100%/ /Warning: matlabpool will be removed in a future release./ /To query the size of an already started parallel pool, query the 'NumWorkers'/ /property of the pool./ /To check if a pool is already started use 'isempty(gcp('nocreate'))'. / /Warning: matlabpool will be removed in a future release./ /To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead. / /Parallel pool using the 'local' profile is shutting down./ // /Summary:/ /Algorithm did not converge at 90.0637 percent of the total number of locations./ /Total elapsed time is 550.1023 minutes./ Also almost all the time the screen showed the following message: /144114: Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i./ /Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i./ /Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i./ /Location 144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448./ /Aproximate percentage of fitted locations: 100%/ / / So my two questions are: 1. Is this problematic? 2. Are there any fixes to this issue? Thank you, Pablo ------------------------------------------------------------------------ Date: Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> CC: pablonajt@hotmail.com <mailto:pablonajt@hotmail.com> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo The error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4); Please make sure that sum(ni) and length(X) give the same value before running the above command. Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* Freesurfer support list <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 15 de octubre de 2015 7:18 *Asunto:* Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Thank you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just in case I am also including a snapshot of my loaded variables at the bottom. Many thanks, Pablo >> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. Use parpool instead. Starting matlabpool using the 'local' profile ... Error using matlabpool (line 148) Failed to start a parallel pool. (For information in addition to the causing error, validate the profile 'local' in the Cluster Profile Manager.) Error in lme_mass_fit (line 128) matlabpool(prs); Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); Caused by: Error using parallel.internal.pool.InteractiveClient/start (line 330) Failed to start pool. Error using parallel.Job/submit (line 304) You requested a minimum of 8 workers, but the cluster "local" has the NumWorkers property set to allow a maximum of 4 workers. To run a communicating job on more workers than this (up to a maximum of 512 for the Local cluster), increase the value of the NumWorkers property for the cluster. The default value of NumWorkers for a Local cluster is the number of cores on the local machine. ------------------------------------------------------------------------ To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> From: mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo, you should run something like this to get the ni: [M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels hope that helps, Martin On 10/14/2015 10:43 AM, pablo najt wrote: Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh'); %---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------% 3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------% lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex); Error using lme_mass_fit (line 108) The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design matrix X Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig'); ------------------------------------------------------------------------ Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset Hi Pablo I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data. Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family. When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects. Hope that helps Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 10 de septiembre de 2015 8:07 *Asunto:* [Freesurfer] A mixed effect model approach in within subject dataset Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo From: michaelnotter@hotmail.com <mailto:michaelnotter@hotmail.com> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto: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 <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu <http://reuter.mit.edu/> _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto: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 <mailto: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 <mailto: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 <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu
_______________________________________________ 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
Hello again,I have now also run the model with 2 random effects and it is definitively worst. The final line says:Algorithm did not converge at 79.5859 percent of the total number of locations. Any suggestion if I should trust in the model with 1 random effect? I know you Martin had doubts, so I am wondering what Jorge thinks of it.ThanksPablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 21 Oct 2015 08:55:04 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} {Disarmed}
Hi Pablo,
23% still seems a lot, but may be OK (maybe Jorge can comment on that)?
If you want to do a model comparison (2 random vs 1) take a look at the wiki. It describes how to do that.
Best, Martin
On 10/21/2015 01:19 AM, pablo najt wrote:
Martin, Doing the changes in the matrix allowed me to run the command. The command with just one random effect for the intercept term give me the following output:
Aproximate percentage of fitted locations: 100%
Summary:
Algorithm did not converge at 23.407 percent of the total number of locations.
Total elapsed time is 116.2184 minutes.
Is this ok? Is there anything I should check? Should I also try two random effects as well? Thank you Pablo
To: freesurfer@nmr.mgh.harvard.edu
From: mreuter@nmr.mgh.harvard.edu
Date: Tue, 20 Oct 2015 09:06:53 -0400
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} {Disarmed}
Hi Pablo,
the zero columns could be outside the cortex and then this would be fine. Do
min(sum(lhY(:,lhcortex)))
if you get 0 you have at least one location where the column is all zero (as we should not have any negative values, you can test that with min(lhY(:))
Now even if you still have a column of 0 , then that vertex is problematic, but who cares about a single vertex. If you do a region approach, it will be part of a region anyway. so you should check how often you get 0 columns (inside cortex).
About your M: I would subtract 1 from gender (to have 0 and 1)
Also you only need 2 columns for the groups: keep only column 3 and 4. A HC is someone who has 0 in both.
Best, Martin
On 10/20/2015 04:37 AM, pablo najt wrote:
Dear FS experts, I am following up on an issue I am having when trying to run linear mixed effects model. I keep getting 'Algorithm did not converge'. So far I have changed the preprocessing steps following literally from the wiki as in message below. Also I tried to use only one random effect for the intercept. Also I have checked that sum of vectors of (ni) is equal to the number of rows in (X).
As none of the above worked I started to have concerns about the dimensions of my lh.thickness_sm10.mgh. When I load in matlab in working space says 'too many variables'. Is this ok?
Also, when I open 'lhY' variable it shows a column of Zeros every nine columns as below. Is this alright?
Perhaps this is fine but I would like to check on anything that is creating this problem. For matrix 'M' I created a matlab script that loads a matrix with intercept (all ones), HC (ones to HC otherwise zero), U_PT (relatives of patients ones otherwise zero), PT (patients one otherwise zero), gender, age (see below). The three columns for groups do not match the qdec.table.dat which have only one column for groups (HC, U_PT, PT). However the order of the rows is identical in M and qdec.table.dat
Please let me know if this should be set differently. Thank you Pablo
To: freesurfer@nmr.mgh.harvard.edu
From: mreuter@nmr.mgh.harvard.edu
Date: Mon, 19 Oct 2015 09:00:57 -0400
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
since you did not run this through the longitudinal stream, you won't have a base ( or subject-template), also you won't have <tpid>.long.<base> directories, so the first command should fail. For that you simply replace the --qdec-long with --qdec and use the same table. Should work. The second command is fine (it just smoothes the thickness stack).
Best, Martin
On 10/19/2015 06:51 AM, pablo najt wrote:
Hi Jorge and FS experts, I have run again the analysis and still get the convergence problem. I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design.
Instructions on lme in the wiki for preprocessing specify the following:
mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mgh
mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape
My data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: 1) for preprocessing should I follow instructions as it would be longitudinal data? 2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ? Thank you Pablo
____________________________________________________________________________________________________________________________________
Date: Fri, 16 Oct 2015 16:24:36 +0000
From: jbernal0019@yahoo.es
To: freesurfer@nmr.mgh.harvard.edu
CC: pablonajt@hotmail.com
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi
Pablo
Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term:
lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4);
If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly.
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com
Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu
Enviado: Viernes 16 de octubre de 2015 1:43
Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thanks you for the replies. Jorge and FS experts, I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects. The analysis gave the following 'error'(?) below: I looked up a previous thread coming across this. At that case you recommended Would you recommend this afain>
Aproximate
percentage of fitted locations: 100%
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead.
Parallel pool using the 'local' profile is shutting down.
Summary:
Algorithm did not converge at 90.0637 percent of the total number of locations.
Total elapsed time is 550.1023 minutes.
Also
almost all the time the screen showed the following message:
144114:
Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i, -241.4153-1.570796i.
Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i.
Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i.
Location
144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448.
Aproximate percentage of fitted locations: 100%
So my two questions are: 1. Is this problematic? 2. Are there any fixes to this issue? Thank
you, Pablo
Date:
Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo
The
error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this:
lhstats
= lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4);
Please
make sure that sum(ni) and length(X) give the same value before running the above command.
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu
Enviado: Jueves 15 de octubre de 2015 7:18 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Thank you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just
in case I am also including a snapshot of my loaded variables at the bottom. Many
thanks, Pablo >>
lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Warning:
matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property
of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning:
matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property
of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning:
matlabpool will be removed in a future release.
Use parpool instead.
Starting
matlabpool using the 'local' profile ...
Error using matlabpool (line 148)
Failed
to start a parallel pool. (For information in addition to the causing error,
validate
the profile 'local' in the Cluster Profile Manager.)
Error in lme_mass_fit (line 128)
matlabpool(prs);
Error in lme_mass_fit_vw (line 73)
[stats1,st1]
= lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
Caused
by:
Error using parallel.internal.pool.InteractiveClient/start (line 330)
Failed to start pool.
Error using parallel.Job/submit
(line 304)
You requested a minimum of 8 workers, but the cluster "local" has the
NumWorkers property set to allow a maximum of 4 workers. To run a
communicating job on more workers than this (up to a maximum of 512 for the
Local cluster), increase the value of the NumWorkers property for the
cluster. The default value of NumWorkers for a Local cluster is the number of
cores on the local machine.
To:
freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On
10/14/2015 10:43 AM, pablo najt wrote:
Dear
FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date:
Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to
control for the non-independence of data contributed by patients and relatives from the same families. Thanks
in advance! Pablo
From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi
everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics
of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael
_______________________________________________ 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 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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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
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 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
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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu
_______________________________________________ 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 Pablo
If you havethoroughly checked that everything is properly setup so that theordering of the design matrix's rows is in proper correspondence withthe ordering of the image data matrix's rows and it is also in propercorrespondence with the ordering of the entries in vector ni (eachentry of this vector has the number of subjects belonging to the samefamily) and the design matrix has enough covariates to describe yourdata then the analysis should be correct.
Bear in mind thatyou are not analyzing longitudinal data for which a proper imageregistration pipeline exists in Freesurfer and there is a clearwithin-subject correlation between cortical thickness values acrosstime points at the same surface vertex. Instead you are analyzingfamily data in which you are trying to model vertex-wise correlationsacross different subjects. Although in principle there should bewithin-family correlation the image registration procedure (and imageprocessing pipeline in general) as well as individual variation couldhave introduced a lot of noise in your data making the vertex-wisewithin-family correlations rather weak. This makes it hard for theLME estimation procedure to converge and give precise parameterestimates.
Best-Jorge
De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 22 de octubre de 2015 1:15 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} {Disarmed}
<!--#yiv0632092676 .yiv0632092676hmmessage P{margin:0px;padding:0px;}#yiv0632092676 body.yiv0632092676hmmessage{font-size:12pt;font-family:Calibri;}-->Hello again,I have now also run the model with 2 random effects and it is definitively worst. The final line says:Algorithm did not converge at 79.5859 percent of the total number of locations. Any suggestion if I should trust in the model with 1 random effect? I know you Martin had doubts, so I am wondering what Jorge thinks of it.ThanksPablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 21 Oct 2015 08:55:04 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} {Disarmed}
Hi Pablo,
23% still seems a lot, but may be OK (maybe Jorge can comment on that)?
If you want to do a model comparison (2 random vs 1) take a look at the wiki. It describes how to do that.
Best, Martin
On 10/21/2015 01:19 AM, pablo najt wrote:
<!--#yiv0632092676 .yiv0632092676ExternalClass .yiv0632092676ecxhmmessage P {padding:0px;}#yiv0632092676 .yiv0632092676ExternalClass body.yiv0632092676ecxhmmessage {font-size:12pt;font-family:Calibri;}--> Martin, Doing the changes in the matrix allowed me to run the command. The command with just one random effect for the intercept term give me the following output: Aproximate percentage of fitted locations: 100% Summary: Algorithm did not converge at 23.407 percent of the total number of locations. Total elapsed time is 116.2184 minutes.
Is this ok? Is there anything I should check? Should I also try two random effects as well? Thank you Pablo To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Tue, 20 Oct 2015 09:06:53 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} {Disarmed}
Hi Pablo,
the zero columns could be outside the cortex and then this would be fine. Do min(sum(lhY(:,lhcortex))) if you get 0 you have at least one location where the column is all zero (as we should not have any negative values, you can test that with min(lhY(:)) Now even if you still have a column of 0 , then that vertex is problematic, but who cares about a single vertex. If you do a region approach, it will be part of a region anyway. so you should check how often you get 0 columns (inside cortex).
About your M: I would subtract 1 from gender (to have 0 and 1) Also you only need 2 columns for the groups: keep only column 3 and 4. A HC is someone who has 0 in both.
Best, Martin
On 10/20/2015 04:37 AM, pablo najt wrote:
<!--#yiv0632092676 .yiv0632092676ExternalClass .yiv0632092676ecxhmmessage P {padding:0px;}#yiv0632092676 .yiv0632092676ExternalClass body.yiv0632092676ecxhmmessage {font-size:12pt;font-family:Calibri;}--> Dear FS experts, I am following up on an issue I am having when trying to run linear mixed effects model. I keep getting 'Algorithm did not converge'. So far I have changed the preprocessing steps following literally from the wiki as in message below. Also I tried to use only one random effect for the intercept. Also I have checked that sum of vectors of (ni) is equal to the number of rows in (X). As none of the above worked I started to have concerns about the dimensions of my lh.thickness_sm10.mgh. When I load in matlab in working space says 'too many variables'. Is this ok?
Also, when I open 'lhY' variable it shows a column of Zeros every nine columns as below. Is this alright?
Perhaps this is fine but I would like to check on anything that is creating this problem. For matrix 'M' I created a matlab script that loads a matrix with intercept (all ones), HC (ones to HC otherwise zero), U_PT (relatives of patients ones otherwise zero), PT (patients one otherwise zero), gender, age (see below). The three columns for groups do not match the qdec.table.dat which have only one column for groups (HC, U_PT, PT). However the order of the rows is identical in M and qdec.table.dat
Please let me know if this should be set differently. Thank you Pablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Mon, 19 Oct 2015 09:00:57 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
since you did not run this through the longitudinal stream, you won't have a base ( or subject-template), also you won't have <tpid>.long.<base> directories, so the first command should fail. For that you simply replace the --qdec-long with --qdec and use the same table. Should work. The second command is fine (it just smoothes the thickness stack).
Best, Martin
On 10/19/2015 06:51 AM, pablo najt wrote:
<!--#yiv0632092676 .yiv0632092676ExternalClass .yiv0632092676ecxhmmessage P {padding:0px;}#yiv0632092676 .yiv0632092676ExternalClass body.yiv0632092676ecxhmmessage {font-size:12pt;font-family:Calibri;}--> Hi Jorge and FS experts, I have run again the analysis and still get the convergence problem. I am assuming the issue has to do with the preprocessing steps, as I had doubts on how to follow from the instructions if I am not using qdec and have a cross sectional design. Instructions on lme in the wiki for preprocessing specify the following: mris_preproc --qdec-long qdec.table.dat --target study_average --hemi lh --meas thickness --out lh.thickness.mgh mri_surf2surf --hemi lh --s study_average --sval lh.thickness.mgh --tval lh.thickness_sm10.mgh --fwhm-trg 10 --cortex --noreshape My data -as I explained below- is cross section but I want to treat it as longitudinal as I want to analyse subjects belonging to the same family. Here is my 2 main questions: 1) for preprocessing should I follow instructions as it would be longitudinal data? 2) Although my design does not allow using qdec (3 groups) do I still create a qdec table and follow literally the two instructions above ? Thank you Pablo ____________________________________________________________________________________________________________________________________
Date: Fri, 16 Oct 2015 16:24:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo Yes, too many locations at which the estimation algorithm didn't converge is problematic. That might mean that having two random effects is not appropriate for your data. You should try to run the command with just a single random effect for the intercept term: lhstats = lme_mass_fit_vw(X, [1], lhY, ni, lhcortex, [], 4); If the result still have too many non-convergence locations then something might be wrong with the ordering of the design matrix and its correspondence with the ordering of the ni vector or the ordering of the image data etc... You will need to check it thoroughly. Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Viernes 16 de octubre de 2015 1:43 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
<!--#yiv0632092676 .yiv0632092676ExternalClass #yiv0632092676ecxyiv7275178360 body.yiv0632092676ecxyiv7275178360hmmessage {font-size:12pt;font-family:Calibri;}--> Thanks you for the replies. Jorge and FS experts, I have run the analysis and first double checked that the sum of vectors of (ni) is equal to the number of rows in (X). Both are 140 which is the number of my subjects. The analysis gave the following 'error'(?) below: I looked up a previous thread coming across this. At that case you recommended Would you recommend this afain>
Aproximate percentage of fitted locations: 100% Warning: matlabpool will be removed in a future release. To query the size of an already started parallel pool, query the 'NumWorkers' property of the pool. To check if a pool is already started use 'isempty(gcp('nocreate'))'. Warning: matlabpool will be removed in a future release. To shutdown a parallel pool use 'delete(gcp('nocreate'))' instead. Parallel pool using the 'local' profile is shutting down. Summary: Algorithm did not converge at 90.0637 percent of the total number of locations. Total elapsed time is 550.1023 minutes.
Also almost all the time the screen showed the following message: 144114: Algorithm did not converge. Initial and final likelihoods: -38.3408-1.5708i,-241.4153-1.570796i. Location 144113: Algorithm did not converge. Initial and final likelihoods: -5.5424-1.5708i, -133.8004-1.570796i. Location 144112: Algorithm did not converge. Initial and final likelihoods: -7.7571-1.5708i, -319.1378-1.570796i. Location 144111: Algorithm did not converge. Initial and final likelihoods: -16.8597-1.5708i, 0.74448. Aproximate percentage of fitted locations: 100%
So my two questions are: 1. Is this problematic? 2. Are there any fixes to this issue? Thank you, Pablo Date: Thu, 15 Oct 2015 13:38:22 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu CC: pablonajt@hotmail.com Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo The error you are getting is because in your Matlab setup you can only request a maximum of 4 matlab parallel workers and by default lme requests 8. So you just need to modify your command like this: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex, [], 4); Please make sure that sum(ni) and length(X) give the same value before running the above command. Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: Freesurfer support list freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 15 de octubre de 2015 7:18 Asunto: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
<!----> Thank you Martin. I am trying to run the following command line and get the error below. Would you have a suggestion to overcome this issue? Just in case I am also including a snapshot of my loaded variables at the bottom. Many thanks, Pablo >> lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
To query the size of an already started parallel pool, query the 'NumWorkers'
property of the pool.
To check if a pool is already started use 'isempty(gcp('nocreate'))'.
Warning: matlabpool will be removed in a future release.
Use parpool instead.
Starting matlabpool using the 'local' profile ...
Error using matlabpool (line 148)
Failed to start a parallel pool. (For information in addition to the causing error,
validate the profile 'local' in the Cluster Profile Manager.)
Error in lme_mass_fit (line 128)
matlabpool(prs);
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
Caused by:
Error using parallel.internal.pool.InteractiveClient/start (line 330)
Failed to start pool.
Error using parallel.Job/submit (line 304)
You requested a minimum of 8 workers, but the cluster "local" has the
NumWorkers property set to allow a maximum of 4 workers. To run a
communicating job on more workers than this (up to a maximum of 512 for the
Local cluster), increase the value of the NumWorkers property for the
cluster. The default value of NumWorkers for a Local cluster is the number of
cores on the local machine.
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
<!----> Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset Hi Pablo I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data. Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family. When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects. Hope that helps Cheers -Jorge De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset <!----> Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach <!----> Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael _______________________________________________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 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 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 -- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu _______________________________________________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
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 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
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 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu _______________________________________________ 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu _______________________________________________ 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
-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web : http://reuter.mit.edu _______________________________________________Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurferThe information in this e-mail is intended only for the person to whom it isaddressed. If you believe this e-mail was sent to you in error and the e-mailcontains patient information, please contact the Partners Compliance HelpLine athttp://www.partners.org/complianceline . If the e-mail was sent to you in errorbut does not contain patient information, please contact the sender and properlydispose of the e-mail. _______________________________________________ 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.
Thank you for your input. I noticed that if I follow literally all the steps in the wiki, my data which is ordered by variable 'family' (instead of subjects, in my case is number of members belonging to e.g. family_1) is shuffled. This happens after I run the command sortData below. Especially I noticed that ni and X do not match sID.It would be really helpful to know what is this command doing. I am wondering whether my data differs in number of columns or else and because of this I end with a shuffled data. Any suggestion or tips to figure what could be happening?ThanksPablo [M, Y, ni] = sortData(M,1,Y,sID) To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see
https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.:
lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label');
2-Read the data file eg.:
[lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.:
lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users,I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families.Thanks in advance!Pablo From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi everybody,
I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem:
1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17].
2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one.
3. I could use a mixed effect model approach and specify site as a random effect.
If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches?
I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows:
Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ...
As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output:
stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex);
Starting matlabpool using the 'local' profile ... connected to 8 workers.
Starting model fitting at each location ...
Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ...
I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence.
Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate?
Thanks, Michael
_______________________________________________ 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
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
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
Hi Pablo,
the sortData function sorts the rows so that entries from the same subject (in your case same family) are blocked and that within each block the time variable (2nd parameter specifies which column that is in your M matrix, in your case the first =1) is increasing. It is important, when creating your design matrix X, that ordering agrees with Y. That is guaranteed if you generate X from M (which is ordered like Y after the sort command).
Best, Martin
On 10/27/2015 01:32 PM, pablo najt wrote:
Thank you for your input. I noticed that if I follow literally all the steps in the wiki, my data which is ordered by variable 'family' (instead of subjects, in my case is number of members belonging to e.g. family_1) is shuffled. This happens after I run the command sortData below. Especially I noticed that ni and X do not match sID. It would be really helpful to know what is this command doing. I am wondering whether my data differs in number of columns or else and because of this I end with a shuffled data. Any suggestion or tips to figure what could be happening? Thanks Pablo
[M, Y, ni] = sortData(M,1,Y,sID)
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh'); %---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------% 3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------% lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex); Error using lme_mass_fit (line 108) The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design matrix X Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig'); ------------------------------------------------------------------------ Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset Hi Pablo I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data. Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family. When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects. Hope that helps Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 10 de septiembre de 2015 8:07 *Asunto:* [Freesurfer] A mixed effect model approach in within subject dataset Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo From: michaelnotter@hotmail.com <mailto:michaelnotter@hotmail.com> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) <http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels%29> but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto: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 <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 <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu _______________________________________________ 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
Dear Martin and FS experts.I have a quick query about how to obtain my design matrix of group by age from the following matlab command listed in the lme wiki. In the wiki example the matrix is created from variables 1 and 2 (group by time). X = [ones(length(M),1) M M(:,1).*M(:,2); In my case I have a combination of variables 1 and 2 for group (relatives of PT and PT) and 5 (age). I would really appreciate if you could kindly advice on how to adapt the command above to my design. Thanks Pablo To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Tue, 27 Oct 2015 17:13:51 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
the sortData function sorts the rows so that entries from the same subject (in your case same family) are blocked and that within each block the time variable (2nd parameter specifies which column that is in your M matrix, in your case the first =1) is increasing.
It is important, when creating your design matrix X, that ordering agrees with Y. That is guaranteed if you generate X from M (which is ordered like Y after the sort command).
Best, Martin
On 10/27/2015 01:32 PM, pablo najt wrote:
Thank you for your input. I noticed that if I follow literally all the steps in the wiki, my data which is ordered by variable 'family' (instead of subjects, in my case is number of members belonging to e.g. family_1) is shuffled. This happens after I run the command sortData below. Especially I noticed that ni and X do not match sID. It would be really helpful to know what is this command doing. I am wondering whether my data differs in number of columns or else and because of this I end with a shuffled data. Any suggestion or tips to figure what could be happening? Thanks Pablo
[M, Y, ni] = sortData(M,1,Y,sID)
To: freesurfer@nmr.mgh.harvard.edu
From: mreuter@nmr.mgh.harvard.edu
Date: Wed, 14 Oct 2015 10:54:41 -0400
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see
https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.:
lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label');
2-Read the data file eg.:
[lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.:
lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users,I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families.Thanks in advance!Pablo From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi everybody,
I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem:
1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17].
2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one.
3. I could use a mixed effect model approach and specify site as a random effect.
If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches?
I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows:
Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ...
As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output:
stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex);
Starting matlabpool using the 'local' profile ... connected to 8 workers.
Starting model fitting at each location ...
Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ...
I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence.
Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate?
Thanks, Michael
_______________________________________________ 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
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
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
Dear Pablo, maybe what you want is a group x age interaction for each of those groups? (so 1x5 and 2x5)? Not sure. It is your hypothesis, your model and we usually don't give advise on model building. Since you have a special case (within family analysis) I would recommend you try to get some advise from a bio-statistician, if unsure what you are doing.
Best, Martin
On 11/03/2015 12:01 PM, pablo najt wrote:
Dear Martin and FS experts. I have a quick query about how to obtain my design matrix of /group by age/ from the following matlab command listed in the lme wiki. In the wiki example the matrix is created from variables 1 and 2 (group by time).
X = [ones(length(M),1) M M(:,1).*M(:,2);In my case I have a combination of variables 1 and 2 for group (relatives of PT and PT) and 5 (age). I would really appreciate if you could kindly advice on how to adapt the command above to my design. Thanks Pablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Tue, 27 Oct 2015 17:13:51 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
the sortData function sorts the rows so that entries from the same subject (in your case same family) are blocked and that within each block the time variable (2nd parameter specifies which column that is in your M matrix, in your case the first =1) is increasing. It is important, when creating your design matrix X, that ordering agrees with Y. That is guaranteed if you generate X from M (which is ordered like Y after the sort command).
Best, Martin
On 10/27/2015 01:32 PM, pablo najt wrote:
Thank you for your input. I noticed that if I follow literally all the steps in the wiki, my data which is ordered by variable 'family' (instead of subjects, in my case is number of members belonging to e.g. family_1) is shuffled. This happens after I run the command sortData below. Especially I noticed that ni and X do not match sID. It would be really helpful to know what is this command doing. I am wondering whether my data differs in number of columns or else and because of this I end with a shuffled data. Any suggestion or tips to figure what could be happening? Thanks Pablo [M, Y, ni] = sortData(M,1,Y,sID) ------------------------------------------------------------------------ To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> From: mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo, you should run something like this to get the ni: [M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels hope that helps, Martin On 10/14/2015 10:43 AM, pablo najt wrote: Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh'); %---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------% 3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------% lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex); Error using lme_mass_fit (line 108) The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design matrix X Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig'); ------------------------------------------------------------------------ Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset Hi Pablo I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data. Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family. When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects. Hope that helps Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 10 de septiembre de 2015 8:07 *Asunto:* [Freesurfer] A mixed effect model approach in within subject dataset Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo From: michaelnotter@hotmail.com <mailto:michaelnotter@hotmail.com> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) <http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels%29> but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto: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 <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 <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu _______________________________________________ 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
Thank you Martin. My intention was not to enquire about advice on the statistical model but instead on how to run the matlab commad with two experimental groups. My decision is to test group by age interaction for each group. The way to proceed for this would be to create two separate matrices? Thanks Pablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Tue, 3 Nov 2015 12:10:17 -0500 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Dear Pablo,
maybe what you want is a group x age interaction for each of those groups? (so 1x5 and 2x5)?
Not sure. It is your hypothesis, your model and we usually don't give advise on model building. Since you have a special case (within family analysis) I would recommend you try to get some advise from a bio-statistician, if unsure what you are doing.
Best, Martin
On 11/03/2015 12:01 PM, pablo najt wrote:
Dear Martin and FS experts. I have a quick query about how to obtain my design matrix of group by age from the following matlab command listed in the lme wiki. In the wiki example the matrix is created from variables 1 and 2 (group by time).
X = [ones(length(M),1) M M(:,1).*M(:,2);
In my case I have a combination of variables 1 and 2 for group (relatives of PT and PT) and 5 (age).
I would really appreciate if you could kindly advice on how to adapt the command above to my design.
Thanks
Pablo
To: freesurfer@nmr.mgh.harvard.edu
From: mreuter@nmr.mgh.harvard.edu
Date: Tue, 27 Oct 2015 17:13:51 -0400
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
the sortData function sorts the rows so that entries from the same subject (in your case same family) are blocked and that within each block the time variable (2nd parameter specifies which column that is in your M matrix, in your case the first =1) is increasing.
It is important, when creating your design matrix X, that ordering agrees with Y. That is guaranteed if you generate X from M (which is ordered like Y after the sort command).
Best, Martin
On 10/27/2015 01:32 PM, pablo najt wrote:
Thank you for your input. I noticed that if I follow literally all the steps in the wiki, my data which is ordered by variable 'family' (instead of subjects, in my case is number of members belonging to e.g. family_1) is shuffled. This happens after I run the command sortData below. Especially I noticed that ni and X do not match sID. It would be really helpful to know what is this command doing. I am wondering whether my data differs in number of columns or else and because of this I end with a shuffled data. Any suggestion or tips to figure what could be happening? Thanks Pablo
[M, Y, ni] = sortData(M,1,Y,sID)
To: freesurfer@nmr.mgh.harvard.edu
From: mreuter@nmr.mgh.harvard.edu
Date: Wed, 14 Oct 2015 10:54:41 -0400
Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Hi Pablo,
you should run something like this to get the ni:
[M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see
https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
hope that helps, Martin
On 10/14/2015 10:43 AM, pablo najt wrote:
Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo
Start analysis as follows:
1-Read your label eg.:
lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label');
2-Read the data file eg.:
[lhY, lhmri] = fs_read_Y('lh.thickness.mgh');
%---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------%
3-Fit a vertex-wise lme model with random effects.:
lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex);
Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------%
lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex);
Error using lme_mass_fit (line 108)
The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design
matrix X
Error in lme_mass_fit_vw (line 73)
[stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e);
My matrix is organised in "family", "group", Sex" and "age" columns".
4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig');
Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es To: freesurfer@nmr.mgh.harvard.edu Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset
Hi Pablo
I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data.
Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family.
When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects.
Hope that helps
Cheers -Jorge
De: pablo najt pablonajt@hotmail.com Para: "freesurfer@nmr.mgh.harvard.edu" freesurfer@nmr.mgh.harvard.edu Enviado: Jueves 10 de septiembre de 2015 8:07 Asunto: [Freesurfer] A mixed effect model approach in within subject dataset
Dear Freesurfer users,I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families.Thanks in advance!Pablo From: michaelnotter@hotmail.com To: freesurfer@nmr.mgh.harvard.edu Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach
Hi everybody,
I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem:
1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17].
2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one.
3. I could use a mixed effect model approach and specify site as a random effect.
If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches?
I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows:
Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ...
As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output:
stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex);
Starting matlabpool using the 'local' profile ... connected to 8 workers.
Starting model fitting at each location ...
Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ...
I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence.
Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate?
Thanks, Michael
_______________________________________________ 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
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
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
Hi Pablo,
setting up that matrix is designing your model. You would probably add the 2 columns (1x5 and 2x5) to your matrix. Having independent matrices for each group, is a different model (actually 2 different ones).
Best, Martin
On 11/03/2015 04:33 PM, pablo najt wrote:
Thank you Martin. My intention was not to enquire about advice on the statistical model but instead on how to run the matlab commad with two experimental groups. My decision is to test group by age interaction for each group. The way to proceed for this would be to create two separate matrices? Thanks Pablo
To: freesurfer@nmr.mgh.harvard.edu From: mreuter@nmr.mgh.harvard.edu Date: Tue, 3 Nov 2015 12:10:17 -0500 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed}
Dear Pablo, maybe what you want is a group x age interaction for each of those groups? (so 1x5 and 2x5)? Not sure. It is your hypothesis, your model and we usually don't give advise on model building. Since you have a special case (within family analysis) I would recommend you try to get some advise from a bio-statistician, if unsure what you are doing.
Best, Martin
On 11/03/2015 12:01 PM, pablo najt wrote:
Dear Martin and FS experts. I have a quick query about how to obtain my design matrix of /group by age/ from the following matlab command listed in the lme wiki. In the wiki example the matrix is created from variables 1 and 2 (group by time). X = [ones(length(M),1) M M(:,1).*M(:,2); In my case I have a combination of variables 1 and 2 for group (relatives of PT and PT) and 5 (age). I would really appreciate if you could kindly advice on how to adapt the command above to my design. Thanks Pablo ------------------------------------------------------------------------ To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> From: mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> Date: Tue, 27 Oct 2015 17:13:51 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo, the sortData function sorts the rows so that entries from the same subject (in your case same family) are blocked and that within each block the time variable (2nd parameter specifies which column that is in your M matrix, in your case the first =1) is increasing. It is important, when creating your design matrix X, that ordering agrees with Y. That is guaranteed if you generate X from M (which is ordered like Y after the sort command). Best, Martin On 10/27/2015 01:32 PM, pablo najt wrote: Thank you for your input. I noticed that if I follow literally all the steps in the wiki, my data which is ordered by variable 'family' (instead of subjects, in my case is number of members belonging to e.g. family_1) is shuffled. This happens after I run the command sortData below. Especially I noticed that ni and X do not match sID. It would be really helpful to know what is this command doing. I am wondering whether my data differs in number of columns or else and because of this I end with a shuffled data. Any suggestion or tips to figure what could be happening? Thanks Pablo [M, Y, ni] = sortData(M,1,Y,sID) ------------------------------------------------------------------------ To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> From: mreuter@nmr.mgh.harvard.edu <mailto:mreuter@nmr.mgh.harvard.edu> Date: Wed, 14 Oct 2015 10:54:41 -0400 Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset {Disarmed} Hi Pablo, you should run something like this to get the ni: [M,Y,ni] = sortData(M,1,Y,sID); # (sorts the data) see https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels hope that helps, Martin On 10/14/2015 10:43 AM, pablo najt wrote: Dear FS experts. I have query about a relating to a previous email (below). I am aiming to run a LME analysis on cross-sectional data from different families and have variable 'family' (number of families) as my NI vector. My design has three groups and therefore I am not able to use qdec. I am running the matlab commands below and finding some difficulty would really appreciate if you could help out. Thanks Pablo Start analysis as follows: 1-Read your label eg.: lhcortex = fs_read_label('freesurfer/subjects/fsaverage/label/lh.cortex.label'); 2-Read the data file eg.: [lhY, lhmri] = fs_read_Y('lh.thickness.mgh'); %---------------------I input the concatenated .mgh image from preproc and mris_surf2surf-----------------------------------------------------------------------% 3-Fit a vertex-wise lme model with random effects.: lhstats = lme_mass_fit_vw(X, [1 2], lhY, ni, lhcortex); Here I am getting the following problems: %-------------------- If I use number of families as my ni get the following------------------------------------------------------------------------------------------------% lhstats = lme_mass_fit_vw(X, [1 2], lhY, 82, lhcortex); Error using lme_mass_fit (line 108) The total number of measurements, indicated by sum(ni), mustbe the same as the number of rows of the design matrix X Error in lme_mass_fit_vw (line 73) [stats1,st1] = lme_mass_fit(X,[],Xrows,Zcols,Y,ni,prs,e); My matrix is organised in "family", "group", Sex" and "age" columns". 4-Perform vertex-wise inference eg.: CM.C = [your contrast matrix]; F_lhstats = lme_mass_F(lhstats, CM); 5-Save results eg.: fs_write_fstats(F_lhstats, lhmri,' sig.mgh', 'sig'); ------------------------------------------------------------------------ Date: Thu, 10 Sep 2015 13:44:36 +0000 From: jbernal0019@yahoo.es <mailto:jbernal0019@yahoo.es> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] A mixed effect model approach in within subject dataset Hi Pablo I think you can use LME to analyze your data by ordering the rows of your design matrix appropriately. You can consider all subjects belonging to the same family as if they were a single subject in a longitudinal analysis. You can put in your design matrix all subjects belonging to family1 first, then all subjects belonging to family 2 and so on. Then the 'ni' required by lme_mass_fit_vw is a vector with the number of subjects in each family as its entries (ordered according to your design matrix). So the length of the 'ni' vector is equal to the number of different families in your data. Now you can go further and additionally order the rows of your design matrix within each family by age. This will allow you to test the effect of age within family. When choosing the random effects for your statistical model remember that a random effect can only be the intercept term or any covariate that varies within family. For example you can compare a model with a single random effect for the intercept term against the same model but considering both the intercept term and age as random effects. Hope that helps Cheers -Jorge ------------------------------------------------------------------------ *De:* pablo najt <pablonajt@hotmail.com> <mailto:pablonajt@hotmail.com> *Para:* "freesurfer@nmr.mgh.harvard.edu" <mailto:freesurfer@nmr.mgh.harvard.edu> <freesurfer@nmr.mgh.harvard.edu> <mailto:freesurfer@nmr.mgh.harvard.edu> *Enviado:* Jueves 10 de septiembre de 2015 8:07 *Asunto:* [Freesurfer] A mixed effect model approach in within subject dataset Dear Freesurfer users, I wanted to enquire if anyone had successfully been able to implement Bernal's Linear Mixed Effects (LME) Models in cross-section dataset *not longitudinal* (please see previous thread below). I am willing to perform a LME (3 groups (HC, PT and Unaffected_relatives) and 3 covariates (sex, age, and family) with "family" variable been a within-subject factor. LME will allow to control for the non-independence of data contributed by patients and relatives from the same families. Thanks in advance! Pablo From: michaelnotter@hotmail.com <mailto:michaelnotter@hotmail.com> To: freesurfer@nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Date: Wed, 19 Feb 2014 13:10:09 +0100 Subject: [Freesurfer] Analysis of structural data acquired from multiple sites by using a mixed effect model approach Hi everybody, I want to compare the surface data of 3 groups (GroupA, GroupB and Controlls) but have the problem that they were acquired from 4 different scanner sites. As I can see it, there are three ways how I could tackle this problem: 1. I could use mri_glmfit and create a qdec table / fsgd-file with 12 classes: Class GroupA_site1; Class GroupA_site2,... And then use the contrasts [0.25 0.25 0.25 0.25 0 0 0 0 -0.25 -0.25 -0.25 -0.25] to compare GroupA to the Controlls. My Problem with this approach is, that the sites don't contribute the same amount of subjects to the analysis. I'm not sure if this could be handled by simply using a weighted contrast. Meaning, if Site1 and Site2 had twice as many subjects than Site3 and Site4, I could modify the contrast to [0.33 0.33 0.17 0.17 0 0 0 0 -0.33 -0.33 -0.17 -0.17]. 2. I could create dummy variables to account for the variability between sites. In this case, I only need to specify 3 classes (Class GroupA; Class GroupB; Class Controlls) in my fsgd-file. And I use a design matrix that has 4 dummy variables at the end, which specify to which site a subject belongs. This approach might work, but I'm not confident that it is the right one. 3. I could use a mixed effect model approach and specify site as a random effect. If I understand it correctly, the mixed effect model approach would be the best one, as it accounts for the variability within sites. Is that correct or are there other issues/better approaches? I tried to implement a mixed effect model by using Bernal's Linear Mixed Effects (LME) Models (http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) <http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels%29> but run into some problems. I'm not sure if LME can only be applied on longitudinal data or if my implementation is wrong. I have a design matrix X that specifies the characteristics of each subject per row as follows: Intercept GroupA GroupB Controll Age IQ Site1 Site2 Site3 Site4 1 1 0 0 11.1 99 0 0 1 0 1 0 1 0 11.1 101 0 0 1 0 1 1 0 0 11.4 95 1 0 0 0 1 0 0 1 12.4 100 1 0 0 0 ... As I have no repeated measures, 'ni' is just a vector with length X containing '1's. If I do now the vertex-wise linear mixed-effects estimation, I get the following output: >> stats = lme_mass_fit_vw(X,[7 8 9 10],Y,ni,lhcortex); Starting matlabpool using the 'local' profile ... connected to 8 workers. Starting model fitting at each location ... Location 24994: Index exceeds matrix dimensions. Location 24994: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. Location 62484: Index exceeds matrix dimensions. Location 62484: Algorithm did not converge. Initial and final likelihoods: -10000000000, -10000000000. ... I've checked the matrix dimensions of X, Y, ni and lhcortex and compared them to the LME mass_univariate example stored in ADNI_Long_50sMCI_vs_50cMCI.mat but couldn't find any divergence. Has anybody encountered similar problems? Is my approach of specifying 'ni' as a vector of'1's even legitimate? Thanks, Michael _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 *MailScanner has detected a possible fraud attempt from "www.partners.org" claiming to be* http://www.partners.org/complianceline <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 <mailto: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 *MailScanner has detected a possible fraud attempt from "www.partners.org" claiming to be* http://www.partners.org/complianceline <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 <mailto: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 *MailScanner has detected a possible fraud attempt from "www.partners.org" claiming to be* http://www.partners.org/complianceline <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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 *MailScanner has detected a possible fraud attempt from "www.partners.org" claiming to be* http://www.partners.org/complianceline <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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <mailto: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 *MailScanner has detected a possible fraud attempt from "www.partners.org" claiming to be* http://www.partners.org/complianceline <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 <mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer-- Martin Reuter, PhD Assistant Professor of Radiology, Harvard Medical School Assistant Professor of Neurology, Harvard Medical School A.A.Martinos Center for Biomedical Imaging Massachusetts General Hospital Research Affiliate, CSAIL, MIT Phone: +1-617-724-5652 Web :http://reuter.mit.edu _______________________________________________ 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
freesurfer@nmr.mgh.harvard.edu
-
jorge luis -
Martin Reuter -
pablo najt