External Email - Use Caution
Hello,
I recently performed an LME longitudinal mass-univariate (surface) analysis and am now trying to do something similar, but with subcortical (aseg) ROIs instead.
I first created an aseg table with the following command:
* asegstats2table \ --sd $subjects_dir \ --qdec-long /path/to/long.qdec.table.dat \ --segno 4 10 11 12 17 18 26 43 49 50 51 53 54 58 251 253 255 \ --stats aseg.stats \ --tablefile $output_file*
The LME tutorial ( https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) provides information and examples of both a univariate and mass-univariate approach. While univariate seems most appropriate in this case, the example details design matrix setup and testing of a single ROI (hippocampus), whereas I would like to assess several subcortical ROIs.
Would I simply create a loop structure to set up design matrices for each ROI (both hemispheres) and test them individually, or can I perform a mass-univariate approach to assess all ROIs at once?
Thanks,
Dan
Hi Dan,
You could do either one. I would recommend the mass univariate for multiple independent test with the same design.
Best, Martin
On 26. Jun 2024, at 17:20, Dan Levitas djlevitas208@gmail.com wrote:
I recently performed an LME longitudinal mass-univariate (surface) analysis and am now trying to do something similar, but with subcortical (aseg) ROIs instead.
I first created an aseg table with the following command: asegstats2table \ --sd $subjects_dir \ --qdec-long /path/to/long.qdec.table.dat \ --segno 4 10 11 12 17 18 26 43 49 50 51 53 54 58 251 253 255 \ --stats aseg.stats \ --tablefile $output_file
The LME tutorial (https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) provides information and examples of both a univariate and mass-univariate approach. While univariate seems most appropriate in this case, the example details design matrix setup and testing of a single ROI (hippocampus), whereas I would like to assess several subcortical ROIs.
Would I simply create a loop structure to set up design matrices for each ROI (both hemispheres) and test them individually, or can I perform a mass-univariate approach to assess all ROIs at once?
Thanks,
Dan
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
External Email - Use Caution
Hi Martin,
Thanks for your reply, it's good to know that the mass-univariate approach is still appropriate here.
Since the aseg segmentations are volumetric, can I still use the *mris_preproc* and *mri_surf2surf* functions? If so, would the mri_preproc function look something like *mri_preproc --qdec-long /path/to/asegstats2table.long.table --target fsaverage --hemi rh --meas volume --out output_dir/rh.volume.mgh*
If not, and I need separate functions, would *mri_surf2surf* be replaced by *mri_vol2surf*?
On Thu, Jun 27, 2024 at 4:51 AM Reuter, Martin,Ph.D. < MREUTER@mgh.harvard.edu> wrote:
Hi Dan,
You could do either one. I would recommend the mass univariate for multiple independent test with the same design.
Best, Martin
On 26. Jun 2024, at 17:20, Dan Levitas djlevitas208@gmail.com wrote:
I recently performed an LME longitudinal mass-univariate (surface) analysis and am now trying to do something similar, but with subcortical (aseg) ROIs instead.
I first created an aseg table with the following command:
- asegstats2table \ --sd $subjects_dir \ --qdec-long
/path/to/long.qdec.table.dat \ --segno 4 10 11 12 17 18 26 43 49 50 51 53 54 58 251 253 255 \ --stats aseg.stats \ --tablefile $output_file*
The LME tutorial ( https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) provides information and examples of both a univariate and mass-univariate approach. While univariate seems most appropriate in this case, the example details design matrix setup and testing of a single ROI (hippocampus), whereas I would like to assess several subcortical ROIs.
Would I simply create a loop structure to set up design matrices for each ROI (both hemispheres) and test them individually, or can I perform a mass-univariate approach to assess all ROIs at once?
Thanks,
Dan
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer 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 Mass General Brigham Compliance HelpLine at https://www.massgeneralbrigham.org/complianceline < https://www.massgeneralbrigham.org/complianceline%3E .
External Email - Use Caution
Hi Martin,
I wanted to ping this thread again, to confirm that my proposed mass-univariate workflow using the subcortical segmented ROIs is appropriate for the LME longitudinal analysis.
Thanks again.
Dan
On Thu, Jun 27, 2024 at 5:44 AM Dan Levitas djlevitas208@gmail.com wrote:
Hi Martin,
Thanks for your reply, it's good to know that the mass-univariate approach is still appropriate here.
Since the aseg segmentations are volumetric, can I still use the *mris_preproc* and *mri_surf2surf* functions? If so, would the mri_preproc function look something like *mri_preproc --qdec-long /path/to/asegstats2table.long.table --target fsaverage --hemi rh --meas volume --out output_dir/rh.volume.mgh*
If not, and I need separate functions, would *mri_surf2surf* be replaced by *mri_vol2surf*?
On Thu, Jun 27, 2024 at 4:51 AM Reuter, Martin,Ph.D. < MREUTER@mgh.harvard.edu> wrote:
Hi Dan,
You could do either one. I would recommend the mass univariate for multiple independent test with the same design.
Best, Martin
On 26. Jun 2024, at 17:20, Dan Levitas djlevitas208@gmail.com wrote:
I recently performed an LME longitudinal mass-univariate (surface) analysis and am now trying to do something similar, but with subcortical (aseg) ROIs instead.
I first created an aseg table with the following command:
- asegstats2table \ --sd $subjects_dir \ --qdec-long
/path/to/long.qdec.table.dat \ --segno 4 10 11 12 17 18 26 43 49 50 51 53 54 58 251 253 255 \ --stats aseg.stats \ --tablefile $output_file*
The LME tutorial ( https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) provides information and examples of both a univariate and mass-univariate approach. While univariate seems most appropriate in this case, the example details design matrix setup and testing of a single ROI (hippocampus), whereas I would like to assess several subcortical ROIs.
Would I simply create a loop structure to set up design matrices for each ROI (both hemispheres) and test them individually, or can I perform a mass-univariate approach to assess all ROIs at once?
Thanks,
Dan
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer 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 Mass General Brigham Compliance HelpLine at https://www.massgeneralbrigham.org/complianceline < https://www.massgeneralbrigham.org/complianceline%3E .
Hi Dan,
No, mris_preproc and also surf2surf are functions that work on surfaces. You would just import the volume statistics into Matlab and fill the corresponding columns in your matrix directly, similar to using co-variates such as eTIV.
Best, Martin
On 1. Jul 2024, at 14:03, Dan Levitas djlevitas208@gmail.com wrote:
Hi Martin,
I wanted to ping this thread again, to confirm that my proposed mass-univariate workflow using the subcortical segmented ROIs is appropriate for the LME longitudinal analysis.
Thanks again.
Dan
On Thu, Jun 27, 2024 at 5:44 AM Dan Levitas <djlevitas208@gmail.commailto:djlevitas208@gmail.com> wrote: Hi Martin,
Thanks for your reply, it's good to know that the mass-univariate approach is still appropriate here.
Since the aseg segmentations are volumetric, can I still use the mris_preproc and mri_surf2surf functions? If so, would the mri_preproc function look something like mri_preproc --qdec-long /path/to/asegstats2table.long.table --target fsaverage --hemi rh --meas volume --out output_dir/rh.volume.mgh
If not, and I need separate functions, would mri_surf2surf be replaced by mri_vol2surf?
On Thu, Jun 27, 2024 at 4:51 AM Reuter, Martin,Ph.D. <MREUTER@mgh.harvard.edumailto:MREUTER@mgh.harvard.edu> wrote: Hi Dan,
You could do either one. I would recommend the mass univariate for multiple independent test with the same design.
Best, Martin
On 26. Jun 2024, at 17:20, Dan Levitas <djlevitas208@gmail.commailto:djlevitas208@gmail.com> wrote:
I recently performed an LME longitudinal mass-univariate (surface) analysis and am now trying to do something similar, but with subcortical (aseg) ROIs instead.
I first created an aseg table with the following command: asegstats2table \ --sd $subjects_dir \ --qdec-long /path/to/long.qdec.table.dat \ --segno 4 10 11 12 17 18 26 43 49 50 51 53 54 58 251 253 255 \ --stats aseg.stats \ --tablefile $output_file
The LME tutorial (https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) provides information and examples of both a univariate and mass-univariate approach. While univariate seems most appropriate in this case, the example details design matrix setup and testing of a single ROI (hippocampus), whereas I would like to assess several subcortical ROIs.
Would I simply create a loop structure to set up design matrices for each ROI (both hemispheres) and test them individually, or can I perform a mass-univariate approach to assess all ROIs at once?
Thanks,
Dan
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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 Mass General Brigham Compliance HelpLine at https://www.massgeneralbrigham.org/complianceline https://www.massgeneralbrigham.org/complianceline .
External Email - Use Caution
Thanks Martin,
I just want to clarify that this then is the correct workflow:
1). Import the table generated from *asegstats2table* (asegstats2table.long.table), and select the column pertaining to the ROIs and eTIV. 2). Add to this table any covariates, as well as the group and time columns. 3). Specify a group x time interaction contrast (e.g. CM.C = [0 0 0 0 0 0 0 0 0 0 0 0 1];). 4). Use the table (design matrix) to perform LME.
Assuming the above is correct, Is there a FreeSurfer (or Matlab) function for point #4, or could this even be performed in a statistical package such as R?
Dan
On Mon, Jul 1, 2024 at 8:40 AM Reuter, Martin,Ph.D. MREUTER@mgh.harvard.edu wrote:
Hi Dan,
No, mris_preproc and also surf2surf are functions that work on surfaces. You would just import the volume statistics into Matlab and fill the corresponding columns in your matrix directly, similar to using co-variates such as eTIV.
Best, Martin
On 1. Jul 2024, at 14:03, Dan Levitas djlevitas208@gmail.com wrote:
Hi Martin,
I wanted to ping this thread again, to confirm that my proposed mass-univariate workflow using the subcortical segmented ROIs is appropriate for the LME longitudinal analysis.
Thanks again.
Dan
On Thu, Jun 27, 2024 at 5:44 AM Dan Levitas djlevitas208@gmail.com wrote:
Hi Martin,
Thanks for your reply, it's good to know that the mass-univariate approach is still appropriate here.
Since the aseg segmentations are volumetric, can I still use the *mris_preproc* and *mri_surf2surf* functions? If so, would the mri_preproc function look something like *mri_preproc --qdec-long /path/to/asegstats2table.long.table --target fsaverage --hemi rh --meas volume --out output_dir/rh.volume.mgh*
If not, and I need separate functions, would *mri_surf2surf* be replaced by *mri_vol2surf*?
On Thu, Jun 27, 2024 at 4:51 AM Reuter, Martin,Ph.D. < MREUTER@mgh.harvard.edu> wrote:
Hi Dan,
You could do either one. I would recommend the mass univariate for multiple independent test with the same design.
Best, Martin
On 26. Jun 2024, at 17:20, Dan Levitas djlevitas208@gmail.com wrote:
I recently performed an LME longitudinal mass-univariate (surface) analysis and am now trying to do something similar, but with subcortical (aseg) ROIs instead.
I first created an aseg table with the following command:
- asegstats2table \ --sd $subjects_dir \ --qdec-long
/path/to/long.qdec.table.dat \ --segno 4 10 11 12 17 18 26 43 49 50 51 53 54 58 251 253 255 \ --stats aseg.stats \ --tablefile $output_file*
The LME tutorial ( https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) provides information and examples of both a univariate and mass-univariate approach. While univariate seems most appropriate in this case, the example details design matrix setup and testing of a single ROI (hippocampus), whereas I would like to assess several subcortical ROIs.
Would I simply create a loop structure to set up design matrices for each ROI (both hemispheres) and test them individually, or can I perform a mass-univariate approach to assess all ROIs at once?
Thanks,
Dan
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer 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 Mass General Brigham Compliance HelpLine at https://www.massgeneralbrigham.org/complianceline < https://www.massgeneralbrigham.org/complianceline%3E .
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 Mass General Brigham Compliance HelpLine at https://www.massgeneralbrigham.org/complianceline .
Hi Dan,
Yes, this is correct. Usually you have a design where the ROI measurement (e.g. hippocampal volume) is your dependent variable and the independent variables are age, sex, eTIV, group etc and of course the group time interaction etc.
Best, Martin
On 1. Jul 2024, at 15:15, Dan Levitas djlevitas208@gmail.com wrote:
External Email - Use Caution
Thanks Martin,
I just want to clarify that this then is the correct workflow:
1). Import the table generated from asegstats2table (asegstats2table.long.table), and select the column pertaining to the ROIs and eTIV. 2). Add to this table any covariates, as well as the group and time columns. 3). Specify a group x time interaction contrast (e.g. CM.C = [0 0 0 0 0 0 0 0 0 0 0 0 1];). 4). Use the table (design matrix) to perform LME.
Assuming the above is correct, Is there a FreeSurfer (or Matlab) function for point #4, or could this even be performed in a statistical package such as R?
Dan
On Mon, Jul 1, 2024 at 8:40 AM Reuter, Martin,Ph.D. <MREUTER@mgh.harvard.edumailto:MREUTER@mgh.harvard.edu> wrote: Hi Dan,
No, mris_preproc and also surf2surf are functions that work on surfaces. You would just import the volume statistics into Matlab and fill the corresponding columns in your matrix directly, similar to using co-variates such as eTIV.
Best, Martin
On 1. Jul 2024, at 14:03, Dan Levitas <djlevitas208@gmail.commailto:djlevitas208@gmail.com> wrote:
Hi Martin,
I wanted to ping this thread again, to confirm that my proposed mass-univariate workflow using the subcortical segmented ROIs is appropriate for the LME longitudinal analysis.
Thanks again.
Dan
On Thu, Jun 27, 2024 at 5:44 AM Dan Levitas <djlevitas208@gmail.commailto:djlevitas208@gmail.com> wrote: Hi Martin,
Thanks for your reply, it's good to know that the mass-univariate approach is still appropriate here.
Since the aseg segmentations are volumetric, can I still use the mris_preproc and mri_surf2surf functions? If so, would the mri_preproc function look something like mri_preproc --qdec-long /path/to/asegstats2table.long.table --target fsaverage --hemi rh --meas volume --out output_dir/rh.volume.mgh
If not, and I need separate functions, would mri_surf2surf be replaced by mri_vol2surf?
On Thu, Jun 27, 2024 at 4:51 AM Reuter, Martin,Ph.D. <MREUTER@mgh.harvard.edumailto:MREUTER@mgh.harvard.edu> wrote: Hi Dan,
You could do either one. I would recommend the mass univariate for multiple independent test with the same design.
Best, Martin
On 26. Jun 2024, at 17:20, Dan Levitas <djlevitas208@gmail.commailto:djlevitas208@gmail.com> wrote:
I recently performed an LME longitudinal mass-univariate (surface) analysis and am now trying to do something similar, but with subcortical (aseg) ROIs instead.
I first created an aseg table with the following command: asegstats2table \ --sd $subjects_dir \ --qdec-long /path/to/long.qdec.table.dat \ --segno 4 10 11 12 17 18 26 43 49 50 51 53 54 58 251 253 255 \ --stats aseg.stats \ --tablefile $output_file
The LME tutorial (https://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels) provides information and examples of both a univariate and mass-univariate approach. While univariate seems most appropriate in this case, the example details design matrix setup and testing of a single ROI (hippocampus), whereas I would like to assess several subcortical ROIs.
Would I simply create a loop structure to set up design matrices for each ROI (both hemispheres) and test them individually, or can I perform a mass-univariate approach to assess all ROIs at once?
Thanks,
Dan
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto:Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
_______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edumailto: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 Mass General Brigham Compliance HelpLine at https://www.massgeneralbrigham.org/complianceline https://www.massgeneralbrigham.org/complianceline .
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 Mass General Brigham Compliance HelpLine at https://www.massgeneralbrigham.org/complianceline .
freesurfer@nmr.mgh.harvard.edu
-
Dan Levitas -
Reuter, Martin,Ph.D.