Hi everyone,
My QDEC analysis ran into an error message that seems to be a "classic" (and, so far, it seems unsolved), from mri_concat, which I've described in a separate message, just sent.
I have, however, more theoretical questions about QDEC, which I hope some kind soul will help me with:
- why does QDEC only provide the volume of *subcortical* structures under StatsDataImport/aseg.volume, when the automatic GM/WM segmentation is done for the entire brain, not just for subcortical regions?
- how are the volume measurements provided in aseg.volume different from the ones done obtained with a VBM analysis?
- why is *volume *the result of *segmentation *(aseg.volume) whereas *thickness *derives from *parcelation *(rh.aparc.thickness)? Is it not the case that both operations (segmentation+parcellation) are necessary to calculate volume as well as thickness?
- how is a QDEC thickness analysis different from a regular AN(C)OVA in which you are interested in main and interaction effects of the IVs on the DV? I ask this because some of the questions that appear in QDEC's Analysis Results tab – e.g. "is the correlation between (DV) and (IV) different from zero?" – would not (I think) be directly answerable by an ANOVA
- would the same type of ANOVA done by QDEC be doable by extracting the values of the DV for each subject (using e.g. *aseg.volume*) and then doing the ANOVA in e.g. SPSS? If so, when would you do one versus the other?
- why is it that only continuous factors (e.g. age) can be taken in the analysis as nuisance factors, when discrete variables (e.g. gender) might also be irrelevant for a particular analysis and thus belong to the Nuisance Factor list?
Many thanks in advance!!
Tudor
I was wondering if anybody is able to help with (any one of) these questions? That would be really helpful, as I could not find an answer on the wiki.
Many thanks in advance!
On 26 March 2013 20:44, Tudor Popescu tudor3@gmail.com wrote:
Hi everyone,
My QDEC analysis ran into an error message that seems to be a "classic" (and, so far, it seems unsolved), from mri_concat, which I've described in a separate message, just sent.
I have, however, more theoretical questions about QDEC, which I hope some kind soul will help me with:
- why does QDEC only provide the volume of *subcortical* structures under
StatsDataImport/aseg.volume, when the automatic GM/WM segmentation is done for the entire brain, not just for subcortical regions?
- how are the volume measurements provided in aseg.volume different from
the ones done obtained with a VBM analysis?
- why is *volume *the result of *segmentation *(aseg.volume) whereas *thickness
*derives from *parcelation *(rh.aparc.thickness)? Is it not the case that both operations (segmentation+parcellation) are necessary to calculate volume as well as thickness?
- how is a QDEC thickness analysis different from a regular AN(C)OVA in
which you are interested in main and interaction effects of the IVs on the DV? I ask this because some of the questions that appear in QDEC's Analysis Results tab – e.g. "is the correlation between (DV) and (IV) different from zero?" – would not (I think) be directly answerable by an ANOVA
- would the same type of ANOVA done by QDEC be doable by extracting the
values of the DV for each subject (using e.g. *aseg.volume*) and then doing the ANOVA in e.g. SPSS? If so, when would you do one versus the other?
- why is it that only continuous factors (e.g. age) can be taken in the
analysis as nuisance factors, when discrete variables (e.g. gender) might also be irrelevant for a particular analysis and thus belong to the Nuisance Factor list?
Many thanks in advance!!
Tudor
On 03/28/2013 03:53 PM, Tudor Popescu wrote:
I was wondering if anybody is able to help with (any one of) these questions? That would be really helpful, as I could not find an answer on the wiki.
Many thanks in advance!
On 26 March 2013 20:44, Tudor Popescu <tudor3@gmail.com mailto:tudor3@gmail.com> wrote:
Hi everyone, My QDEC analysis ran into an error message that seems to be a "classic" (and, so far, it seems unsolved), from mri_concat, which I've described in a separate message, just sent. I have, however, more theoretical questions about QDEC, which I hope some kind soul will help me with: - why does QDEC only provide the volume of /subcortical/ structures under StatsDataImport/aseg.volume, when the automatic GM/WM segmentation is done for the entire brain, not just for subcortical regions?
No reason in particular. You can add them into the qdec table.
- how are the volume measurements provided in aseg.volume different from the ones done obtained with a VBM analysis?
Does VBM give segmentation volumes? I thought it was voxel-wise exploratory.
- why is /volume /the result of /segmentation /(aseg.volume) whereas /thickness /derives from /parcelation /(rh.aparc.thickness)? Is it not the case that both operations (segmentation+parcellation) are necessary to calculate volume as well as thickness?
You can get volume from parcellations too (this is supplied in the ?h.apac.stats file). You can't get thickness from non-surface structures like amygdala.
- how is a QDEC thickness analysis different from a regular AN(C)OVA in which you are interested in main and interaction effects of the IVs on the DV? I ask this because some of the questions that appear in QDEC's Analysis Results tab – e.g. "is the correlation between (DV) and (IV) different from zero?" – would not (I think) be directly answerable by an ANOVA
Often times they are the same, depends on the contrast. I don't know what IV and DV are so I can't answer your specific question.
- would the same type of ANOVA done by QDEC be doable by extracting the values of the DV for each subject (using e.g. /aseg.volume/) and then doing the ANOVA in e.g. SPSS? If so, when would you do one versus the other?
Yes. QDEC is an exploratory voxel-based method so effects don't have to lie cleanly within an ROI boundary.
- why is it that only continuous factors (e.g. age) can be taken in the analysis as nuisance factors, when discrete variables (e.g. gender) might also be irrelevant for a particular analysis and thus belong to the Nuisance Factor list?
It is just a matter of convenience in making the contrasts. ie, it is easy to extend our contrast making code to add nuisance continuous variables doug
Many thanks in advance!! Tudor
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Thanks Doug, just to follow-up on some of these:
- why does QDEC only provide the volume of /subcortical/structures under StatsDataImport/aseg.volume, when the automatic GM/WM segmentation is done for the entire brain, not just for subcortical regions?No reason in particular. You can add them into the qdec table.
Do you mean aseg.volume also contains the volume of cortical structures as well, not just subcortical? If so, how can one retrieve those?
- how are the volume measurements provided in aseg.volumedifferent from the ones done obtained with a VBM analysis?Does VBM give segmentation volumes? I thought it was voxel-wise exploratory.
I was making a confusion here: in FSL, it's not FSL_VBM that gives you the volume of a given structure, but FSLSTATS, when used with a certain mask/ROI. My question is, then, whether the volume measurements provided in aseg.volume are similar to what you would get with FSLSTATS, with the exception that aseg.volume uses automatic segmentation rather than a cutom-defined mask
- how is a QDEC thickness analysis different from a regularAN(C)OVA in which you are interested in main and interaction effects of the IVs on the DV? I ask this because some of the questions that appear in QDEC's Analysis Results tab – e.g. "is the correlation between (DV) and (IV) different from zero?" – would not (I think) be directly answerable by an ANOVAOften times they are the same, depends on the contrast. I don't know what IV and DV are so I can't answer your specific question.
IV = independent variable; DV = dependent variable. Since in QDEC you can't define the contrast vector directly (it's created automatically, if I understand it correctly), then it seems to me that the analysis is just an AN(C)OVA, so I still don't understand how questions such as the one about the correlation being non-zero can be answered..
Thanks again very much Doug!!
On 28 March 2013 20:07, Douglas N Greve greve@nmr.mgh.harvard.edu wrote:
On 03/28/2013 03:53 PM, Tudor Popescu wrote:
I was wondering if anybody is able to help with (any one of) these questions? That would be really helpful, as I could not find an answer on the wiki.
Many thanks in advance!
On 26 March 2013 20:44, Tudor Popescu <tudor3@gmail.com mailto:tudor3@gmail.com> wrote:
Hi everyone, My QDEC analysis ran into an error message that seems to be a "classic" (and, so far, it seems unsolved), from mri_concat, which I've described in a separate message, just sent. I have, however, more theoretical questions about QDEC, which I hope some kind soul will help me with: - why does QDEC only provide the volume of /subcortical/ structures under StatsDataImport/aseg.volume, when the automatic GM/WM segmentation is done for the entire brain, not just for subcortical regions?No reason in particular. You can add them into the qdec table.
- how are the volume measurements provided in aseg.volume different from the ones done obtained with a VBM analysis?Does VBM give segmentation volumes? I thought it was voxel-wise exploratory.
- why is /volume /the result of /segmentation /(aseg.volume) whereas /thickness /derives from /parcelation /(rh.aparc.thickness)? Is it not the case that both operations (segmentation+parcellation) are necessary to calculate volume as well as thickness?You can get volume from parcellations too (this is supplied in the ?h.apac.stats file). You can't get thickness from non-surface structures like amygdala.
- how is a QDEC thickness analysis different from a regular AN(C)OVA in which you are interested in main and interaction effects of the IVs on the DV? I ask this because some of the questions that appear in QDEC's Analysis Results tab – e.g. "is the correlation between (DV) and (IV) different from zero?" – would not (I think) be directly answerable by an ANOVAOften times they are the same, depends on the contrast. I don't know what IV and DV are so I can't answer your specific question.
- would the same type of ANOVA done by QDEC be doable by extracting the values of the DV for each subject (using e.g. /aseg.volume/) and then doing the ANOVA in e.g. SPSS? If so, when would you do one versus the other?Yes. QDEC is an exploratory voxel-based method so effects don't have to lie cleanly within an ROI boundary.
- why is it that only continuous factors (e.g. age) can be taken in the analysis as nuisance factors, when discrete variables (e.g. gender) might also be irrelevant for a particular analysis and thus belong to the Nuisance Factor list?It is just a matter of convenience in making the contrasts. ie, it is easy to extend our contrast making code to add nuisance continuous variables doug
Many thanks in advance!! Tudor
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422
Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: www.nmr.mgh.harvard.edu/facility/filedrop/index.html Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
On 03/29/2013 04:34 PM, Tudor Popescu wrote:
Thanks Doug, just to follow-up on some of these:
> - why does QDEC only provide the volume of /subcortical/ > structures under StatsDataImport/aseg.volume, when the automatic > GM/WM segmentation is done for the entire brain, not just for > subcortical regions? > No reason in particular. You can add them into the qdec table.Do you mean aseg.volume also contains the volume of cortical structures as well, not just subcortical? If so, how can one retrieve those?
No, I mean you can add them into the qdec.dat file that you load into qdec.
> - how are the volume measurements provided in aseg.volume > different from the ones done obtained with a VBM analysis? > Does VBM give segmentation volumes? I thought it was voxel-wise exploratory.I was making a confusion here: in FSL, it's not FSL_VBM that gives you the volume of a given structure, but FSLSTATS, when used with a certain mask/ROI. My question is, then, whether the volume measurements provided in aseg.volume are similar to what you would get with FSLSTATS, with the exception that aseg.volume uses automatic segmentation rather than a cutom-defined mask
I don't know how FSLSTATS works. Ours counts up the voxels but also uses partial volume correction.
> - how is a QDEC thickness analysis different from a regular > AN(C)OVA in which you are interested in main and interaction > effects of the IVs on the DV? I ask this because some of the > questions that appear in QDEC's Analysis Results tab – e.g. "is > the correlation between (DV) and (IV) different from zero?" – > would not (I think) be directly answerable by an ANOVA > Often times they are the same, depends on the contrast. I don't know what IV and DV are so I can't answer your specific question.IV = independent variable; DV = dependent variable. Since in QDEC you can't define the contrast vector directly (it's created automatically, if I understand it correctly), then it seems to me that the analysis is just an AN(C)OVA, so I still don't understand how questions such as the one about the correlation being non-zero can be answered..
The correlation just tests whether the slope of a DV vs IV is equal to 0. I don't know whether this fits into an ANCOVA framework or not. doug
Thanks again very much Doug!!
On 28 March 2013 20:07, Douglas N Greve <greve@nmr.mgh.harvard.edu mailto:greve@nmr.mgh.harvard.edu> wrote:
On 03/28/2013 03:53 PM, Tudor Popescu wrote: > I was wondering if anybody is able to help with (any one of) these > questions? That would be really helpful, as I could not find an answer > on the wiki. > > Many thanks in advance! > > On 26 March 2013 20:44, Tudor Popescu <tudor3@gmail.com <mailto:tudor3@gmail.com> > <mailto:tudor3@gmail.com <mailto:tudor3@gmail.com>>> wrote: > > Hi everyone, > > > My QDEC analysis ran into an error message that seems to be a > "classic" (and, so far, it seems unsolved), from mri_concat, which > I've described in a separate message, just sent. > > > I have, however, more theoretical questions about QDEC, which I > hope some kind soul will help me with: > > > - why does QDEC only provide the volume of /subcortical/ > structures under StatsDataImport/aseg.volume, when the automatic > GM/WM segmentation is done for the entire brain, not just for > subcortical regions? > No reason in particular. You can add them into the qdec table. > > > - how are the volume measurements provided in aseg.volume > different from the ones done obtained with a VBM analysis? > Does VBM give segmentation volumes? I thought it was voxel-wise exploratory. > > - why is /volume /the result of /segmentation /(aseg.volume) > whereas /thickness /derives from /parcelation > /(rh.aparc.thickness)? Is it not the case that both operations > (segmentation+parcellation) are necessary to calculate volume as > well as thickness? > You can get volume from parcellations too (this is supplied in the ?h.apac.stats file). You can't get thickness from non-surface structures like amygdala. > > > - how is a QDEC thickness analysis different from a regular > AN(C)OVA in which you are interested in main and interaction > effects of the IVs on the DV? I ask this because some of the > questions that appear in QDEC's Analysis Results tab – e.g. "is > the correlation between (DV) and (IV) different from zero?" – > would not (I think) be directly answerable by an ANOVA > Often times they are the same, depends on the contrast. I don't know what IV and DV are so I can't answer your specific question. > > - would the same type of ANOVA done by QDEC be doable by > extracting the values of the DV for each subject (using e.g. > /aseg.volume/) and then doing the ANOVA in e.g. SPSS? If so, when > would you do one versus the other? > Yes. QDEC is an exploratory voxel-based method so effects don't have to lie cleanly within an ROI boundary. > > - why is it that only continuous factors (e.g. age) can be taken > in the analysis as nuisance factors, when discrete variables (e.g. > gender) might also be irrelevant for a particular analysis and > thus belong to the Nuisance Factor list? > It is just a matter of convenience in making the contrasts. ie, it is easy to extend our contrast making code to add nuisance continuous variables doug > > Many thanks in advance!! > > Tudor > > > > > _______________________________________________ > Freesurfer mailing list > Freesurfer@nmr.mgh.harvard.edu <mailto:Freesurfer@nmr.mgh.harvard.edu> > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Douglas N. Greve, Ph.D. MGH-NMR Center greve@nmr.mgh.harvard.edu <mailto:greve@nmr.mgh.harvard.edu> Phone Number: 617-724-2358 Fax: 617-726-7422 Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting <http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting> FileDrop: www.nmr.mgh.harvard.edu/facility/filedrop/index.html <http://www.nmr.mgh.harvard.edu/facility/filedrop/index.html> Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu <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
freesurfer@nmr.mgh.harvard.edu
-
Douglas N Greve -
Tudor Popescu