Dear Freesurfers, I recently heard that there are several disadvantages to using a measure of cortical surface area in Freesurfer, specifically due to conducting a vertex-wise analysis. Can someone help me understand why a measure of surface area using this approach might be flawed? Thanks!
Hi Emma,
I might not be the only one who is unsure what you mean by a vertex-wise cortical surface area measure. Do you mean something like what is illustrated in figures 2 and 3 here:
http://www.pnas.org/content/107/29/13135.figures-only
… which is similar, but not identical to the local gyrification index?
If so, would you use this measure on surfaces before or after registration to a target atlas?
Donna
On Oct 10, 2014, at 4:19 PM, Emma Thompson voneconomos@gmail.com wrote:
Dear Freesurfers, I recently heard that there are several disadvantages to using a measure of cortical surface area in Freesurfer, specifically due to conducting a vertex-wise analysis. Can someone help me understand why a measure of surface area using this approach might be flawed? Thanks!
Hi Donna,
In that same paper we also comment on the differences between expansion/contraction and absolute areal measurements assessed at each face of the surface (i.e., facewise), which can, after interpolation, be converted to vertexwise to facilitate analysis using current tools (e.g., mri_glmfit or qdec), without loss of areal quantities.
All the best,
Anderson
On 11 October 2014 04:05, Donna Dierker donna@brainvis.wustl.edu wrote:
Hi Emma,
I might not be the only one who is unsure what you mean by a vertex-wise cortical surface area measure. Do you mean something like what is illustrated in figures 2 and 3 here:
http://www.pnas.org/content/107/29/13135.figures-only
... which is similar, but not identical to the local gyrification index?
If so, would you use this measure on surfaces before or after registration to a target atlas?
Donna
On Oct 10, 2014, at 4:19 PM, Emma Thompson voneconomos@gmail.com wrote:
Dear Freesurfers, I recently heard that there are several disadvantages to using a measure
of cortical surface area in Freesurfer, specifically due to conducting a vertex-wise analysis. Can someone help me understand why a measure of surface area using this approach might be flawed? Thanks!
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 Anderson, Thank you for your paper. I guess I am confused. I have analyzed my structural data with qdec, specifically looking at a measure of cortical thickness and volume. I'm considering looking at surface area as well but was told it wasn't a very good measure using qdec, I have no other information really. Could you speak to whether or not this measure is valid in so far as one runs it using qdec?
Also, in your opinion, would it be redundant to look at surface area in addition to measures of cortical thickness and cortical volume?
Thanks for your help and apologizes for my silly questions.
On Sun, Oct 12, 2014 at 3:00 PM, Anderson M. Winkler <winkler@fmrib.ox.ac.uk
wrote:
Hi Donna,
In that same paper we also comment on the differences between expansion/contraction and absolute areal measurements assessed at each face of the surface (i.e., facewise), which can, after interpolation, be converted to vertexwise to facilitate analysis using current tools (e.g., mri_glmfit or qdec), without loss of areal quantities.
All the best,
Anderson
On 11 October 2014 04:05, Donna Dierker donna@brainvis.wustl.edu wrote:
Hi Emma,
I might not be the only one who is unsure what you mean by a vertex-wise cortical surface area measure. Do you mean something like what is illustrated in figures 2 and 3 here:
http://www.pnas.org/content/107/29/13135.figures-only
… which is similar, but not identical to the local gyrification index?
If so, would you use this measure on surfaces before or after registration to a target atlas?
Donna
On Oct 10, 2014, at 4:19 PM, Emma Thompson voneconomos@gmail.com wrote:
Dear Freesurfers, I recently heard that there are several disadvantages to using a
measure of cortical surface area in Freesurfer, specifically due to conducting a vertex-wise analysis. Can someone help me understand why a measure of surface area using this approach might be flawed? Thanks!
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 Emma,
Please, see below:
On 13 October 2014 20:11, Emma Thompson voneconomos@gmail.com wrote:
Hi Anderson, Thank you for your paper. I guess I am confused. I have analyzed my structural data with qdec, specifically looking at a measure of cortical thickness and volume. I'm considering looking at surface area as well but was told it wasn't a very good measure using qdec, I have no other information really. Could you speak to whether or not this measure is valid in so far as one runs it using qdec?
The data that is used in qdec comes from mris_preproc, which uses vertexwise interpolation with Jacobian correction for local stretches/shrinkages. The amount of area is preserved globally, which is desirable, although locally it's something difficult to achieve. If without any smoothing, the spatial correlation between mris_preproc and the method in the paper is about 0.65, and the correlation at a given vertex, across subjects, varies spatially. Note, however, that the smoothing -- that is applied by default -- should increase the correlation vastly, but I don't have the exact numbers at hand to provide yet.
For area, as well as for volume, the method shown in the paper is really the gold standard: it's pycnophylactic and thus guarantees that the amount of area is preserved locally everywhere in the brain during the interpolation. It's the same method used in GIS applications for geoprocessing. The downside is that the current implementation, using Matlab, is too slow (for the ico7, it can well take 10-15h per hemisphere in a single CPU).
Also, in your opinion, would it be redundant to look at surface area in addition to measures of cortical thickness and cortical volume?
Actually, the redundant bit here is volume. Area and thickness are uncorrelated phenotypically and genetically, and each provide different pieces of information (this is shown here http://dx.doi.org/10.1093/cercor/bhp026 and here http://www.sciencedirect.com/science/article/pii/S1053811909013160). Volume is a mixture of both, in proportions that vary across the cortical mantle. I would not consider using volume given that thickness and area are available. Moreover, volume is also an "areal quantity" in the sense that it requires the same methods for interpolation as surface area itself, so all that is said above about area applies also to cortical volumes.
Hope this helps!
All the best,
Anderson
On Sun, Oct 12, 2014 at 3:00 PM, Anderson M. Winkler <winkler@fmrib.ox.ac.uk
wrote:
Hi Donna,
In that same paper we also comment on the differences between expansion/contraction and absolute areal measurements assessed at each face of the surface (i.e., facewise), which can, after interpolation, be converted to vertexwise to facilitate analysis using current tools (e.g., mri_glmfit or qdec), without loss of areal quantities.
All the best,
Anderson
On 11 October 2014 04:05, Donna Dierker donna@brainvis.wustl.edu wrote:
Hi Emma,
I might not be the only one who is unsure what you mean by a vertex-wise cortical surface area measure. Do you mean something like what is illustrated in figures 2 and 3 here:
http://www.pnas.org/content/107/29/13135.figures-only
... which is similar, but not identical to the local gyrification index?
If so, would you use this measure on surfaces before or after registration to a target atlas?
Donna
On Oct 10, 2014, at 4:19 PM, Emma Thompson voneconomos@gmail.com wrote:
Dear Freesurfers, I recently heard that there are several disadvantages to using a
measure of cortical surface area in Freesurfer, specifically due to conducting a vertex-wise analysis. Can someone help me understand why a measure of surface area using this approach might be flawed? Thanks!
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.
This helps immensely, thank you!
On Tue, Oct 14, 2014 at 4:52 AM, Anderson M. Winkler <winkler@fmrib.ox.ac.uk
wrote:
Hi Emma,
Please, see below:
On 13 October 2014 20:11, Emma Thompson voneconomos@gmail.com wrote:
Hi Anderson, Thank you for your paper. I guess I am confused. I have analyzed my structural data with qdec, specifically looking at a measure of cortical thickness and volume. I'm considering looking at surface area as well but was told it wasn't a very good measure using qdec, I have no other information really. Could you speak to whether or not this measure is valid in so far as one runs it using qdec?
The data that is used in qdec comes from mris_preproc, which uses vertexwise interpolation with Jacobian correction for local stretches/shrinkages. The amount of area is preserved globally, which is desirable, although locally it's something difficult to achieve. If without any smoothing, the spatial correlation between mris_preproc and the method in the paper is about 0.65, and the correlation at a given vertex, across subjects, varies spatially. Note, however, that the smoothing -- that is applied by default -- should increase the correlation vastly, but I don't have the exact numbers at hand to provide yet.
For area, as well as for volume, the method shown in the paper is really the gold standard: it's pycnophylactic and thus guarantees that the amount of area is preserved locally everywhere in the brain during the interpolation. It's the same method used in GIS applications for geoprocessing. The downside is that the current implementation, using Matlab, is too slow (for the ico7, it can well take 10-15h per hemisphere in a single CPU).
Also, in your opinion, would it be redundant to look at surface area in addition to measures of cortical thickness and cortical volume?
Actually, the redundant bit here is volume. Area and thickness are uncorrelated phenotypically and genetically, and each provide different pieces of information (this is shown here http://dx.doi.org/10.1093/cercor/bhp026 and here http://www.sciencedirect.com/science/article/pii/S1053811909013160). Volume is a mixture of both, in proportions that vary across the cortical mantle. I would not consider using volume given that thickness and area are available. Moreover, volume is also an "areal quantity" in the sense that it requires the same methods for interpolation as surface area itself, so all that is said above about area applies also to cortical volumes.
Hope this helps!
All the best,
Anderson
On Sun, Oct 12, 2014 at 3:00 PM, Anderson M. Winkler < winkler@fmrib.ox.ac.uk> wrote:
Hi Donna,
In that same paper we also comment on the differences between expansion/contraction and absolute areal measurements assessed at each face of the surface (i.e., facewise), which can, after interpolation, be converted to vertexwise to facilitate analysis using current tools (e.g., mri_glmfit or qdec), without loss of areal quantities.
All the best,
Anderson
On 11 October 2014 04:05, Donna Dierker donna@brainvis.wustl.edu wrote:
Hi Emma,
I might not be the only one who is unsure what you mean by a vertex-wise cortical surface area measure. Do you mean something like what is illustrated in figures 2 and 3 here:
http://www.pnas.org/content/107/29/13135.figures-only
… which is similar, but not identical to the local gyrification index?
If so, would you use this measure on surfaces before or after registration to a target atlas?
Donna
On Oct 10, 2014, at 4:19 PM, Emma Thompson voneconomos@gmail.com wrote:
Dear Freesurfers, I recently heard that there are several disadvantages to using a
measure of cortical surface area in Freesurfer, specifically due to conducting a vertex-wise analysis. Can someone help me understand why a measure of surface area using this approach might be flawed? Thanks!
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.
If you are using --qcache, then there is nothing wrong specifically with the qdec implementation
On 10/13/2014 03:11 PM, Emma Thompson wrote:
Hi Anderson, Thank you for your paper. I guess I am confused. I have analyzed my structural data with qdec, specifically looking at a measure of cortical thickness and volume. I'm considering looking at surface area as well but was told it wasn't a very good measure using qdec, I have no other information really. Could you speak to whether or not this measure is valid in so far as one runs it using qdec?
Also, in your opinion, would it be redundant to look at surface area in addition to measures of cortical thickness and cortical volume?
Thanks for your help and apologizes for my silly questions.
On Sun, Oct 12, 2014 at 3:00 PM, Anderson M. Winkler <winkler@fmrib.ox.ac.uk mailto:winkler@fmrib.ox.ac.uk> wrote:
Hi Donna, In that same paper we also comment on the differences between expansion/contraction and absolute areal measurements assessed at each face of the surface (i.e., facewise), which can, after interpolation, be converted to vertexwise to facilitate analysis using current tools (e.g., mri_glmfit or qdec), without loss of areal quantities. All the best, Anderson On 11 October 2014 04:05, Donna Dierker <donna@brainvis.wustl.edu <mailto:donna@brainvis.wustl.edu>> wrote: Hi Emma, I might not be the only one who is unsure what you mean by a vertex-wise cortical surface area measure. Do you mean something like what is illustrated in figures 2 and 3 here: http://www.pnas.org/content/107/29/13135.figures-only … which is similar, but not identical to the local gyrification index? If so, would you use this measure on surfaces before or after registration to a target atlas? Donna On Oct 10, 2014, at 4:19 PM, Emma Thompson <voneconomos@gmail.com <mailto:voneconomos@gmail.com>> wrote: > Dear Freesurfers, > I recently heard that there are several disadvantages to using a measure of cortical surface area in Freesurfer, specifically due to conducting a vertex-wise analysis. Can someone help me understand why a measure of surface area using this approach might be flawed? Thanks! _______________________________________________ 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 https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
Hi Emma,
Could you clarify what would be the disadvantages that you've heard about? I ask because we've studied and tried to clarify many aspects of the analysis of surface area in this paper http://www.sciencedirect.com/science/article/pii/S1053811912002996. Maybe if there's something not covered there we could try to comment then.
Thanks!
All the best,
Anderson
On 10 October 2014 22:19, Emma Thompson voneconomos@gmail.com wrote:
Dear Freesurfers, I recently heard that there are several disadvantages to using a measure of cortical surface area in Freesurfer, specifically due to conducting a vertex-wise analysis. Can someone help me understand why a measure of surface area using this approach might be flawed? Thanks!
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@nmr.mgh.harvard.edu
-
Anderson M. Winkler -
Donna Dierker -
Douglas N Greve -
Emma Thompson