I have seen that the official recommendation for making a mid-gray surface is to use mris_expand. This seems to be a very slow procedure on my machine. I have also tried just loading the ?h.white and ?h.pial surfaces into matlab, averaging their vertices, and exporting the result. This seems to result in a nearly identical surface as mris_expand. On a test subject, only 48 out of 127,998 (0.0375%) vertices differed by more than 0.1mm, and these were almost entirely around the edges of the corpus callosum.
Is there a reason this alternative is a bad idea? I imagine mris_expand does additional topology checks along the way, but if the vertices vary so little, can that be a major problem?
I'm inclined to go with the faster solution, but since I'm relatively new to this procedure, I wanted to see what the community's experience might be.
Thanks! Keith
_______________________________ Keith Jamison Graduate Research Assistant
Department of Biomedical Engineering University of Minnesota 7-105 NHH, 312 Church St. SE Minneapolis, MN 55455
Office: 6-112 Nils Hasselmo Hall Mobile: 607-227-0696 kjamison@umn.edu
Hi Keith
the average white/pial surface doesn't have any topological or smoothness constraints, which is why we don't really use it. Yes, mris_expand is slow, but that's precisely because it prohibits self-intersection. The good nes is that compared to the rest of the recon stream it doesn't add that much time :)
Bruce
On Sat, 6 Apr 2013, Keith Jamison wrote:
I have seen that the official recommendation for making a mid-gray surface is to use mris_expand. This seems to be a very slow procedure on my machine. I have also tried just loading the ?h.white and ?h.pial surfaces into matlab, averaging their vertices, and exporting the result. This seems to result in a nearly identical surface as mris_expand. On a test subject, only 48 out of 127,998 (0.0375%) vertices differed by more than 0.1mm, and these were almost entirely around the edges of the corpus callosum.
Is there a reason this alternative is a bad idea? I imagine mris_expand does additional topology checks along the way, but if the vertices vary so little, can that be a major problem?
I'm inclined to go with the faster solution, but since I'm relatively new to this procedure, I wanted to see what the community's experience might be.
Thanks! Keith
Keith Jamison Graduate Research Assistant
Department of Biomedical Engineering University of Minnesota 7-105 NHH, 312 Church St. SE Minneapolis, MN 55455
Office: 6-112 Nils Hasselmo Hall Mobile: 607-227-0696 kjamison@umn.edu
Thanks for the reply. Maybe I'm not thinking deeply enough about this, but since there is a 1-to-1 mapping from white to pial vertices, is there a possibility of self-intersection in the average?
-Keith
On Sat, Apr 6, 2013 at 7:59 PM, Bruce Fischl fischl@nmr.mgh.harvard.eduwrote:
Hi Keith
the average white/pial surface doesn't have any topological or smoothness constraints, which is why we don't really use it. Yes, mris_expand is slow, but that's precisely because it prohibits self-intersection. The good nes is that compared to the rest of the recon stream it doesn't add that much time :)
Bruce
On Sat, 6 Apr 2013, Keith Jamison wrote:
I have seen that the official recommendation for making a mid-gray surface
is to use mris_expand. This seems to be a very slow procedure on my machine. I have also tried just loading the ?h.white and ?h.pial surfaces into matlab, averaging their vertices, and exporting the result. This seems to result in a nearly identical surface as mris_expand. On a test subject, only 48 out of 127,998 (0.0375%) vertices differed by more than 0.1mm, and these were almost entirely around the edges of the corpus callosum.
Is there a reason this alternative is a bad idea? I imagine mris_expand does additional topology checks along the way, but if the vertices vary so little, can that be a major problem?
I'm inclined to go with the faster solution, but since I'm relatively new to this procedure, I wanted to see what the community's experience might be.
Thanks! Keith
______________________________**_ Keith Jamison Graduate Research Assistant
Department of Biomedical Engineering University of Minnesota 7-105 NHH, 312 Church St. SE Minneapolis, MN 55455
Office: 6-112 Nils Hasselmo Hall Mobile: 607-227-0696 kjamison@umn.edu
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/**compliancelinehttp://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.
yes, I think it can happen, although it's probably pretty rare On Sat, 6 Apr 2013, Keith Jamison wrote:
Thanks for the reply. Maybe I'm not thinking deeply enough about this, but since there is a 1-to-1 mapping from white to pial vertices, is there a possibility of self-intersection in the average?
-Keith
On Sat, Apr 6, 2013 at 7:59 PM, Bruce Fischl fischl@nmr.mgh.harvard.edu wrote: Hi Keith
the average white/pial surface doesn't have any topological or smoothness constraints, which is why we don't really use it. Yes, mris_expand is slow, but that's precisely because it prohibits self-intersection. The good nes is that compared to the rest of the recon stream it doesn't add that much time :) Bruce On Sat, 6 Apr 2013, Keith Jamison wrote: I have seen that the official recommendation for making a mid-gray surface is to use mris_expand. This seems to be a very slow procedure on my machine. I have also tried just loading the ?h.white and ?h.pial surfaces into matlab, averaging their vertices, and exporting the result. This seems to result in a nearly identical surface as mris_expand. On a test subject, only 48 out of 127,998 (0.0375%) vertices differed by more than 0.1mm, and these were almost entirely around the edges of the corpus callosum. Is there a reason this alternative is a bad idea? I imagine mris_expand does additional topology checks along the way, but if the vertices vary so little, can that be a major problem? I'm inclined to go with the faster solution, but since I'm relatively new to this procedure, I wanted to see what the community's experience might be. Thanks! Keith _______________________________ Keith Jamison Graduate Research Assistant Department of Biomedical Engineering University of Minnesota 7-105 NHH, 312 Church St. SE Minneapolis, MN 55455 Office: 6-112 Nils Hasselmo Hall Mobile: 607-227-0696 kjamison@umn.eduThe 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
-
Bruce Fischl -
Keith Jamison