Hi Katie,
I don't have a lot of experience in that age range but... are you talking about the wm.mgz or the ?h.white in terms of accuracy? You should use the latter, as the wm.mgz is just an initial estimate that gets refined. If it's unmyelinated white matter control points are probably not the right thing to do, unless you can find some partially myelinated regions nearby that are darker than the 110 we try to center the wm at. If you send us a dataset I can take a look. Oh, and don't worry about the inflated surface - just the accuracy of the ?h.white and ?h.pial surfs.
cheers, Bruce
On Tue, 22 Jul 2008, Katie Travis wrote:
Hello,
I have some basic questions regarding some Freesurfer reconstructions that I've made on a set of infant brains (12-15month olds). Overall, the reconstructions have been successful. The exception being that the gray matter of the anterior temporal lobes are being underestimated, typically bilaterally. For example, when I look at an inflated surface, the temporal pole(s) appears pointed and has an almost triangular edge as opposed to a more "rounded" pole. To my eye, it appears that the the gray matter is being estimated very close to the border of the white matter in the anterior temporal lobe. I believe this to be happening since the wm in this age is not well-defined and it appears darker than most other wm in the cortex. I'm also finding in a few other areas (mostly motor and some anterior/ frontal cortex) that the white matter is also being overestimated (extending into gray matter areas). For my analyses, I'm hoping to estimate most accurately the wm/gm surfaces for source analyses I am doing with MEG data. The temporal lobes are of particular importance given the auditory nature of my task. I am new to the Freesurfer editing process, and was wondering if you have any suggestions about editing these surfaces. I would also be happy to try and send you data/screen shots of what I'm describing in terms of the surface segmentations if that would better assist you with my question.
Thank you for your help,
Katie Travis
Neuroscience Doctoral Student Center for Research in Language and Multimodal Imaging Laboratory 9500 Gilman Drive, Mail Code 0562 La Jolla, CA 92093-0562 email: ktravis@ucsd.edu
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer