Hi Sherry,

that's pretty tough. The dura is very bright and quite close to the gm. You may need to erase more of the dura

Bruce

On Tue, 9 Feb 2010, Sherri Novis wrote:

Hello,

I have a series of datasets with skull stripping issues - namely left over dura/vasculature in the occipital cortex, and some in frontal areas.(Attached is a jpeg of the type of issue. It's pervasive throughout the slices, across a handful of subject data.)

I have rerun skull stripping changing the watershed value, however it's not making the distinction between dura and gray matter. (recon-all -skullstrip -wsthresh [tried 0..35] -clean-bm -no-wsgcaatlas -subjid [subject_id] - FS version: freesurfer-Linux-centos4-stable-pub-v4.3.0)

I also attempted to edit 50% of the slices by erasing voxels to make a clear boundary between the dura and cortex and then re-ran autorecon2 to see if it would at least adjust the pial surfaces - however that did not work (pial surfaces still crossed the one pixel boundary I made into the dura).

Can you recommend any other flags or things to try before I edit/erase the excess by hand?

Thanks for your time - Sherri