Hi Gurus,

I know that residual dura in skull stripped anatomical datasets are not usually considered an issue, but I've been noticing that nonlinear registration methods are a lot less forgiving of such cruft because they have enough degrees of freedom to try to align bits of dura with the gray matter in a template without disturbing the gross alignment of the rest of the brain, as would occur in a linear registration. Do you have any tips for making 3dSkullStrip more aggressive? In the two T1 images that are giving me trouble in this case, the dura is slightly brighter than the adjacent gray matter.

Thanks,
Isaac