Hi Michael

I think the 3rd is the best. It's guaranteed to not self-intersect and be smooth, which the other two methods are not

cheers Bruce On Wed, 11 May 2016, Michael Waskom wrote:

Hi Freesurfers, I have seen at least three ways for identifying an intermediate surface between the white and pial surfaces:

• Multiply the surface normal vector for each white surface vertex by half

of the cortical thickness and add it to each vertex coordinate

• Average the white and pial coordinates for each vertex
• Use mris_expand, e.g. `mris_expand -thickness lh.smoothwm 0.5 lh.graymid`

I have found that these approaches all give similar, but not identical results.

Is there a consensus that one approach is best relative to the others?

Thanks, Michael