[Freesurfer] ICV smaller than the brain

Hello,

I asked a question similar to this last year when working with FS 4.3.1, and I'm hoping that there is something new in FS 5 that can help.

I am having a problem where a lot of my subjects' brains are larger than the eICV. In FS 4.3.1, it is happening in about 70/140 subjects, and in FS 5, it is happening in about 30/140 subjects. I checked them in tkregister2, and they look ok. I can use the --old-etiv-only command, but then the problem is that my ICV shoots up to 1.7L, and my brain/ICV ratio drops to .65, way to low to be reasonable. What could the reason for this be, and are there any other tricks that I could use?

Thank you, Jeff Sadino