Hi,
I've also come across similar situations in my FS 6.0 dataset.
I have tried tweaking the seg-wlo and seg-ghi values and these steps have led to marked improvements in some cases.
Curious to hear from the FS team too.
Thanks.
Best Wishes, Elijah
Elijah Mak, Research Associate Department of Psychiatry, Old Age Psychiatry Group University of Cambridge Trinity College, CB21TQ, UK http://www.neuroscience.cam.ac.uk/directory/profile.php?fkm24