We have been having consistent errors, where the extreme left and extreme right of the brain are mis-segmented. Basically, it looks like the intensity normalization is not working well. I'm sending a screenshot to illustrate the problem. We added control points to fix the errors before, but this one seems like a lot of work.
Is there a solution you can offer? Such as changing parameters for segmentation etc.
FYI: image is acquired on a Skyra scanner with FA 9 degrees, TI 930, TR 2300, TE 2.39. We may be able to change the parameters in the future, but we still need to get these scans processed.