[Freesurfer] hyperintense artifact causing problems for segmentation

Hi,

I'm working with some poor quality images and having a lot of trouble with FreeSurfer segmentation. The main issue is that for a lot of images, there is a bright "strip" of hyperintense signal forming a ring around the cortex. I'm not sure what this is or where it comes from. I've attached an example image that shows where FreeSurfer has misinterpreted the GM/WM border due to this artefact. It will be quite labour intensive to manually remove the artefact, as it appears across most of the cortex.

Has anyone seen this before and/or could offer advice as to the best way to proceed? Using the 3T flag in version 5 helps a little but far from perfect.

Thanks,

Sarah