Hello,
We're working with a pediatric data set and have found that Freesurfer's skullstripping process is inadequate for our purposes. We're interested in using FSL's bet, which we have found to do a better job of skullstripping, in combination with the other steps of Freesurfer. In a previous email, Bruce recommended that the best way to do this would be to replace the watershed step with bet. How would we go about doing that while still obtaining the brainmask.mgz necessary for the rest of the stream?
We've unsuccessfully tried using bet at different points in the stream. When using bet prior to initiating the Freesurfer stream, we've found that the Talairach transform is irreparably bad (leading to a severely altered ICV). We have also encountered difficulties when using bet on the skullstripped brainmask.mgz. So, we think that replacing Freesurfer's skullstrip with bet is the option, but we are unsure about how to proceed.
Thank you for your help, Heather