I would recommend registering them with 6 parameters (FLIRT -dof 6) to one
of the MNI152 templates with skull, and that should reduce some of the
variability in outcomes with BET, once you have tweaked the settings to get
good extractions with your data.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Jennie Lambert
Sent: Tuesday, May 17, 2011 6:20 PM
To: [log in to unmask]
Subject: [FSL] BET With A Large Data Sample
Hello,
We are beginning to embark on a project that will involve applying the Brain
Extraction Tool to about 900 brains. These brains have a substantial amount
neurodegenerative disease, which typically makes BETing them cumbersome.
Could you please suggest the fastest approach in getting an accurate BET
image in the most timely manner? The BET images will be used for gray matter
analyses.
Thank you very much,
~ Jennie
|