Hello
I have a dataset of patients with hydrocephalus. Some of the patients have severe ventricular enlargement which is problematic for BET. The output from a standard BET analysis extracts too much neck and cuts the top of the head. Therefore the scans were cropped by 100 voxels in the z plane, and different BET options (-f, -g, -R...) have been used to get the optimal output for each individual.
My problem arises in the more severe cases (ventricles are very large!). Random black holes appear in various locations within the extracted brain and I can't seem to solve this by changing -f or -g options, for example. I was tempted to exclude these patients from further analysis but thought I would check whether anyone has experienced something similar before?
Thanks
Katie
|