We’ve been using Brainsuite for more than a year, with great results.
Recently we started analyzing subjects that have some abnormal brain structure, and have encountered improper labeling by SVReg. Specifically, the subject I processed today has very large lateral ventricular regions. Those ventricles were labeled very poorly — big sections of the ventricles were labeled as thalamus and caudate nucleus.
Any suggestions on how to improve this SVReg behavior on brains with abnormal internal structures?
Thanks, Tod