We're using Freesurfer version 4.5.0_64bit to delineate and calculate cortical thickness. The images we are working with are poor quality and when we run the Freesurfer program, we have problems with the segmentation of prefrontal and temporal areas. We read through previous posts on the mailing list and found that we should apply control points to normalize the white matter on about 20 subjects and possibly use these to apply bias field correction on the remainder. However, it seems that after multiple times of applying control points (about 25 control points per subject) and running the recon-all command, the problem is not corrected. The segmentation does not extend to the farthest part of the temporal poles. The command we are using is:
recon-all -autorecon-all -autorecon2-cp -autorecon3 -subjid (subj) -sd (directory) -no-isrunning
Are there any better ways of going about this? I have attached screenshots of one of the subjects in two different views.
