I am attempting to segment a brain with large ventricles. I was able to create a decent brainmask.mgz but the aseg segmentation after running autorecon2 and autorecon3 is neglecting a part of one hemisphere (see a picture attached). I tried to readjust the talairach parameters with tkregister without success. Running autorecon2 with "-bigventricles" flag did yield a similar result as running without it. Lastly, for your information, I use FS v5.3. I am also attaching the recon-all.log from the most recent attempt.
Can you please advise me how to get the correct segmentation of such a brain?