I have run recon-all on two timepoints of a subject (MPS1011-4 and MPS1011-5). I added control points to MPS1011-4 after the initial recon-all processing because large areas of white matter were excluded. Unfortunately, ~50% of the control points I added were apparently ignored, as they did not cause the missed wm areas to be segmented (see attached images). 

I have run into this problem in a few other subjects as well, where large wm areas are missed even after adding control points and running recon-all -autorecon2-cp -autorecon3

The other timepoint exited with errors, leaving the brain unsegmented. 

Any advice on how to accurately analyze these timepoints would be greatly appreciated. Their recon-all logs are also attached.

Thank you!

Victor