yes, that is correct. An alternative would be editing the aseg directly.
Editing the wm will help the surfaces but not correct the aseg.

But as Antonin suggests, try running with -bigventricles and see if that
fixes your problems

cheers
Bruce


On Sun, 23 Apr 2017, Shane Schofield wrote:

> Thank you. I will try that. Do I run that command after editing the wm.mgz
> to fill in the ventricles? I noticed that the voxels (value = 255) are still
> 255 in the new WM.mgz after autorecon2-wm. Is this correct?
>
>
> On Sunday, April 23, 2017 1:33 PM, Antonin Skoch <ansk@ikem.cz> wrote:
>
>
> Dear Shane,
>
> I would suggest to run:
>
> recon-all -bigventricles -s subject -autorecon2 -autorecon3 (-T2pial or
> -FLAIRpial if you have 3D T2 or 3D FLAIR). This should help to correct aseg
> in case of such large ventricles.
>
> Antonin Skoch
>
>
> From: Shane Schofield <shane.schofield@yahoo.com>
> To: Freesurfer Support List <freesurfer@nmr.mgh.harvard.edu>, Antonin Skoch
> <ansk@ikem.cz>
> Sent: 4/22/2017 11:30 PM
> Subject: Filling in missing ventricles in WM
>
>      Hi Freesurfer Experts and Antonin,
>
> About an earlier question on XL Defects
>
> I have used your suggestions to check the orig.nofix surfaces and I
> think the skull and missing ventricle voxels are causing the problem.
> Picture pasted.
>
> Inline image
> I have  removed the WM voxels in the skull, filled in the ventricles
> by adding WM voxels (brush value = 255 ), and ran the command
> recon-all -s subject -autorecon2-wm -autorecon3
> Recon-all was completed successfully and the surfaces look good.
> However, the aseg volume in the lateral ventricles are still wrong.
> Inline image