I would reprocess all the way through first and see if the remaining dura affects the surfaces, then only edit in those locations that it does

cheers Bruce

On Wed, 29 Nov 2017, Boyd, Emma wrote:

Hi Arsenije,

If you've already tested out different watershed thresholds and gcut, then yes, the next step would be to manually erase these voxels and reprocess.

The ventricle labeling is a separate problem - which version of FreeSurfer are you running?

Best,

Emma

---

Emma Boyd

Research Technician II

Laboratory for Computational Neuroimaging

Martinos Center for Biomedical Imaging

---

From: freesurfer-bounces@nmr.mgh.harvard.edu freesurfer-bounces@nmr.mgh.harvard.edu on behalf of Arsenije Subotic arsenije.subotic1@ucalgary.ca Sent: Tuesday, November 28, 2017 5:29 PM To: freesurfer@nmr.mgh.harvard.edu Subject: [Freesurfer] Erroneous output after recon-all: Bad segmentation and surfaces

Dear experts,

I recently finished processing a subject for whom I had performed a skull-st rip on the brainmask.mgz volume. I decreased the watershed parameter and re- examined the brainmask and the portion of the skull was gone, although some dura was left over. I did a recon-all using the following command as the troubleshoo ting tutorial suggested to regenerate the surfaces based on the new brainmas k.mgz volume:

recon-all -autorecon-pial -subjid <subject name>

However, after the subject finished processing the surfaces were generated/o utlined on the dura as well. I tried performing a g-cut command but it did n ot get rid of the dura. I also had issues with the subcortical segmentation as the ventricles were not completely segmented. I've attached screenshots for you to see as well. Would it be necessary to m anually erase the extra dura and then re-process the subject? Please let me know.

Thank you for your help.

Best,

Arsenije