This is from Freesurfer wiki -  https://surfer.nmr.mgh.harvard.edu/fswiki/UserContributions/FAQ

Q. I have already skull-stripped data. Can I submit it to recon-all?

A: If your skull-stripped volume does not have the cerebellum, then no. If it does, then yes, however you will have to run the data a bit differently.

First you must run only -autorecon1 like this:
recon-all -autorecon1 -noskullstrip -s <subjid>

Then you will have to make a symbolic link or copy T1.mgz to brainmask.auto.mgz and a link from brainmask.auto.mgz to brainmask.mgz. Finally, open this brainmask.mgz file and check that it looks okay (there is no skull, cerebellum is intact; use the sample subject bert that comes with your FreeSurfer installation to make sure it looks comparable). From there you can run the final stages of recon-all:
recon-all -autrecon2 -autorecon3 -s <subjid>

Hope it helps.

Best,

Yujing

From: freesurfer-bounces@nmr.mgh.harvard.edu <freesurfer-bounces@nmr.mgh.harvard.edu> On Behalf Of James Hengenius
Sent: Thursday, December 28, 2023 5:06 PM
To: Freesurfer support list <freesurfer@nmr.mgh.harvard.edu>
Subject: [Freesurfer] Recon-all failing with pre-skullstripped input data (mri_em_register can't find brainmask.mgz)

        External Email - Use Caution        

Hello FS experts,

I downloaded some anatomical data. I have defaced full images and pre-skullstripped images for each participant. I tried the defaced images first, but FS skullstripping looks a little off (possibly because the defacing hides bits of the cranium in some images and causes problems in mri_watershed). So I would like to use the pre-skullstripped files as my initial inputs.

I have placed the pre-skullstripped file for each participant in $SUBJECTS_DIR/$subjid/mri/orig/. Each participant's skullstripped input volume is named 001.mgz.

I then submit the following command in a script to my slurm queue (after setting appropriate FS env vars), where the script input ${1} is the $subjid mentioned above:

recon-all \
    -subject ${1} \
    -all \
    -openmp 40 \
    -noskullstrip

The output log from FS ends with the following error. The function mri_em_register seems to be looking for brainmask.mgz, and fails to open it. If the input image is already extracted from the skull, do I need to manually call it brainmask.mgz rather than 001.mgz? Or am I missing something else in my function call (like order of flags)? Does -noskullstrip need to come before -openmp, for example?

Thanks!

Jim H.

End of recon-all output:

writing output to T1.mgz
3D bias adjustment took 1 minutes and 20 seconds.
@#@FSTIME  2023:12:28:16:07:04 mri_normalize N 7 e 80.39 S 0.66 U 102.70 P 128% M 581068 F 0 R 24411 W 0 c 32 w 229 I 0 O 0 L 1.00 1.38 9.22
@#@FSLOADPOST 2023:12:28:16:08:25 mri_normalize N 7 1.99 1.66 8.64
#-------------------------------------
#@# EM Registration Thu Dec 28 16:08:25 EST 2023
/gpfs/fs0/scratch/t/tpaus/jhengeni/subjects_dir_ukb/1000108_20252_2_0/mri

 mri_em_register -uns 3 -mask brainmask.mgz nu.mgz /home/t/tpaus/jhengeni/fs_versions/7.4.1/freesurfer/average/RB_all_2020-01-02.gca transforms/talairach.lta

setting unknown_nbr_spacing = 3
using MR volume brainmask.mgz to mask input volume...

== Number of threads available to mri_em_register for OpenMP = 40 ==
reading 1 input volumes...
logging results to talairach.log
reading '/home/t/tpaus/jhengeni/fs_versions/7.4.1/freesurfer/average/RB_all_2020-01-02.gca'...
GCAread took 0 minutes and 1 seconds.
average std = 7.2   using min determinant for regularization = 5.2
0 singular and 884 ill-conditioned covariance matrices regularized
reading 'nu.mgz'...
error: mghRead(/gpfs/fs0/scratch/t/tpaus/jhengeni/subjects_dir_ukb/1000108_20252_2_0/mri/brainmask.mgz, -1): could not open file^@
error: mri_em_register: could not open mask volume brainmask.mgz.
^@
@#@FSTIME  2023:12:28:16:08:25 mri_em_register N 7 e 1.42 S 0.25 U 1.43 P 118% M 446804 F 0 R 108260 W 0 c 2 w 56 I 0 O 0 L 1.99 1.66 8.64
@#@FSLOADPOST 2023:12:28:16:08:26 mri_em_register N 7 1.99 1.66 8.64
Linux nia0918.scinet.local 3.10.0-1160.102.1.el7.x86_64 #1 SMP Tue Oct 17 15:42:21 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux

recon-all -s 1000108_20252_2_0 exited with ERRORS at Thu Dec 28 16:08:26 EST 2023

For more details, see the log file /gpfs/fs0/scratch/t/tpaus/jhengeni/subjects_dir_ukb/1000108_20252_2_0/scripts/recon-all.log
To report a problem, see MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting