Freesurfer

freesurfer@nmr.mgh.harvard.edu

10 participants
30058 discussions

Re: Recon-all for multiple subjects on parallel
by Finn, Quentin 09 Jun '26

09 Jun '26

External Email - Use Caution Not a developer, but this is something that comes up frequently in most analysis situations. For qsub, you should be able to just make separate scripts for each subject and submit them all at once and let the scheduler node handle the rest. E.g. If you have sub1.sh, sub2.sh,....subn.sh as the submission scripts you can just submit them all: for j in sub*.sh; do qsub ${j} done You may consider writing a wrapper that will create all the subscripts based on an input...e.g.: SUBNAME=$1 Echo "#!/bin/bash #$ -cwd #$ -o /data/holly-host/UserName/logs/ #$ -j y #$ -S /bin/bash #$ -N reconall export FREESURFER_HOME=/usr/local/freesurfer/8.2.0-1 export FS_LICENSE=/usr/local/freesurfer/8.2.0-1/license.txt export SUBJECTS_DIR=/data/holly-host/hliu/freesurfer_subjects source $FREESURFER_HOME/SetUpFreeSurfer.sh mkdir -p $SUBJECTS_DIR env | grep -E "FREESURFER|FS_LICENSE|SUBJECTS_DIR" recon-all -all -i /data/holly-host/UserName/anat/${SUBNAME}_t1w.nii -s sub-01 SegmentAAN.sh sub-01 " > ${SUBNAME}.sh ________________________________ From: freesurfer-request(a)nmr.mgh.harvard.edu <freesurfer-request(a)nmr.mgh.harvard.edu> Sent: Tuesday, June 9, 2026 9:36 AM To: freesurfer(a)nmr.mgh.harvard.edu <freesurfer(a)nmr.mgh.harvard.edu> Subject: [EXTERNAL] Freesurfer Digest, Vol 268, Issue 8 Caution: This message came from an external sender outside your organization. Send Freesurfer mailing list submissions to freesurfer(a)nmr.mgh.harvard.edu To subscribe or unsubscribe via email, send a message with subject or body 'help' to freesurfer-request(a)nmr.mgh.harvard.edu You can reach the person managing the list at freesurfer-owner(a)nmr.mgh.harvard.edu When replying, please edit your Subject line so it is more specific than "Re: Contents of Freesurfer digest..." Today's Topics: 1. Re: Recon-all for multiple subjects on parallel (Huaiyu Liu) ---------------------------------------------------------------------- Message: 1 Date: Tue, 9 Jun 2026 14:35:49 +0000 From: Huaiyu Liu <huaiyu.liu(a)ucl.ac.uk> Subject: [Freesurfer] Re: Recon-all for multiple subjects on parallel To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu> Message-ID: <DB8PR01MB5899E8997D209D3F51B01799DB1D2(a)DB8PR01MB5899.eurp rd01.prod.exchangelabs.com> Content-Type: multipart/alternative; boundary="_000_DB8PR01MB5899E8 997D209D3F51B01799DB1D2DB8PR01MB5899eurp_" External Email - Use Caution Dear FS Developers, I recently asked a question about how to run recon-all in parallel for multiple subjects (like 4-8). Could someone give me some clues, I would be really apperciate. Best, Huaiyu ________________________________ From: Huaiyu Liu <huaiyu.liu(a)ucl.ac.uk> Sent: Thursday, May 28, 2026 18:21 To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu> Subject: Recon-all for multiple subjects on parallel Dear FreeSurfer Developers, I am using the following code on my department's HPC to (1) reconstruct a single participant’s T1w image and (2) perform brainstem segmentation. The code works well, but I have 40 participants in total. Hence, I wonder how I can run recon-all and brainstem segmentation for multiple subjects in parallel. Could you tell me the code? 1) FreeSurfer version: --freesurfer-linux 8.2.0-1 on my department's HPC (high-performance cluster) 2) My shell script for recon-all and brainstem segmentation for single subject: #!/bin/bash #$ -cwd #$ -o /data/holly-host/UserName/logs/ #$ -j y #$ -S /bin/bash #$ -N reconall export FREESURFER_HOME=/usr/local/freesurfer/8.2.0-1 export FS_LICENSE=/usr/local/freesurfer/8.2.0-1/license.txt export SUBJECTS_DIR=/data/holly-host/hliu/freesurfer_subjects source $FREESURFER_HOME/SetUpFreeSurfer.sh mkdir -p $SUBJECTS_DIR env | grep -E "FREESURFER|FS_LICENSE|SUBJECTS_DIR" recon-all -all -i /data/holly-host/UserName/anat/sub-01_t1w.nii -s sub-01 SegmentAAN.sh sub-01 3) My command for submit the jobs to the cluster with 32 GB of RAM request qsub -l vf=32G -l h_vmem=32G reconall Thank you very much for your time and consideration, I look forward to hearing from you. Best, Huaiyu _________________________________________ Huaiyu Liu (Ted) Postdoctoral Research Fellow Department of Imaging Neuroscience, Functional Imaging Laboratory University College London 12 Queen Square, London WC1N 3AR

1 0

Recon-all for multiple subjects on parallel
by Huaiyu Liu 09 Jun '26

09 Jun '26

External Email - Use Caution Dear FreeSurfer Developers, I am using the following code on my department's HPC to (1) reconstruct a single participant’s T1w image and (2) perform brainstem segmentation. The code works well, but I have 40 participants in total. Hence, I wonder how I can run recon-all and brainstem segmentation for multiple subjects in parallel. Could you tell me the code? 1) FreeSurfer version: --freesurfer-linux 8.2.0-1 on my department's HPC (high-performance cluster) 2) My shell script for recon-all and brainstem segmentation for single subject: #!/bin/bash #$ -cwd #$ -o /data/holly-host/UserName/logs/ #$ -j y #$ -S /bin/bash #$ -N reconall export FREESURFER_HOME=/usr/local/freesurfer/8.2.0-1 export FS_LICENSE=/usr/local/freesurfer/8.2.0-1/license.txt export SUBJECTS_DIR=/data/holly-host/hliu/freesurfer_subjects source $FREESURFER_HOME/SetUpFreeSurfer.sh mkdir -p $SUBJECTS_DIR env | grep -E "FREESURFER|FS_LICENSE|SUBJECTS_DIR" recon-all -all -i /data/holly-host/UserName/anat/sub-01_t1w.nii -s sub-01 SegmentAAN.sh sub-01 3) My command for submit the jobs to the cluster with 32 GB of RAM request qsub -l vf=32G -l h_vmem=32G reconall Thank you very much for your time and consideration, I look forward to hearing from you. Best, Huaiyu _________________________________________ Huaiyu Liu (Ted) Postdoctoral Research Fellow Department of Imaging Neuroscience, Functional Imaging Laboratory University College London 12 Queen Square, London WC1N 3AR

2 2

Re: Question About 8.2.0 Arm64 Mac Release Update
by fsbuild 08 Jun '26

08 Jun '26

External Email - Use Caution https://secure-web.cisco.com/16COGs9Xf9FKM8v1Ci_oDQXeV8-AvFh3_rWW0AtsOrgd4L… … lists there is an update for mri_tumorsynth included (for all OS’s).  It looks to me like the updates do not change the output but makes running it more reliable/stable.  For darwin_arm64,  the SynthSeg update is needed if you intend to run recon-all (and there is also an update to the recon-all script for all OS's). - R.

2 1

Question About 8.2.0 Arm64 Mac Release Update
by nooshin Safari 08 Jun '26

08 Jun '26

External Email - Use Caution Greeting, I saw the notice about the “Important 8.2.0 release updates for arm64 Mac machines.” I wanted to check whether this update could affect our current setup or workflow for TumorSynth and FastSurfer-LIT. Could you please confirm whether any changes are needed on our side, or whether these tools should continue to run normally after the update? Best, Nooshin

1 0

Important 8.2.0 release updates for arm64 Mac machines
by fsbuild 05 Jun '26

05 Jun '26

External Email - Use Caution If you are a Mac user with a silicon (arm64) machine with an M series chip (M1 … M4) running the 8.2.0 Freesurfer release, then please apply the latest updates by downloading the 8.2.0 patch release script and running it as described here,https://secure-web.cisco.com/17cvaN4h3NMsl7K7bR6zvjEQxNQNiLWoGZ7144oqZ… This should provide a working/updated version of SynthSeg, e..g., which is used by recon-all.  Please be sure to accept all the requests to update files. This update only applies to the 8.2.0 release and does not apply to Linux users of Mac users with older hardware using Intel processors. - R.

1 0

bringing longitudinal pipeline results back to native T1w space
by Yakupov, Renat /DZNE 04 Jun '26

04 Jun '26

External Email - Use Caution Good day all, I'd like to revisit the topic of bringing the segmentation and parcellation results of the longitudinal processing back to native T1w image space. I've used the suggested commands from this wiki https://secure-web.cisco.com/12JK6OH3yluIisG2SSs-MiTTRFfsBTwA7OHLDD0T4LKePO… with a small modification of using transformation files from <baseid>/mri/transforms directory (since they are absent in <tpid>mri/transforms directory). However, the output *-in-rawavg fies still have 256x256x256 resolution, rather than the native T1w resolution. How can I get the aseg and aparc results in native T1w space and resolution? Thank you in advance! Best regards, Renat.

1 0

Re: comparison of FS v 8.0.0 with earlier versions
by fsbuild 02 Jun '26

02 Jun '26

External Email - Use Caution Starting with 8.0.0, the recon-all pipeline was changed by default to use python code with deep learning algorithms via use of the SynthSeg, SynthStrip and SynthMorph commands.  This change should reduce the time needed to run a recon-all on a single CPU from about 8 hours down to ~2.5 hours.  This change however requires more RAM, e.g., at least 32G total system RAM; that may not be enough for recon-all running on some laptops including machines running MacOS.  Without enough memory, the OS will likely kill the recon-all process causing it to prematurely exit, and you will get some kind of system/fault error in the recon-all.log file. You can read about all the changes in the release notes page for 8.2.0, 8.1.0 and 8.0.0 (in that order from the top and then scrolling on down),https://secure-web.cisco.com/1D1wxOT0U-uFIizZOp_WPrw9ovGJ3pBOyXu3NR73… Releases can be downloaded from,https://secure-web.cisco.com/1J_8lzKD8I8vBHFQBalZC_mEPr3j-YcAz2NXavu9u… If you are running on Linux, I would start with the current 8.2.0 release as that will have the most recent bug fixes. There is currently an issue with SynthSeg on arm64 Macs, so for darwin_arm64 try the 8.1.0 release instead.   If you install 8.2.0 you should apply the latest patch updates as described under,https://secure-web.cisco.com/1o1ymBvn7L-cM9TppcvD--YG4ZaDPL4Dw4P1yr_8… I think it still works to revert your installed 8.x.x release back to the old recon-all pipeline behavior by passing the argument --no-v8  on the recon-all command line.  Keep in mind this should increase the recon-all run time back up to however long it used to take on your machine with older Freesurfer releases.  - R. On Jun 2, 2026, at 17:55, CRISTIAN MOJICA <comojica319(a)g.ucla.edu> wrote:        External Email - Use Caution        Hello FreeSurfer experts, Is anyone aware of any resources/studies that compare FreeSurfer 8.0.0 to earlier versions, or can anyone provide their experience with FreeSurfer 8.0.0 compared to earlier versions? I'm primarily curious about recon -all and its performance. Thank you in advance._______________________________________________ Freesurfer mailing list -- freesurfer(a)nmr.mgh.harvard.edu To unsubscribe send an email to freesurfer-leave(a)nmr.mgh.harvard.edu https://secure-web.cisco.com/1CQvmv0hUEWLZ1hizdo4vLfOCdBdp_sOh3Acuu_AGc-w9Q…

1 0

comparison of FS v 8.0.0 with earlier versions
by CRISTIAN MOJICA 02 Jun '26

02 Jun '26

External Email - Use Caution Hello FreeSurfer experts, Is anyone aware of any resources/studies that compare FreeSurfer 8.0.0 to earlier versions, or can anyone provide their experience with FreeSurfer 8.0.0 compared to earlier versions? I'm primarily curious about recon -all and its performance. Thank you in advance.

1 0

3 PhD positions in the BRIDGE-AI Doctoral Network
by Koen Van Leemput 02 Jun '26

02 Jun '26

External Email - Use Caution Are you passionate about pushing the limits of AI in neuroscience, and eager to make a real-world clinical impact? The BRIDGE-AI Doctoral Network is hiring ambitious PhD candidates across Europe to advance trustworthy AI for longitudinal neuroimaging, improving the way neurological diseases are monitored over time.* We are currently inviting applications for the last 3 open PhD positions, located in Germany and Finland*: * Deep collaboration between hospitals, universities & industry * Training events in leading European innovation hubs * Ample clinical neuroimaging data and expert guidance * A competitive salary while you research. Apply by 30 Jun 2026 (but applying earlier increases your chances!) Learn more & apply: https://secure-web.cisco.com/1UvMp3hoZMJb4yFegzuGdbiiNdOoG-BIuUKZp8LYxjdcxR…

1 0

Question on FreeSurfer recon-all workflow for brain tumor MRI
by nooshin Safari 01 Jun '26

01 Jun '26

External Email - Use Caution Hello, We are processing T1-weighted MRI from brain tumor patients and aim to run high-quality FreeSurfer recon-all for morphometric analysis. We are considering this workflow: TumorSynth tumor mask → VBG/virtual-brain correction → recon-all. Could you advise whether this is methodologically sound? Should tumor cases be handled with VBG/lesion filling before recon-all, by passing a mask directly, or another approach? Any recommended flags, QC steps, manual edits, example commands, or references would be appreciated. Best, Nooshin

2 2

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Freesurfer