Freesurfer

freesurfer@nmr.mgh.harvard.edu

17 participants
30128 discussions

Re: SynthSR - White Block Issue in the Processed T1 image
by fsbuild 13 Jan '26

13 Jan '26

External Email - Use Caution I’m wondering if this is a graphics issue causing the screen not to update properly, related to the fact that you are using T1 images.  You could try one of the subjects that displays the erroneous output on a different system and see if you get the same result.  Maybe you could let us know what the system is you are running on including the OS and what your graphics hardware is.  - R. On Jan 12, 2026, at 00:57, Tejendra Singh <tejendra.singh.cp(a)gmail.com> wrote:        External Email - Use Caution        This is the issueOn Mon, Jan 12, 2026 at 11:25 AM Tejendra Singh <tejendra.singh.cp(a)gmail.com> wrote: Hello everyone, I am using SynthSR (FreeSurfer 8.1.0, mri_synthsr) to standardize T1-weighted MRI scans across a multi-protocol clinical dataset. The majority of subjects process correctly; however, for a small subset of scans, the SynthSR output contains prominent white rectangular/block artefacts over the brain volume.Key observations: The artefacts are not present in the original T1 imagesThey appear only in the SynthSR outputThe affected original scans are already 1 × 1 × 1 mm isotropicOriginal images otherwise appear visually normal (no obvious truncation or corruption)Artefacts look axis-aligned (rectangular slabs), suggesting a reconstruction failure rather than anatomyExample workflow:mri_synthsr \   --i sub-XX_T1w.nii \   --o sub-XX_synthsr.nii.gzQuestions:Are these white block artefacts a known failure mode of SynthSR for certain inputs?Are there recommended preprocessing checks (e.g., header/orientation fixes) to reduce this issue?Would trying the --v1 model be advisable in such cases?If the original scan is already clean and isotropic, is it methodologically acceptable to exclude SynthSR output and retain the original image for downstream processing?I am currently treating SynthSR as an exploratory preprocessing step and excluding affected subjects from the SynthSR-based pipeline, but I would appreciate guidance on best practices.Thank you for any insights <Screenshot 2026-01-12 at 11.26.45 AM.png>_______________________________________________Freesurfer mailing list -- freesurfer@nmr..mgh.harvard.eduTo unsubscribe send an email to freesurfer-leave@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman3/lists/freesurfer@nmr.mgh.harvard.edu/

2 1

Synthstrip plus manual edits
by Zeng, Victor (BIDMC - Keshavan - Psychiatry) 12 Jan '26

12 Jan '26

External Email - Use Caution Hi, I enjoy the output of synthstrip, thank you all for developing it. I am asking 2 questions to better understand how it interacts with FreeSurfer. 1. If you run the output of synthstrip on a T1.mgz in recon1 to create a better brainmask.mgz, there is no effects of synthstrip further downstream, because the brainmask is in of itself a mask only, and the nu.mgz, which was untouched by synthstrip is used to create the brain.mgz and etc. So the intensity normalization that is done within synthstrip does not determine anything downstream. Using the brainmask.mgz for visual quality check after recon-all finishes only creates an illusion that this much better branimask was used to create the aparc+aseg.mgz/pial/white lines etc. So for example, say I have a nu.mgz/T1.mgz that has a motion artifact and synthstrip fixes it. Yet recon2/3 is running the aparc+aseg.mgz on the unfixed nu.mgz/T1.mgz. Likewise, hippocampal subfield is run on the unfixed nu.mgz/T1.mgz If I wanted to apply the intensity normalization procedures that synthstrip.mgz uses, I would use it to overwrite the T1.mgz. However it is unwise to have a skull-stripped image such as the output of synthstrip be put in that stage. Is this all correct 2) Considering that I am correct for q1, if you are manual editing the brain, using the post-synthstrip brainmask.mgz and T1.mgz as the reference, you are in fact drawing back pre-synthstrip voxels into the brain. But it doesn't matter you're mixing pre or post-synthstrip voxels because brainmask.mgz is only used as a binary mask. Is this correct? Victor Zeng Beth Israel Deaconess Medical Center Vial of Victory LLC -- ---------------------------------------------------------------------- PLEASE NOTE: This message is intended for the use of the person to whom it is addressed. It may contain information that is privileged, confidential and exempt from disclosure under applicable law. If you are not the intended recipient, your use of this message for any purpose is strictly prohibited. If you have received this communication in error, please delete the message and notify the sender so that we may correct our records. See our web page at http://secure-web.cisco.com/15nnoIgymyRRQJfA8M1EZ7dQcjsn0pbFPCgUE1G5mQBzPoZ… for a full directory of Beth Israel Lahey Health sites, staff, services and career opportunities.

1 0

SynthSR - White Block Issue in the Processed T1 image
by Tejendra Singh 12 Jan '26

12 Jan '26

External Email - Use Caution Hello everyone, I am using SynthSR (FreeSurfer 8.1.0, mri_synthsr) to standardize T1-weighted MRI scans across a multi-protocol clinical dataset. The majority of subjects process correctly; however, for a small subset of scans, the SynthSR output contains prominent white rectangular/block artefacts over the brain volume. Key observations: - The artefacts are not present in the original T1 images - They appear only in the SynthSR output - The affected original scans are already 1 × 1 × 1 mm isotropic - Original images otherwise appear visually normal (no obvious truncation or corruption) - Artefacts look axis-aligned (rectangular slabs), suggesting a reconstruction failure rather than anatomy Example workflow: mri_synthsr \ --i sub-XX_T1w.nii \ --o sub-XX_synthsr.nii.gz *Questions:* 1. Are these white block artefacts a known failure mode of SynthSR for certain inputs? 2. Are there recommended preprocessing checks (e.g., header/orientation fixes) to reduce this issue? 3. Would trying the --v1 model be advisable in such cases? 4. If the original scan is already clean and isotropic, is it methodologically acceptable to exclude SynthSR output and retain the original image for downstream processing? I am currently treating SynthSR as an exploratory preprocessing step and excluding affected subjects from the SynthSR-based pipeline, but I would appreciate guidance on best practices. Thank you for any insights

2 2

FreeSurfer license.txt request
by Rohit Kumar 12 Jan '26

12 Jan '26

External Email - Use Caution Dear FreeSurfer Team, I am a member of the FreeSurfer mailing list and would like to request a license.txt file for academic research use. Name: Rohit Kumar Email: rkumar_me24(a)thapar.edu Institution: Thapar Institute of Engineering and Technology Operating System: Linux Thank you

2 1

2 Tenure-track positions at UTRGV
by Anderson Winkler 08 Jan '26

08 Jan '26

External Email - Use Caution Dear all, We have two tenure track positions open in the areas of computational genomics/statistical genetics and neuroscience. These are in the Department of Human Genetics at UTRGV: * FRGV1939 - Assistant Professor in Bioinformatics and Computational Genomics - Human Genetics<https://secure-web.cisco.com/1_921e5oeJ5FkCUsz-J7j2b8_nA-ws0xQmEmh3YRvlQfet…> - https://secure-web.cisco.com/1_921e5oeJ5FkCUsz-J7j2b8_nA-ws0xQmEmh3YRvlQfet… * FRGV1940 - Assistant Professor in Statistical Genetics and Neuroscience - Human Genetics<https://secure-web.cisco.com/13Bs3h7UnYDJX9s5iA_KiU2_6MdYN5kHdUvXrApPsFey-m…> - https://secure-web.cisco.com/13Bs3h7UnYDJX9s5iA_KiU2_6MdYN5kHdUvXrApPsFey-m… Anyone interested and with the qualifications is encouraged to apply. All the best, Anderson [Logo Description automatically generated] Anderson M. Winkler, MD, DPhil Associate Professor Department of Human Genetics 1 West University Boulevard - BCHGB 1.110 Brownsville, Texas 78520 +1 (956) 882-7563 • anderson.winkler(a)utrgv.edu<mailto:anderson.winkler@utrgv.edu> utrgv.edu<https://secure-web.cisco.com/1VT5YG_SDSlt11vCklJbDj_aWIsF5WRGLQJZtWOvtLEFtO…> Brownsville • Edinburg • Harlingen

1 0

FS Clinical 8.1.0 question
by Sudhakar, Tejaswi 07 Jan '26

07 Jan '26

External Email - Use Caution Hi FreeSurfer Team, I am trying to run the clinical pipeline using the 8.1.0 version. I was able to successfully run the pipeline using the 7.4 version using the approach (recon-all-clinical.sh /path/to/T1 [subject id] [threads] /path/to/SUBJECTS_DIR, without providing the input, subject, threads and sdir flags) but am getting the following errors when running it on the newest release. I have set the subjects directory as an environmental variable on attempted terminal test runs. recon-all-clinical.sh ~T1//T1.nii.gz 1 6 ${SUBJECTS_DIR} Unknown option: ~/T1/T1.nii.gz Use -h or --help for usage information. recon-all-clinical.sh ~T1/T1.nii.gz 1 6 ${SUBJECTS_DIR} then: then/endif not found. recon-all-clinical.sh -i ~/T1/T1.nii.gz -subjid 1 -threads 6 -sdir ~/fullcohort then: then/endif not found. recon-all-clinical.sh -i ~/T1/T1.nii.gz -subjid 1 -threads 6 SUBJECTS_DIR variable is empty. Please provide a valid directory using the -sdir flag or set it as an environment variable. Can you recommend any appropriate changes to my call to run the clinical pipeline? I can provide my full script if necessary.

2 1

Registration Open: IST Winter School 2026 (Hands-on Tractography) – Verona, Italy
by Graham Little 07 Jan '26

07 Jan '26

External Email - Use Caution Dear colleagues, We are pleased to announce that registration is now open for the IST Winter School 2026, organized by the International Society for Tractography. Dates: February 24–27, 2026 Location: University of Verona, Italy Website & Registration: https://secure-web.cisco.com/1o-6EAvBjNuHa_ukaP0TS0Kq3Gqh09WzXhvE5uFE1Wjt_e… This comprehensive, hands-on workshop is designed to take participants through the complete tractography workflow—from diffusion MRI preprocessing and model fitting to streamline reconstruction and quantitative analysis. Highlights: - Expert Instructors: Alessandro Daducci, Francois Rheault, Chiara Maffei, Alexander Leemans, Pamela Guevara, Elena Borra, and Alessandra Griffa - Curriculum: Covers fundamentals to advanced applications suitable for users of FSL, MRtrix, and other major software packages. - Format: Bootcamp-style with theoretical talks and guided practical sessions. Registration Deadline: January 31, 2026 Space is limited. We encourage early registration to secure your spot. Best regards, The IST Winter School Organizing Committee

1 0

Clarification on running -autorecon2-wm, -autorecon-pial, and -autorecon3 after manual white matter and pial edits
by Camilo Castelblanco Riveros 05 Jan '26

05 Jan '26

External Email - Use Caution Hello FreeSurfer Developers and Community, I am currently refining surface reconstructions for my subjects by performing manual edits on both the white matter segmentation and pial surfaces using FreeSurfer. According to the FreeSurfer documentation, after making white matter edits, the recommended command is: recon-all -autorecon2-wm -subjid <subject> and after pial surface edits, it is recommended to run: recon-all -autorecon-pial -subjid <subject> followed by: recon-all -autorecon3 -subjid <subject> My main goal is to ensure that both white matter and pial manual edits are correctly incorporated without losing either. However, I have seen in the FreeSurfer tutorial the combined command: recon-all -autorecon2-wm -autorecon3 -subjid <subject> which runs both stages sequentially. I have two related questions: 1. If I run `recon-all -autorecon2-wm -autorecon3` after both white matter and pial edits, will my manual pial edits be preserved, or will they be overwritten by the `-autorecon2-wm` step? 2. Is the recommended workflow to run `-autorecon2-wm` first, then `-autorecon-pial` as a separate command, and finally `-autorecon3`? Or can some of these stages be combined safely in a single command line? For reference, I am using: - FreeSurfer version: 7.4.1 - Platform: Red Hat Enterprise Linux With gratitude, Camilo =================================== *Camilo Castelblanco Riveros (He/His)* Neuroscience Ph.D. Candidate (G4) Robertson Lab <https://secure-web.cisco.com/1VGQeQdcjwLtBi3C9UBo55c7SqVai-UrjQm1lF9QaAVlL0…> & Bujarski Lab <https://secure-web.cisco.com/1y_9Zgr30ubJ5JlJFc0I6-yJcgPm8sSkph5TvzTpqhl6nv…> Student Interest Group in Neurology (SIGN <https://secure-web.cisco.com/1uGgfMp5acPV-NG-koj4ZIyk0_wqPKAvcoF_BCuwGLKcw_…>) Co-Leader Global Health Fellow e: ccastelblanco.gr(a)dartmouth.edu Zoom <https://secure-web.cisco.com/1i_IGOhLxqGYKgaY8rs_7Ga3801HHvx4C0j37X5CbmatR6…>

2 1

dev builds for Ubuntu 24 and Rocky 9 Linux (and caveat)
by fsbuild 02 Jan '26

02 Jan '26

External Email - Use Caution Please note *development* (post 8.1.0) builds are available for Rocky 9 and Ubuntu 24 Linux (along with the same build for other platforms) under,https://secure-web.cisco.com/18lDFyiOZI7F--fMpU0ilANw1HPsulu5D_fzx2JS… Development builds are not releases and so they are potentially unstable and not fully tested to the point where they are suitable for long term work.   We would be interested in feedback from running known to be good recon-all, etc. command line sequences using the Rocky 9 and Ubuntu 24 builds on their respective OS's. - R.Caveat from the Freesurfer release page, Important Note: When processing a group of subjects for your study, it is essential to process all your subjects with the same version of FreeSurfer, on the same OS platform and vendor, and for safety, even the same version of the OS. While we continue to work to ensure that results match across platforms, there are none-the-less system-level libraries that are OS dependent. An exception to this rule is that you may view and edit files across any platform or version, and run some post-processing tools (outside the recon-all stream) if you check with us first (for instance you may run the longitudinal processing with newer versions).

1 0

Re: Freesurfer Digest, Vol 262, Issue 31
by Jerome Maller 30 Dec '25

30 Dec '25

External Email - Use Caution Thanks Matt. However I am wondering why the -expert flag is not working? Regards, Jerome On Tue, 30 Dec 2025 at 11:23, <freesurfer-request(a)nmr.mgh.harvard.edu> wrote: > 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. Problem using my own mask in recon-all (Jerome Maller) > 2. Re: Problem using my own mask in recon-all (Glasser, Matthew) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Tue, 30 Dec 2025 08:45:24 -0600 > From: Jerome Maller <jerome.maller(a)monash.edu> > Subject: [Freesurfer] Problem using my own mask in recon-all > To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu> > Message-ID: > <CAMsE-JKS3KGXtTFGqGfU0ysnCwn-C8= > ngZ5uf+3v-h-dZ4SHpA(a)mail.gmail.com> > Content-Type: multipart/alternative; > boundary="000000000000a7a85006472c69f9" > > External Email - Use Caution > > Hi Freesurfer experts. > > > > I am trying to process a T1-MPRAGE 7T scan with 0.5mm isotropic resolution > with the recon-all -highres flag. FS is struggling to make a good mask even > though I’ve tried various combinations in the recon-all command (e.g. > different -skullstrip -wsthresh numbers, -wsmore, different -watershed > atlases). I have tried with and without neck removal (using FSL’s > robustfov) as well as bias field correction (via SPM and also FSL’s FAST), > and feeding in with and without the scan already brain extracted (via > robex). I am now trying to feed in my pre-made mask using the -expert flag. > > > > * I have named my pre-made mask as brainmask.mgz and placed it in the > directory where my t1.nii.gz is located > > * I have made a text file (called expert.opts) that contains: > > > mri_mask brainmask.mgz > > > > My recon-all command calls the file using -expert expert.opts > > > > However, recon-all seems to ignore the calling of the expert.opts and so > tries to make its own mask… > > > > So my question is: am I correctly using the =expert flag and expert.opts to > use my pre-made brainmask.mgz? > > > > PS. I am using FS version 7.1 (as I am running Ubuntu 18.04 and can’t > upgrade due to lots of other software running on the computer) > > > > Thankyou! > > > > Jerome > -- >

1 0

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