Freesurfer

freesurfer@nmr.mgh.harvard.edu

6 participants
30054 discussions

Re: NextBrain atlas: Missing hippocampal subfields in FreeSurfer 8.1
by Iglesias Gonzalez, Juan E. 19 Jan '26

19 Jan '26

Thank for the comprehensive feedback, Luis! * ROIs: A lot of small structures were merged with others when building (and then further simplifying) the atlas, including (pro-)subicular labels. Sorry for the inconvenience. Even with histology, it can be tricky to reliably label some of these. I’m CCing our co-author Martina in case she can shed further light on this. * Thanks for detecting the bug. Trying to keep all the versions has led to rather messy code. We are in the process of reorganizing / refactoring the code and hope to have a clean version out soon (2-3 weeks). * The error with the missing atlas should be solved by now in the development version; apologies for the inconvenience. Again: thanks for you feedback and for your patience with our new software! /E -- Juan Eugenio Iglesias http://www.jeiglesias.com From: Luis Alberto Barreiro Araújo <pg54004(a)alunos.uminho.pt> Date: Saturday, January 17, 2026 at 14:11 To: Iglesias Gonzalez, Juan E. <JIGLESIASGONZALEZ(a)mgh.harvard.edu> Cc: freesurfer(a)nmr.mgh.harvard.edu <freesurfer(a)nmr.mgh.harvard.edu>, Tiago Gil Rodrigues Oliveira <tiago(a)med.uminho.pt>, Francisco Almeida <franciscojcga(a)gmail.com> Subject: NextBrain atlas: Missing hippocampal subfields in FreeSurfer 8.1 External Email - Use Caution Dear Dr. Iglesias, I'm a researcher at the University of Minho - ICSV Institute, conducting research on hippocampal patterns in Alzheimer's disease. I've used NextBrain for detailed hippocampal and full segmentation on NACC dataset subjects. Issue: I noticed your NextBrain demo website shows hippocampal structures including: prosubiculum (uncal, rostral, caudal) presubiculum (rostral, caudal) indusium_griseum supracallosal_subiculum However, these structures don't appear in either atlas_simplified or atlas_full YAML configuration files in FreeSurfer 8.1. I've verified this by checking: data_simplified/atlas_names_and_labels.yaml data_full/atlas_names_and_labels.yaml Additionally, when attempting to use atlas_full, the segmentation crashes with "IndexError: list index out of range" at label 266, suggesting the data_full/combined_atlas_labels_1mm.yaml config is incomplete (maps only 265 of 334 labels). Questions: Are these additional hippocampal subfields available in FreeSurfer dev or a more recent atlas version? If not yet released, are they planned for a future update? Would you recommend an alternative approach for segmenting these specific structures? Should I post these issues in the official FreeSurfer GitHub repository for broader advice ? Note: I'm running NextBrain via FreeSurfer 8.1 Docker container, due to incompatibility issues in our workstation with the OS. The Docker image was missing both the atlas files and the Python segmentation scripts, so I had to manually download: * atlas_simplified.zip and atlas_full.zip from the FTP server * ERC_bayesian_segmentation scripts from the GitHub dev branch * SuperSynth_August_2025.pth model (1GB) After mounting these files into the container, atlas_simplified works perfectly. However, atlas_full crashes as described above, and neither atlas contains the additional hippocampal subfields visible in your demo. From my understanding, recently (in the past week) a new version with the bayesian segmentation folder was put available but the issues I have remain. Another small note: I had to manually correct the mri_histo_atlas_segment_fireants.sh script because some of the parameters for the python scripts where missing. I'm not sure if that's supposed to happen or if it's a version error. Thank you for your attention. I look forward to receiving a response. Best regards, Luís Researcher, Informatics Engineering University of Minho - ICVS

1 0

NextBrain atlas: Missing hippocampal subfields in FreeSurfer 8.1
by Luis Alberto Barreiro Araújo 17 Jan '26

17 Jan '26

External Email - Use Caution Dear Dr. Iglesias, I'm a researcher at the University of Minho - ICSV Institute, conducting research on hippocampal patterns in Alzheimer's disease. I've used NextBrain for detailed hippocampal and full segmentation on NACC dataset subjects. Issue: I noticed your NextBrain demo website shows hippocampal structures including: prosubiculum (uncal, rostral, caudal) presubiculum (rostral, caudal) indusium_griseum supracallosal_subiculum However, these structures don't appear in either atlas_simplified or atlas_full YAML configuration files in FreeSurfer 8.1. I've verified this by checking: data_simplified/atlas_names_and_labels.yaml data_full/atlas_names_and_labels.yaml Additionally, when attempting to use atlas_full, the segmentation crashes with "IndexError: list index out of range" at label 266, suggesting the data_full/combined_atlas_labels_1mm.yaml config is incomplete (maps only 265 of 334 labels). Questions: Are these additional hippocampal subfields available in FreeSurfer dev or a more recent atlas version? If not yet released, are they planned for a future update? Would you recommend an alternative approach for segmenting these specific structures? Should I post these issues in the official FreeSurfer GitHub repository for broader advice ? Note: I'm running NextBrain via FreeSurfer 8.1 Docker container, due to incompatibility issues in our workstation with the OS. The Docker image was missing both the atlas files and the Python segmentation scripts, so I had to manually download: * atlas_simplified.zip and atlas_full.zip from the FTP server * ERC_bayesian_segmentation scripts from the GitHub dev branch * SuperSynth_August_2025.pth model (1GB) After mounting these files into the container, atlas_simplified works perfectly. However, atlas_full crashes as described above, and neither atlas contains the additional hippocampal subfields visible in your demo. From my understanding, recently (in the past week) a new version with the bayesian segmentation folder was put available but the issues I have remain. Another small note: I had to manually correct the mri_histo_atlas_segment_fireants.sh script because some of the parameters for the python scripts where missing. I'm not sure if that's supposed to happen or if it's a version error. Thank you for your attention. I look forward to receiving a response. Best regards, Luís Researcher, Informatics Engineering University of Minho - ICVS

1 0

mri_cc segmentation fault: Possibly due to M5 silicon incompatibility?
by Ajay Peddada 17 Jan '26

17 Jan '26

External Email - Use Caution Hello FreeSurfer Developers, I am trying to run recon-all on a T1.nii file on my M5 macbook pro. I have set my environment variables per documentation instructions: export FREESURFER_HOME=/Applications/freesurfer/8.1.0 source $FREESURFER_HOME/SetUpFreeSurfer.sh export SUBJECTS_DIR=$FREESURFER_HOME/subjects export FS_LICENSE=$FREESURFER_HOME/license.txt I am concerned if the issue is due to me trying to run freesurfer on an M5 macbook. Also, I'm getting this recurring warning: if: Empty if. then: Command not found. I've searched the list and no similar errors have been reported. Does anyone have any thoughts on how to trouble-shoot this one? Also, Ive attached the recon-all.log in case it's of any use. 1) FreeSurfer version: freesurfer-Linux-centos6_x86_64-stable-pub-v5.3.0 2) Platform: MacOSX Tahoe 26.2 (25C56) 3) uname -a: Darwin Ajays-MBP.attlocal.net 25.2.0 Darwin Kernel Version 25.2.0: Tue Nov 18 21:09:49 PST 2025; root:xnu-12377.61.12~1/RELEASE_ARM64_T8142 arm64 4) recon-all.log: see attached

1 0

freeview overlay_custom: how to create a transparent/“no overlay” band around 0 for diverging maps?
by Camilo Castelblanco Riveros 15 Jan '26

15 Jan '26

External Email - Use Caution Hello Freesurfer Community, I’m visualising a surface overlay in Freeview (fsaverage inflated surface) using custom JSON colourmaps. For sequential features (mean/std), this works fine, but for signed significance-like maps, I want a diverging display with a “dead zone” around 0 (e.g., do not colour values between -2 and +2, so the underlying sulc/curv grey shading remains visible). With built-in colourmaps, Freeview can effectively truncate the middle, but with overlay_custom, the overlay always paints a colour (no alpha/transparent option), so I can’t keep the native curvature shading in the middle range. *Example command*: freeview -f $SUBJECTS_DIR/fsaverage/surf/rh.inflated:annot=aparc.annot:annot_outline=1:overlay=/path/to/sig.mgh:overlay_custom="/path/to/custom.json" -viewport 3d -layout 1 *Question*: Is there a supported way in Freeview to make a value interval (e.g. [-2,2]) render as “no overlay/transparent” when using custom overlays, or is the intended workflow to pre-mask the overlay (e.g., set values in that interval to NaN) before loading? [image: image.png] Thank you! P.S. Is there a way to save images in vector format or PDF? The saving option hasn't allowed me =================================== *Camilo Castelblanco Riveros (He/His)* Neuroscience Ph.D. Candidate (G4) Robertson Lab <https://secure-web.cisco.com/11IId-FSIj2OB9o07kaQlCFuuH7EcJfRBMfniDZE2ADa-k…> & Bujarski Lab <https://secure-web.cisco.com/1dyd3QN20M10b3cHLFbP2mCJe0t7Zk0cwZJ25BnaffnKVn…> Student Interest Group in Neurology (SIGN <https://secure-web.cisco.com/1F4yBWY2i0PLHYagzW9cPA92DGKJvl4QglUz0VOM7rIVDd…>) Co-Leader Global Health Fellow e: ccastelblanco.gr(a)dartmouth.edu Zoom <https://secure-web.cisco.com/1TpnrTMiSXuR0AZwK-n65X5F9L-iGGIlwdoOm1tvaFTtsG…>

2 1

Fwd: Clarification on FreeSurfer recon-all Output and Preprocessing Steps
by Akanksha Upadhyay 15 Jan '26

15 Jan '26

External Email - Use Caution ---------- Forwarded message --------- From: Akanksha Upadhyay <aupadhyay(a)cs.du.ac.in> Date: Sun, Dec 21, 2025 at 6:09 PM Subject: Clarification on FreeSurfer recon-all Output and Preprocessing Steps To: <freesurfer(a)nmr.mgh.harvard.edu> Dear FreeSurfer Support Team, Greetings of the day. I am attempting to replicate a previously published work and have applied the recon-all pipeline on our structural MRI data (T1w.nii) using FreeSurfer. After segmentation and normalization, the output we are obtaining appears unusual to us. I would like to confirm whether these results are expected or if additional preprocessing steps are required prior to running recon-all . I have also attempted reorientation of the input image; however, I was only able to find GUI-based orientation options. I would appreciate your guidance on the following points: 1. Whether the observed segmentation and normalization outputs are expected, or if further preprocessing steps (such as reorientation, intensity correction, or format conversion) should be performed before running recon-all. 2. If manual reorientation is recommended, how should the orientation be determined (e.g., reference axes, alignment criteria, or specific angles)? 3. Any best practices for preparing T1w.nii files to ensure optimal processing with FreeSurfer. I am attaching screenshots of the normalization and segmentation results (aseg output) for your reference. Kindly guide me on how to proceed further. Thank you very much for your time and support.

1 0

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

13 Jan '26

External Email - Use Caution Wondering now what version of MacOS you are running.  While Freesurfer 8.1.0 was not tested on MacOS Tahoe (26.X), we’ve noticed that some mri_synth* commands did not always work correctly on the older MacOS Monterey (12.X) and Ventura (13.X) systems.  At least for FS 8.1.0, we saw some of those commands run more successfully on macOS Sonoma (14.X) and Sequoia (15.X), presumably from bug fixes and updates, including GPU performance.  Please let us know if you find the output is markedly different on Linux compared to (a specific version of) MacOS. - R. On Jan 13, 2026, at 00:09, Tejendra Singh <tejendra.singh.cp(a)gmail.com> wrote:        External Email - Use Caution        Thank you for the suggestion. To clarify, the white block artefacts are present in the saved SynthSR NIfTI output itself, not just during visualization. I observe the same rectangular regions when opening the output in multiple viewers (e.g., Freeview, MRIcron), and they persist after reloading the file.For context, I am running: OS: macOS (Apple Silicon); System: MacBook Air (M-4); FreeSurfer: 8.1.0 (arm64 build); Graphics: Integrated Apple GPU; SynthSR command: mri_synthsr (default model and also tested with --lowfield). I will try re-running one of the affected subjects on a different system (Linux workstation) and compare the SynthSR output to see if the artefact reproduces. I will report back once I have tested this.Thanks again for the helpful suggestion. On Mon, Jan 12, 2026 at 1:35 PM fsbuild <fsbuild(a)contbay.com> wrote:         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@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(a)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/ _______________________________________________ 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/1mpI2pDlMT7dsUzFPkVklGSe5u0LlWDd2br91FPaYO6NZG… The information in this e-mail is intended only for the person to whom it is addressed.  If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Mass General Brigham Compliance HelpLine at https://secure-web.cisco.com/1u4Q66NudVvgBCN5_TnfNHEjisLGf3rtd-0sBX… . _______________________________________________ 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/1mpI2pDlMT7dsUzFPkVklGSe5u0LlWDd2br91FPaYO6NZG…

2 1

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

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