Freesurfer

freesurfer@nmr.mgh.harvard.edu

21 participants
30071 discussions

Help
by Gantz, Abbey 10 Aug '25

10 Aug '25

External Email - Use Caution Hello, I’m just contacting to ask for advice with the use of recon-all on a participant that has an abnormally structured brain, specifically a compressed Anterior temporal lobe. Recon-all finished without error, but didn’t completely encapsulate the anatomy properly. I ran recon-all using this code: recon-all -s omni07 -i /mnt/sml_share/OMNI/derivatives/fpp/sub-omni0 7/anat/sub-omni07_space-individual_res-1_desc-preproc_Tlw.nii.gz -all -deface -T2 /mnt/sml. share/OMNI/derivatives/fpp/sub-omni07/anat/sub-omni07_space-individual_res-1_desc-preproc_7 2w.nii.gz -T2pial The final recon-all brain mask.mgz file came out like this: [cid:10e2f9c0-f6e6-4f8f-bf38-405a0b106646] Questions: 1. Are there flags to or steps to manually adjust surface place for such cases? 2. Would you recommend disabling certain steps/running in stages to improve the outcome? Thank you for your help, and let me know if there is anything else I can provide to clarify the issue Abigail Gantz PhD Student | Social Memory Lab Tulane Department of Psychology agantz1(a)tulane.edu

2 1

Issue with Pial and WM Surface Distortion
by Mahta Abbaspour 10 Aug '25

10 Aug '25

External Email - Use Caution Hello FreeSurfer team, While performing QC on my recon-all output, I spotted an issue in a scan where the pial surface intersects with the white matter surface, causing distortions across several slices in that region. I overlaid the filled.mgz and white matter segmentation, and the area appears to be fully filled in both masks, so I'm not quite sure why the surfaces are inaccurate here. I'd appreciate any insights and the best way to edit this. I've attached a few example slices below for reference. [image: image.png] [image: image.png] [image: image.png] [image: image.png]

2 3

Consistent Dura Inclusion in QC
by Mahta Abbaspour 10 Aug '25

10 Aug '25

External Email - Use Caution Hello Freesurfer team, I'm currently performing QC on the FreeSurfer outputs and noticed that in many scans, there’s consistent inclusion of dura near the occipital lobe and cerebellum, spanning approximately 10–40 slices sometimes. No dura is included in the anterior or middle slices. I had two questions: - Do you have any idea why this might be happening? The pial and white matter surfaces look fine throughout. - What would be the best way to address this? Should I manually remove the dura and regenerate the surfaces — though this is quite labor-intensive, especially when it affects many slices — or should I consider adjusting the -h parameter for more aggressive skull stripping? However, brainmask also look fine in all other regions. I’ve attached a screenshot from one of the slices for reference. I’d appreciate any suggestions you might have. Thank you. [image: image.png]

2 14

License not received after confirmation
by 강다정 07 Aug '25

07 Aug '25

External Email - Use Caution Dear FreeSurfer team, I successfully confirmed my email address for FreeSurfer license registration, but I have not received the license.txt file yet. Could you please send it again? My registered email: doramisister(a)naver.com Affiliation: Yonsei Unsiversity Yoon lab Thank you very much. Best regards, Dajeong Kang

1 0

Discrete pial surface
by Yujia Huang 06 Aug '25

06 Aug '25

External Email - Use Caution Hi Freesurfers, I have a question reagrding to how the pial surface is generated. For an oversimplified description, I assume each subject's pial surface was initially created with a spherical shaped 3D-topology model. Then each verticie on the sphere is assigned to specific location of the pial surface to cover the subject's brain. If that is the case, then I don't understand the result illustrated below in the figure. The figure is showing anatomical T1w image overlay with left and right pial surface from recon-all (command used at the end), while the red cross is targeting a pial surface located within cortical region. Although there is some mis-segmentation error nearby, I don't see the pial surface inside of the cortical region is connecting to any outer surface when I move the image slice back and forth. Is this allowed or even possible if the initial spherical surface can generate such discrete surface that is disconnected from the whole outer brain region? [cid:20a64222-bc84-4c74-9935-ffd295eac833] Does anyone know what is the cause of this and how to trouble shoot? Thanks in advance! 1. Freesurfer version: freesurfer-linux-ubuntu22_x86_64-7.4.1-20230614-7eb8460 2. Platform: Ubuntu 24.04.2 LTS 3. uname -a: Linux hawaii 6.11.0-26-generic #26~24.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Apr 17 19:20:47 UTC 2 x86_64 x86_64 x86_64 GNU/Linux 4. Command and steps used for recon-all: * recon-all -autorecon1 -s <subj_id> -i <mp2rage_input> * check and verify skull strip quality * recon-all -autorecon2 -autorecon3 -s <subj_id> Best, Yujia Huang Research Engineer Advanced Imaging Research Center UT Southwestern Medical Center 2201 Inwood Rd. Dallas, TX, USA, 75390 Office: 214.645.5241 Pronouns: He, Him, His ________________________________ UT Southwestern Medical Center The future of medicine, today.

3 3

Re: Reading mca-dura.both-lh.nstd21.fhs.h5 error
by fsbuild 05 Aug '25

05 Aug '25

External Email - Use Caution   This looks like a file system access/permissions problem, i.e., the hdf5 python code being run from where freesurfer is installed, namely the code in /usr/local/freesurfer/8.1.0/python/lib/python3.8/site-packages/h5py/_hl/files.py claims it cannot see or read the file via the absolute path /models/mca-dura.both-lh.nstd21.fhs.h5 Such an error might occur if /models/mca-dura.both-lh..nstd21.fhs.h5 is not a valid hdf5 file, or /models is a mount point or soft link in your computers file system.  It does however look like some previous code was able to open /models/mca-dura.both-lh.nstd21.fhs.h5 and load in the weights. You could try cd’ing to /usr/local/freesurfer/8.1.0/python/lib/python3.8/site-packages and see if you can read the file, i.e., “cat /models/mca-dura.both-lh.nstd21.fhs.h5”.  I would at least try re-locating that model file to be under a known path where you always have read/write permission like your home directory. - R. /usr/local/freesurfer/8.1.0/python/scripts/mri_sclimbic_seg --model /models/mca-dura.both-lh.nstd21.fhs.h5 --ctab /data2/FST1T2Res/002_S_0295_006/mri/tmp.mri_mcadura_seg.271283/dumca.lh.ctab --keep_ac --percentile 99.9 --vmp --output-base mcadura --conform --logfile mri_mcadura_seg.log --no-cite-sclimbic --threads 1 --fov 72 --i /data2/FST1T2Res/002_S_0295_006/mri/tmp.mri_mcadura_seg.271283/invol.lh.crop.nii.gz --o /data2/FST1T2Res/002_S_0295_006/mri/tmp.mri_mcadura_seg.271283/seg.lh.mgz Keeping anterior commissure in vols and stats Using CPU Using 1 thread(s) Loaded lookup table /data2/FST1T2Res/002_S_0295_006/mri/tmp.mri_mcadura_seg.271283/dumca.lh.ctab Loaded population stats /usr/local/freesurfer/8.1.0/models/sclimbic..volstats.csv nb_labels 2 inshape (72, 72, 72) features 24   (None, 72, 72, 72, 1) Loading weights from /models/mca-dura.both-lh.nstd21.fhs.h5 -----------------------     segmenter = LimbicSegmenter(model_file=model_file,   File "/usr/local/freesurfer/8.1.0/python/scripts/mri_sclimbic_seg", line 425, in __init__     self.model.load_weights(model_file, by_name=True)   File "/usr/local/freesurfer/8.1.0/python/lib/python3.8/site-packages/keras/src/utils/traceback_utils.py", line 70, in error_handler     raise e.with_traceback(filtered_tb) from None   File "/usr/local/freesurfer/8.1.0/python/lib/python3..8/site-packages/h5py/_hl/files.py", line 562, in __init__     fid = make_fid(name, mode, userblock_size, fapl, fcpl, swmr=swmr)   File "/usr/local/freesurfer/8.1.0/python/lib/python3..8/site-packages/h5py/_hl/files.py", line 235, in make_fid     fid = h5f.open(name, flags, fapl=fapl)   File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper   File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper   File "h5py/h5f.pyx", line 102, in h5py.h5f.open FileNotFoundError: [Errno 2] Unable to synchronously open file (unable to open file: name = '/models/mca-dura.both-lh.nstd21.fhs.h5', errno = 2, error message = 'No such file or directory', flags = 0, o_flags = 0) -- This is a copy of a message which has been originally sent to: "To" field: freesurfer(a)nmr.mgh.harvard.edu The message has been sent by https://secure-web.cisco.com/1K-uMBwgIPgOmDbmGlFSw0j7kzReQeeg5ieEo0Di0x-0Ti… and the used account name was: SSL  

1 0

Individual pre- and post-values for cortical thickness after a paired analysis
by Franziska Mund 05 Aug '25

05 Aug '25

External Email - Use Caution Dear FreeSurfer Developers, we have a longitudinal study design with one group and ran a paired analysis in FreeSurfer V6 (as described here: https://secure-web.cisco.com/1V0gSNyIvhCtX3j8K-mzIp_bu0OnK1ZDZxKlOeenlA2T-l…) Afterwards we corrected for multiple comparisons (as described here: https://secure-web.cisco.com/1kF3xxriMcabTqMwlV5PjrV7Bz3-ihiWEkX08AlNLpslyH…) One cluster showed a significant change for cortical thickness and the average values of each subject were located in the “y.ocn.dat”. However, we are now wondering how it is possible to get the individual pre- and post-values for that specific cluster for all subjects. Is there any method to do so? Any assistance is very much appreciated! Thanks in advance. Best, Franziska Dr. Franziska Mund Wissenschaftliche Mitarbeiterin Sportmedizinisches Institut [cid:image001.png@01DC016F.4B7B4640]<http://secure-web.cisco.com/1MXBasQV5Cu2ljjndrv0cARnfi2yk8RGCa8zldfWWltrJ-c…> Warburger Straße 100 33098 Paderborn Raum SP.0.508 Telefon +49 5251 60-3187 Mail mund(a)sportmed.uni-paderborn.de<mailto:mund@sportmed.uni-paderborn.de> Web http://secure-web.cisco.com/1mtAP8Z87ARwP9B1vJH2CI98DdfTwRbHMRsFSxOHk8YZF_j…<http://secure-web.cisco.com/1HjqZvGkQ5IUZLtZDOdlUAFPmJpXhcqgZc6dhIvSG816p-t…> [UPB Instagram]<https://secure-web.cisco.com/1Nd--jQ71ePt5LJ_Wmk5O8d4Kr2kWpH55lAu0v9n9LQVjG…> [UPB Facebook] <https://secure-web.cisco.com/1qI6cvPugysabKPf7WJIn-pK1ebwXTYqTkS1Cq7h9IJDbH…> [HNI LinkedIn] <https://secure-web.cisco.com/1QjEIu8F4_OFs-qMy1zfRJ09N60yJv7f8jN4vl0JbHSAUa…> [HNI Twitter] <https://secure-web.cisco.com/1RiPhNxpWoH_W7GY4WC-7wEUB04OrzGVeskAVxjwYN9y6H…> [HNI YouTube] <https://secure-web.cisco.com/11K2yPkMlVnMf6mztj9krvQ4gM44-jTlMplFJReirNNc8t…>

2 5

Freesurfer labels not generated and have some corrupted label files
by Samantha Hendry 05 Aug '25

05 Aug '25

External Email - Use Caution Hello FreeSurfer Developers, I'm attempting to run freesurfer on a resampled T1w scan and it has come up with segmentation fault error at the stages of the atlas. I was wondering if the resampled T1w scan has caused this error and this is the only difference between processing before. I am really stuck and its not a system compatability issue as I have been able to run all of my freesurfer output without error before so anything would be great. 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-v6.0.0-2beb96c 2) Platform: Performed on MobaXterm on windows computer using eddie super computer and Edinburgh university 3) uname -a: Linux node1c08.ecdf.ed.ac.uk 5.14.0-427.33.1.el9_4.x86_64 #1 SMP PREEMPT_DYNAMIC Wed Aug 28 17:34:59 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux 4) recon-all.log: see attached Not sure if the above is fully correct. Any help would be greatly appreciated! Best wishes Samantha Hendry The University of Edinburgh is a charitable body, registered in Scotland, with registration number SC005336. Is e buidheann carthannais a th' ann an Oilthigh Dh?n ?ideann, cl?raichte an Alba, ?ireamh cl?raidh SC005336.

2 3

Segmentation of the Hypothalamus in FreeSurfer 7.4.1
by Juen Guo 04 Aug '25

04 Aug '25

External Email - Use Caution Dear FreeSurfer Expert, I am trying to segment the subunits of the hypothalamus using FreeSurfer 7.4.1, but the segment_subregions command does not include an option for the hypothalamus in this version. I found that version 7.2.0 includes the command mri_segment_hypothalamic_subunits, but this command is not executable in version 7.4.1. My questions are: 1. Do I have to downgrade to 7.2.0 to use this function? 2. Why was this function discontinued in version 7.4? Was it removed due to defects or other issues? 3. Is there any alternative way to segment the hypothalamus in FreeSurfer 7.4.1? Best, Juen

3 4

Best way to map aseg.mgz to denoised T1 (native space)
by Razieh Salesi 01 Aug '25

01 Aug '25

External Email - Use Caution Dear FreeSurfer team, I have a workflow question and would appreciate your guidance: I ran `recon-all` on the original `T1.nii.gz` image (208×240×256), which produced the standard FreeSurfer outputs, including `mri/aseg.mgz`, conformed to 256×256×256. Separately, I performed N4 bias correction using ANTs and denoising (non-local means) on the original `T1.nii.gz`, resulting in a denoised volume (`T1_N4_naonlm3d.nii.gz`) with the original dimensions (208×240×256). My goal is to: 1. Create an accurate white matter (WM) mask aligned with this denoised native-space T1, 2. Use it with an EPC image (T1/T2 ratio) for perivascular space (PVS) mapping using Frangi filtering. My main question is: What is the best approach to transform `aseg.mgz` (in conformed space) to this denoised T1 native space, preserving anatomical accuracy? Should I: - Use `mri_label2vol` with `--regheader` and `--temp T1_N4_naonlm3d.nii.gz`? - Register `T1_N4_naonlm3d` back to `orig.mgz` and invert the matrix? - Or is there a more accurate method (e.g., using `tkregister2`, `bbregister`, or resampling via FreeSurfer)? Any recommendations or practices would be very helpful. Best regards, Razieh Salesi

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