Freesurfer

freesurfer@nmr.mgh.harvard.edu

21 participants
30070 discussions

FS version 8-beta
by Douglas N. Greve 01 Dec '24

01 Dec '24

Hi All, as some of you have figured out, we have a beta version of FS 8.0.0. Please give it a spin and let us know if you have any problems or concerns. The release notes are here https://surfer.nmr.mgh.harvard.edu/fswiki/ReleaseNotes which I have copied below. One bug that we have already found is that you must set environment variable FS_ALLOW_DEEP to 1, eg, export FS_ALLOW_DEEP=1 (for bash) or setenv FS_ALLOW_DEEP 1 (for csh/tcsh) thanks doug * Now uses SynthSeg <https://surfer.nmr.mgh.harvard.edu/fswiki/SynthSeg>, SynthStrip <https://synthstrip.io>, and SynthMorph <https://synthmorph.io> (deep learning algorithms). These make recon-all more robust and much faster. Technically, SynthMorph <https://synthmorph.io> is not needed for recon-all to run, but the nonlinear registration can be used in multimodal analysis to map to the MNI152. * There is now DL automatic segmentation of regions that often cause errors in pial surface placement. These include the major sinuses (sagittal and transverse) as well as a spot between frontal and temporal lobes where the middle cerebral artery (MCA) and dura cause bright areas. * Several fixes in entorhinal cortex. First, the WM in EC is segmented using DL. The WM is often only 0.5mm thick and very hard to see; this often resulted in severe white surface areas in this location requiring manual editing to fix. The new surface now represents gyrus ambiens much better. Second, The white surface is extended to the front of amygdala. Third, the binary cortex of label now covers entorhinal cortex and parts of the temporal pole better allowing the pial surface to extend out to its true edge. * Some bugs where found in and around inferior frontal cortex causes the white surface to extend too far into basal ganglia. This has been fixed. This is above and beyond the issue reported in 7.3.2 below. * All the bug fixes described below are now turned on by default. * These changes will work with longitudinal with the caveat that the volumetric segmentation will not be fully longitudinal. The segmentation will be applied to the time point in the base space, but it will not (and cannot) be initialized with the base segmentation. Initial tests indicated that the change in power was minimal and mixed. * recon-all will now run in about 2h vs 8h on a single CPU. It will, however, required about 24GB of memory to run. * The use of SynthStrip <https://synthstrip.io> should make recon-all much more robust as the old skull stripping method often included extra cerebral tissue. * With synthseg, we are now supplying a segmentation-based total intracrainial volume (sTIV) along with the registration-based estimated TIV (eTIV) * We have applied v8 to several studies (aging, AD crossectional, AD longitudinnal, schizophrenia). There were no major changes in the ability to detect effects, and there was no interaction between effect and version. But let us know how it performs on your studies.

4 7

Re: [Freesurfer] recon-all command stuck before creating pial files
by fsbuild 01 Dec '24

01 Dec '24

External Email - Use Caution  Hello Chemda, Maybe you could let us know what version of freesurfer you are running, what version of Linux or Mac OS your machine is running, and how much memory the machine has.  The recon-all.log could also be useful to send along.    A machine lacking resources like memory, disk space, etc., could hang. - R. On Dec 1, 2024, at 05:41, chemda wiener <chemda10(a)gmail.com> wrote:        External Email - Use Caution         Hi, My name is Chemda Wiener. I am a PhD student in a Neuro-Engineering lab. We are working with MRI scans of patients with epilepsy and I am attempting to use freesurfer to create the mesh .pial files. When I run the freesurfer recon-all command, the command gets stuck at "correcting defects" with the last file it creates being the "lh.defect_chull". It has been stuck for a few days on this part until I finally killed the process. I tried running autorecon2 afterwards on this patient to see if it can pick up from where it stopped but it gets stuck again at the same place. I would appreciate any help in the matter. Thank you in advance for your time and your software!Chemda_______________________________________________ Freesurfer mailing list Freesurfer(a)nmr.mgh.harvard.edu https://secure-web.cisco.com/1s225DDC1H9uV0qs7TECfeaML25tmfh3BDMpJPLr4anEEO…

1 0

Problems with recon-all on M4 iMac
by Torben Ellegaard Lund 27 Nov '24

27 Nov '24

External Email - Use Caution Dear FreeSurfer Developers, I am trying to run the recon-all pipeline on some MP2RAGE images, which have been regularised using the robust reconstruction, in order to avoid the salt and pepper noise in the background. I am running the beta-version of FS 8, and have followed the instructions, including "export FS_ALLOW_DEEP=1”. Synthstrip runs extremely slow, but finishes. I also get a lot of other mgz images : ls /Volumes/SanDiskSSD1/Freesurfer/subjetcs/sub-0271/mri T1.mgz mri_nu_correct.mni.log orig rawavg.mgz synthseg.rca.mgz tmp.mri_mcadura_seg.50836 aseg.auto_noCCseg.mgz nu.mgz orig.mgz rawavg2orig.lta synthstrip.mgz transforms But after a while the job crashes with the message: Node: 'unet/unet_bn_down_0/FusedBatchNormV3' 2 root error(s) found. (0) INVALID_ARGUMENT: input must be 4-dimensional[1,72,72,72,24] [[{{node unet/unet_bn_down_0/FusedBatchNormV3}}]] [[unet/unet_conv_downarm_2_1/BiasAdd/_44]] (1) INVALID_ARGUMENT: input must be 4-dimensional[1,72,72,72,24] [[{{node unet/unet_bn_down_0/FusedBatchNormV3}}]] 0 successful operations. 0 derived errors ignored. [Op:__inference_predict_function_1146] Installation details: 1) FreeSurfer version: MacOS X darwin arm64 (64b) PKG installer 2) Platform:macOS Séquoia version Version 15.1.1 32GB RAM 3) uname -a: Darwin iMac-Fam-Lund.local 24.1.0 Darwin Kernel Version 24.1.0: Thu Nov 14 18:19:02 PST 2024; root:xnu-11215.41.3~13/RELEASE_ARM64_T8132 arm64 4) recon-all.log: see attached I am new to Freesurfer, and really hope you can help. Best Torben Lund Torben Ellegaard Lund Associate Professor, PhD Center of Functionally Integrative Neuroscience (CFIN) Aarhus University Aarhus University Hospital Palle Juul Jensens Boulevard 99 J117 krydspunkt J117 8200 Aarhus N Denmark

1 0

Questions on Using recon-all-clinical for Cortical Thickness Analysis
by 王又劼 26 Nov '24

26 Nov '24

External Email - Use Caution Hello, I have longitudinal imaging data for some participants. Due to the data collection protocol, the T1w image from the first timepoint (TP1) is a clinical sequence (e.g., slice thickness of 5mm or 6.5mm), while subsequent timepoints collected 3D T1w images with a slice thickness of 1mm. My study focuses on analyzing cortical thickness changes in participants over time. I have the following questions: I plan to use recon-all-clinical to analyze the clinical T1w data. However, I noticed that the official documentation mentions that slice thickness significantly impacts cortical thickness estimation. Are there any acceptable methods for comparing cortical thickness changes across different timepoints for the same participants? Can the results of recon-all-clinical be used in FreeSurfer's longitudinal workflow? I have searched archived mails and reviewed studies citing articles related to recon-all-clinical but have not found relevant answers. I would greatly appreciate any help on this matter.

2 1

update from 7.1 to 7.4
by Eszter Boros 23 Nov '24

23 Nov '24

External Email - Use Caution Dear FreeSurfer Team, We did a longitudinal analysis with Freesufer 7.1.1 (two-stage longitudinal model) 2 years ago, and now we would like to add new data/time point. I was wondering if we could use Freesurfer 7.4.1 for this (This is still Freesurfer 7 but 7.4 instead of 7.1). Would that work or we should stick to the exact same version? Please let me know if you need any more information. Thank you for your answer in advance! All the best, Eszter Boros

2 1

mri_glmfit error：ill-conditioned matrix
by Chen Yang 22 Nov '24

22 Nov '24

External Email - Use Caution Dear FreeSurfer experts, I tried to examine the thickness differences between two groups after controlling for the age, gender, and mean thickness, while I got an error message below: ERROR: matrix is ill-conditioned or badly scaled, condno = 1e+08 Possible problem with experimental design: Check for duplicate entries and/or lack of range of continuous variables within a class. If you seek help with this problem, make sure to send: Your command line: mri_glmfit.bin --y glm/FB_demo_lh_CT/lh.FB_demo_lh_CT.thickness.10B.mgh --fsgd glm/FSGD_files/createFSGD/FB_demo_lh_CT/FB_demo_lh_CT.fsgd --C glm/FB_demo_lh_CT/PS-NP.mtx --C glm/FB_demo_lh_CT/NP-PS.mtx --surf 43kids lh --glmdir glm/FB_demo_lh_CT/lh.FB_demo_lh_CT.thickness.10B.glmdir The FSGD file (if using one) And the design matrix above Attempting to diagnose further SumSq: Min=1.258825 (col 10), Max=5.384785 (col 5) ... could not determine the cause of the problem I have searched this problem in the mail archive. I have done the normalization for all covariables (age and mean thickness). Sex and test condition are treated as class variables. So I have no idea about the cause of the error. Here I attached the FSGD file and the matrix. The contrast is 0.5, -0.5, 0.5, -0.5, 0,0,0,0,0,0,0,0. Can someone help me? Thank you! Best regards, Chen -- Chen Yang PhD Student Independent Research Group 'Milestones of Early Childhood Development' Max Planck Institute for Human Cognitive and Brain Sciences Email:yangc@cbs.mpg.de

2 1

mri_synthseg CT segmentation error
by "PD Dr. Susanne Gellißen" 22 Nov '24

22 Nov '24

External Email - Use Caution Hi, When I segment brain CT images using mri_synthseg with the following command line: mri_synthseg — $file_to_segment —o $output — robust —ct —threads 10 —vol FS_vol.csv —qc FS_qc.csv I get segmentation errors concerning the cortex for many CCTs: e.g. some scattered patches of the contralateral right cortex are segmented/labeled as left cortex and vise versa. How can I avoid this error? I think the algorithm should be able to recognize the hemisphere and not scatter patches of left cortex within right cortex? When I use —parchment paper I in the same patients I get similar segmentation errors but different voxels that are segmented with errors. Ho can I improve the segmentation and get at least correct segmentation of the hemispheres? ------------------------------------- PD Dr. med. Susanne Gellißen Department of Diagnostic and Interventional Neuroradiology University Medical Center Hamburg-Eppendorf Martinistraße 52, 20246 Hamburg E-mail: s.siemonsen(a)uke.de -- _____________________________________________________________________ Universitätsklinikum Hamburg-Eppendorf; Körperschaft des öffentlichen Rechts; Gerichtsstand: Hamburg | http://secure-web.cisco.com/1gKbDYZ9Jses5iVoFPW3XeKg6fABwND3QyMrgP5zleSvECU… Vorstandsmitglieder: Prof. Dr. Christian Gerloff (Vorsitzender), Joachim Prölß, Prof. Dr. Blanche Schwappach-Pignataro, Corinna Wriedt _____________________________________________________________________

1 0

Question About Longitudinal TRACULA Analysis
by De Giorgi, Sara 21 Nov '24

21 Nov '24

Dear all, I wanted to ask you a question about longitudinal TRACULA analysis. I'm analyzing a group of patients with different time points. In my configuration file, I used "affine registration from the individual T1 to the MNI T1 template with FLIRT and a correlation ratio cost" as the inter-subject registration method (set interreg = 1). However, I’m considering whether it might be better for my analysis to use "nonlinear registration from the individual FA map to a custom FA template with symmetric normalization from ANTs" (set interreg = 5). I would appreciate your thoughts on this. Thank you in advance. Best regards, Sara

1 0

Longitudinal processing including infant data
by Hendrikse, Chanelle [cbuckle＠sun.ac.za] 21 Nov '24

21 Nov '24

External Email - Use Caution Good day I am writing to enquire regarding the best approach to process longitudinal data acquired across 3 time points. The time points are as follows: * Infants (2-4 weeks old) – only T2 data available for this time point * 2-year-olds – T1 data available * 6-year-olds – T1 data available The approach I’m considering: * run the T2 infant data through synthmorph to create T1 images and then run infant Freesurfer on the synthetic T1 images * run Freesurfer 7.1.1 on 2-year-old data * I have Freesurfer 7.1.1 output for the 6-year-old data * Run the Freesurfer longitudinal pipeline across the 3 time points My questions are: * I know it is highly recommended that the same Freesurfer version be used for longitudinal processing, but this may not be possible given that our dataset includes infant data. What would happen if I ran the normal Freesurfer 7.1.1 on the infant data? Would it be reliable? If not, would it at all be possible to combine the infant Freesurfer output with the version 7.1.1. output in the longitudinal pipeline? * I saw that the beta version of Freesurfer 8 is out and that it includes synthmorph etc. Would it be better to process all the data with this version of Freesurfer? * When do you expect the stable release of Freesurfer 8 to come out? Thank you in advance for your guidance. Kind regards, Chanelle The integrity and confidentiality of this email are governed by these terms. Disclaimer<https://secure-web.cisco.com/1mYiWbfWgJICTMnHZeNDAJNLUO9qWS8T1BhEmMF309NBCg…> Die integriteit en vertroulikheid van hierdie e-pos word deur die volgende bepalings bereël. Vrywaringsklousule<https://secure-web.cisco.com/1mYiWbfWgJICTMnHZeNDAJNLUO9qWS8T1BhEmMF309NBCg…>

1 0

Longitudinal FS Processing
by Yingling, Michael 20 Nov '24

20 Nov '24

External Email - Use Caution Hello, I hope that I am sending this question to the right place. Please let me know if not. I need to do longitudinal FS processing for a study that originally had 3 timepoints (data was processed a couple of years ago using FS for these 3 timepoints). Part 2 of the study has begun, which consists of 1 additional timepoint. What would be the best approach for using FS to process the data at this new timepoint given that the data for the other 3 timepoints have already been processed? Do I need to be careful with using different versions of FS? Thank you! Michael Michael Yingling, MS Senior Statistical Data Analyst Washington University SOM Department of Psychiatry Healthy Mind Lab Phone: (573) 579-1412 ________________________________ The materials in this message are private and may contain Protected Healthcare Information or other information of a sensitive nature. If you are not the intended recipient, be advised that any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited. If you have received this email in error, please immediately notify the sender via telephone or return mail.

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