Freesurfer January 2024

freesurfer@nmr.mgh.harvard.edu

70 participants
111 discussions

Registration of MPF to FLASH 3D image
by ERNEST KISSI 12 Jan '24

12 Jan '24

External Email - Use Caution Dear Freesurfer team, I messaged the team on my MPF map analysis and a suggestion was made that I register the MPF maps on the T1 MPRAGE images. However, I did not acquire MPRAGE but FLASH images. Can I used those instead ? Again, I tried even using only the T1 FLASH images and it did not work so I am not sure whether the registration of the MPF maps on it will work or not. Is there anyway around this? Thank you.

3 5

Re: [Freesurfer] Trouble with ssh tunneling & Freeview
by fsbuild 12 Jan '24

12 Jan '24

External Email - Use Caution A remote connection working to display images from the glxgears or xeyes commands run on the linux host forwarded to your local machine does not mean Freeview graphics will do the same.  If Freeview still displays an empty or black screen, then there could be issues with OpenGL graphics running on or being sent from the linux host.   The system requirements for Freesurfer include the following for running Freeview, Graphics card: 3D graphics card with its own graphics memory & accelerated OpenGL drivers Updating the OpenGL drivers on the linux host will sometimes enable Freeview to display images remotely.  My understanding is Linux servers need Mesa packages installed to support newer versions of OpenGL.  If the Freesurfer Debian or RPM package installer was used to install the Freesurfer distribution on the Linux host, then at least some of those packages should now be installed, i.e., we do not recommend using the tar archive to “install” Freesurfer.  We don’t know how EXPANSE installs/maintains Freesurfer or what their servers support for graphics hardware. GLX is the X-windows extension for OpenGL graphics that can be tunneled thru a remote SSH connection.  But GLX may not work with the newer versions of OpenGL over ssh.  A workaround for this is to install and use VirtualGL on the Linux host.  If VirtualGL is supported/installed on EXPANSE, try prefacing the freeview command with “vglrun” - see https://secure-web.cisco.com/10kzDiIcCSK5wSi3KZ4WHry8djBXAuam589oK…  We use VirtualGL and graphics acceleration on a few Linux workstations to accommodate Freeview running on Linux and viewing its 3D graphics thru a remote VPN connection. XQuartz on the Mac is the equivalent of X-Windows on Linux.   The Mac should be using Xquartz 2.8.2 or newer for use with the Freesurtfer 7.4.X releases.  I think X-Quartz will open automatically to view X-Windows based graphics on the Mac. If the EXPANSE linux machines are setup to support remote connections to view 3D X-windows graphics then the IT support staff will know about the options above.  File/open a ticket with them, ask them what they support, and how to setup your environment.  If they do not support it, an alternative is to install Freesurfer on your Mac and download/copy over the data to view with Freeview. - R. On Jan 11, 2024, at 14:18, Dong, Yilei <yid025(a)health.ucsd.edu> wrote:        External Email - Use Caution        Hi Freesurfer team, I have Freesurfer installed on UCSD's Expanse HPC, which uses the Rocky Linux 8 operating system. Whenever we try to do ssh tunneling and more specifically, logging into our HPC via ssh -X to open up Freeview's GUI, it always aborts. Prior to logging into my HPC from my Mac laptop's terminal, I always open up XQuartz. Another colleague of mine has also encountered similar issues, where they have managed to get X11 and Xeyes to run, but also Freeview's window keeps aborting. We have also tried to follow instructions from the following links when it comes to getting ssh tunneling to work correctly in hopes of making Freeview's GUI pop up: MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to behttps://http://secure-web.cisco.com/12UaY28IjOvcCuzqwhVkemgeWWYzbyjV_gOl5… has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://secure-web.cisco.com/13jhlKniK4ot1jAvrTEdLmUxfeVBEFlM0Xf3YM… we operate from a Mac computer, is it better to use a VNC or XQuartz for ssh tunneling? In addition, can we find an error log in Freesurfer's directory to find out exactly why Freeview keeps aborting in our circumstance?Thank you. Sincerely,Yilei <Screen Shot 2024-01-11 at 11.01.27 AM.png>_______________________________________________Freesurfer mailing listFreesurfer@nmr.mgh.harvard.eduhttps://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

2 1

Including superficial white matter in cortical parcellation
by Thomas, George 12 Jan '24

12 Jan '24

External Email - Use Caution Dear Freesurfer Experts, I am interested in generating a volumetric segmentation and parcellation of the HCP-MMP1 atlas using mri_surf2volseg - and I have previously done this like so: mri_surf2volseg --o $sub/HCP-MMP1_fsSpace.mgz \ --label-cortex --i $sub/mri/aseg.mgz --threads 4 \ --lh-annot $sub/label/lh.HCP-MMP1.annot 1000 --lh-cortex-mask $sub/label/lh.cortex.label \ --lh-white $sub/surf/lh.white --lh-pial $sub/surf/lh.pial \ --rh-annot $sub/label/rh.HCP-MMP1.annot 2000 --rh-cortex-mask $sub/label/rh.cortex.label \ --rh-white $sub/surf/rh.white --rh-pial $sub/surf/rh.pial However rather than use the regular white matter surface generated by recon-all (as above), I want to parcellate between the pial surface and a superficial white matter surface 1mm below the GM/WM boundary. I have generated this superficial white matter (SWM) surface for both hemispheres like so: mris_expand lh.white -1 lh.SWM.1mm My first though was just to replace the --rh-white / --lh-white inputs with the SWM surface, but after running this, the output remains the same. I think I may also need use a different input than aseg.mgz, but not sure what this would be or how to generate it. Any assistance you could offer with the would be greatly appreciated. Many thanks, George

2 1

poor inferior cerebellum segmentation
by Carlos Alcocer 12 Jan '24

12 Jan '24

External Email - Use Caution Hello, Inferior portion of cerebellum are not segmented well. It is included in skullstrip and I tried adding control points in cerebellar WM but this did not help. I am not interested in cerebellum but I do need eTIV. Will this poor cerebellar segmentation affect eTIV calculation? Thank you, Carlos

4 3

Issue with mri_glmfit-sim
by O'Neill, Alexandra Grace 12 Jan '24

12 Jan '24

Hello FS community, To correct for multiple comparisons for the results of a parametric modulation GLM analysis, I have run mri_glmfit-sim using Freesurfer version 6.0 with the following command: mri_glmfit-sim \ --debug --glmdir /cluster/eryilmaz/users/NFE/NFE_Subjects_ago/Analyses_NFE/18_NFE_Patients/SIRPSEC_LOADREG_RH/LoadDepSec/OSGM/osgm \ --perm 1000 2.0 abs \ --cwp 0.05 \ --2spaces \ --bg 1 The code completes without errors, however none of the output files mentioned in this tutorial<https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/MultipleComparisonsV6.…> are created. Below are the contents of the mriglm-fit.log file in case that is helpful. $Id: mri_glmfit.c,v 1.241.2.4 2016/12/08 22:02:40 zkaufman Exp $ cwd /autofs/cluster/eryilmaz/users/NFE/NFE_Subjects_ago cmdline mri_glmfit.bin --y /cluster/eryilmaz/users/NFE/NFE_Subjects_ago/Analyses_NFE/18_NFE_Patients/SIRPSEC_LOADREG_RH/LoadDepNorm/ces.nii.gz --wls /cluster/eryilmaz/users/NFE/NFE_Subjects_ago/Analyses_NFE/18_NFE_Patients/SIRPSEC_LOADREG_RH/LoadDepNorm/cesvar.nii.gz --fsgd /cluster/eryilmaz/users/FSGD/18_NFE_Patients.fsgd --osgm --glmdir /cluster/eryilmaz/users/NFE/NFE_Subjects_ago/Analyses_NFE/18_NFE_Patients/SIRPSEC_LOADREG_RH/LoadDepNorm/OSGM --surf fsaverage rh --fwhm 6 sysname Linux hostname tempest.nmr.mgh.harvard.edu machine x86_64 user ago5 FixVertexAreaFlag = 1 UseMaskWithSmoothing 1 fwhm 6.000000 niters 26.000000 OneSampleGroupMean 1 y /autofs/cluster/eryilmaz/users/NFE/NFE_Subjects_ago/Analyses_NFE/18_NFE_Patients/SIRPSEC_LOADREG_RH/LoadDepNorm/ces.nii.gz logyflag 0 usedti 0 FSGD /cluster/eryilmaz/users/FSGD/18_NFE_Patients.fsgd labelmask /cluster/eryilmaz/users/NFE/NFE_Subjects_ago/fsaverage/label/rh.cortex.label maskinv 0 glmdir /cluster/eryilmaz/users/NFE/NFE_Subjects_ago/Analyses_NFE/18_NFE_Patients/SIRPSEC_LOADREG_RH/LoadDepNorm/OSGM IllCondOK 0 ReScaleX 1 DoFFx 0 wFile /cluster/eryilmaz/users/NFE/NFE_Subjects_ago/Analyses_NFE/18_NFE_Patients/SIRPSEC_LOADREG_RH/LoadDepNorm/cesvar.nii.gz weightinv 1 weightsqrt 1 ResidualFWHM 12.847434 SearchSpace 74490.844410 anattype surface Do you have any idea how to fix this problem? Thanks so much in advance! Alexandra Alexandra G. O'Neill (she/her) Clinical Research Coordinator Massachusetts General Hospital Department of Psychiatry Division of Neuropsychiatry and Neuromodulation 149 13th Street, Charlestown, MA 02129 T: (617) 726-8753 | E: agoneill(a)mgh.harvard.edu<mailto:agoneill@mgh.harvard.edu>

2 1

Tksurfer and freeview
by JAEHYUN KIM 12 Jan '24

12 Jan '24

External Email - Use Caution Hello, I attempted to switch tksurfer to freeview in the script, but it seems like I need to change arguments "tcl" to something that can be applied to freeview. Option 'tcl' not recognized. What argument could possibly be used to replace *tcl* in the script such as *tmp.tcl* and *lh pial -tcl*?? Best, Ed

2 1

modify intensity normalization pipeline
by Carlos Alcocer 12 Jan '24

12 Jan '24

External Email - Use Caution Hello, Is there a way to modify intensity normalization pipeline in recon-all script? The temporal lobe is being very poorly segmented because white matter is being segmented as gray matter and therefore gray matter does not reach outer boundary of pial surface. Thank you, Carlos

2 1

Re: [Freesurfer] Freesurfer Digest, Vol 239, Issue 1
by Spapens, Jeltje (NP) 12 Jan '24

12 Jan '24

External Email - Use Caution Hi Douglas, I hope you had some nice holidays and a happy beginning of 2024. Thank you for your suggestion. I have added the intermediate picture and changed the MRI_vol2vol target volume from the original scan to the MRI/T1 recon_all output. The latter one seemed to prevent the orientation shift. However, the registration still does not seem of very good quality. I have tried init-fsl, header and spm, but the first one gave the best results. Do you have another advice for me? bbregister --s MRS_200_recon_all --mov /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_MT_tfl_stripped.nii --t2 --int /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/presurf_MPRAGEise/MRS_200_UNI_MPRAGEised_stripped.nii --reg /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_NM2mp2rage.dat --init-fsl mri_vol2vol --mov /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_MT_tfl_stripped.nii --targ /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_recon_all/mri/T1.mgz --o /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_MT_tfl_regt1.nii.gz --reg /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_NM2mp2rage.dat --cubic Kind regards, Jeltje Spapens ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ Message: 15 Date: Mon, 1 Jan 2024 16:07:00 -0500 From: "Douglas N. Greve" <dgreve(a)mgh.harvard.edu <mailto:dgreve@mgh.harvard.edu>> Subject: Re: [Freesurfer] FW: Boundary based registration To: freesurfer(a)nmr.mgh.harvard.edu <mailto:freesurfer@nmr.mgh.harvard.edu> Message-ID: <e31f2342-e539-4ecc-b146-2b059dab1b6c(a)mgh.harvard.edu <mailto:e31f2342-e539-4ecc-b146-2b059dab1b6c@mgh.harvard.edu>> Content-Type: text/plain; charset="utf-8" Have you tried using the "intermediate" option (--int) in bbregister? The intermediate volume should be a whole brain volume and it should share a RAS space with the slab volume (eg, if they were acquired in the same session; load both the slab and the intermediate in freeview to see if they are aligned). You can run bbregister with --help to get more info On 12/13/2023 2:48 AM, Spapens, Jeltje (NP) wrote: > > ????????External Email - Use Caution > > Hello Freesurfer developers, > > I want to register an MRI slap to a full FoV UNI scan of the same > subject, but I get a false reorientation with my code (see attached > image). I have tried multiple input and tried the script on different > subjects. > > I also changed som flags such as -init-header and --cubic, but all > seem to give the same results. > > bbregister --s MRS_200_recon_all --mov > /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_MT_tfl_stripped.mgz > --reg > /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_NM2mp2rage.dat > --t2 --init-fsl > > mri_vol2vol--mov > /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_MT_tfl_stripped.mgz--targ > /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_acpc_upsamp.nii > --o > /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_MT_tfl_regt1.nii.gz > --reg > /Users/robspapens/Desktop/Thesis/Subjects/MRS_200/MRS_200_NM2mp2rage.dat > ?trilin > > I would appreciate to hear your thoughts on this problem. > > Have a nice weekend and kind regards, > > Jeltje Spapens >

2 1

FLAIR vs. T1 question
by aliyasuri＠yahoo.com 12 Jan '24

12 Jan '24

External Email - Use Caution Dear FS mailing list, My apologies if this is a naive question. I have FLAIR images that were acquired with much higher resolution than the T1 images for the same subject. Does it make sense to run recon-all using FLAIR as the input image? I tried it, and my segmentations look reasonable. I am mainly interested in volumetric WM segmentations. Am I missing something basic, or is there something else I can try that leverages the FLAIR (other than using it to refine the pial surfaces)? Thank you.

2 1

Bug in FSFAST functional connectivity analysis
by Douglas N. Greve 12 Jan '24

12 Jan '24

There is a bug in the FSFAST <https://surfer.nmr.mgh.harvard.edu/fswiki/FsFast> functional connectivity seed-based analysis (current version and all previous versions). The amplitude of the seed waveforms are not rescaled in any way. This can affect the amplitude of the regression coefficient with the rest of the image (which is rescaled). This will not have any effect under several circumstances: (1) it will have no impact on the sig.nii.gz (p-value) or the Pearson correlation (pcc) at the individual level, (2) the seed uses a pca/svd, (3) if the intensity scale in each image/subject is the same (or close to it), or (4) if you are passing up the Pearson correlation (pcc) to the higher level. Note that the typical FSFAST analysis passes up the contrast effect size (ces) to the higher level which will be affected by this bug. The problem will manifest itself as extra variability at the group level due to the variability of the mean intensity at the individual level (so your group results should probably improve when fixed). This has been fixed in the dev version of FS and will be in the next release. The change will require that you specify the type of rescaling you want. If you run the new fcseed-sess on an old configuration that does not have a scaling method specified, it will exit with a error message (this can be overridden by adding -no-require-rescale-type). Please reach out with any questions and sorry about the inconvenience. doug

1 0

← Newer
1
2
3
4
5
6
7
8
9
...
12
Older →

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Freesurfer January 2024