Freesurfer

freesurfer@nmr.mgh.harvard.edu

19 participants
30088 discussions

Re: [Freesurfer] Lacunes
by Antonin Skoch 02 Nov '18

02 Nov '18

External Email - Use Caution Dear Emily, similar errors I usually try to correct by editing wm.mgz (fill the spots by intensity 255). But it does not work always (mainly when the hypointensities are of large extent, it often does not help either). Antonin Hi Emily it's going to be tough to get it past the gray matter lesions. If you have labels we could implement something to fill them in with surrounding tissue intensities I suppose, but we don't have anything at the moment sorry Bruce On Wed, 31 Oct 2018, Emily Schwartz wrote: External Email - Use Caution Hi Freesurfer team, I was wondering if there is a way to edit incorrect surfaces due to lacunes? As you can see in the image the surfaces are incorrect. I tried editing the aseg, but this did not work. I have attached both images in the coronal and axial view as well as FLAIR (obviously there are other errors that need to be fixed in this subject as well). I am using version 6.0 (freesurfer-Linux-centos6_x86_64-stable-pub-v6.0.0-2beb96c). Any guidance is appreciated. Thank you, Emily Emily Schwartz, B.A. Research Assistant Department of Neurology -- Cognitive and Motor Aging Albert Einstein College of Medicine

1 0

Export stats
by Miguel Ángel Rivas Fernández 02 Nov '18

02 Nov '18

External Email - Use Caution Hi Freesurfer devs, Is there any way to export all the stats using an unique command? For example, export the mean area, volume, thickness, thicknessstd, meancurv values of the Desikian-Killiany atlas using an unique command or is only possible export it using separate commands? Thanks in advance, Cheers -- *Miguel Ángel Rivas Fernández*

3 3

Freeview xcb error
by Emily Levin 02 Nov '18

02 Nov '18

External Email - Use Caution Hi Freesurfer experts, I'm using redhat 7 so I installed the development centos7 version of freesurfer: *https://surfer.nmr.mgh.harvard.edu/pub/dist/freesurfer/dev/ <https://surfer.nmr.mgh.harvard.edu/pub/dist/freesurfer/dev/>* I followed the installation instructions and it is having issues with xcb from qt when I try and run freeview: *freeview* *This application failed to start because it could not find or load the Qt platform plugin "xcb"* *in "".* *Reinstalling the application may fix this problem.* *Aborted (core dumped)* I have a previously installed version of freesurfer (6.0) that is statically linked to qt and it runs freeview fine. I already tried setting the QT_QPA_PLATFORM_PLUGIN_PATH environment variable for my qt/5.10.1 plugins directory that contains xcb. Also see the attached output of ldd on freeview. Do you have any other thoughts or actions I can take? Thank you, Emily -- Emily J. Levin Lab Manager Attention & Perception Neuroimaging Lab Boston University Lab: 617-358-1737

2 1

Lacunes
by Emily Schwartz 02 Nov '18

02 Nov '18

External Email - Use Caution Hi Freesurfer team, I was wondering if there is a way to edit incorrect surfaces due to lacunes? As you can see in the image the surfaces are incorrect. I tried editing the aseg, but this did not work. I have attached both images in the coronal and axial view as well as FLAIR (obviously there are other errors that need to be fixed in this subject as well). I am using version 6.0 (freesurfer-Linux-centos6_x86_64-stable-pub-v6.0.0-2beb96c). Any guidance is appreciated. Thank you, Emily Emily Schwartz, B.A. Research Assistant Department of Neurology -- Cognitive and Motor Aging Albert Einstein College of Medicine

2 1

Re: [Freesurfer] Third party dependencies directory not found when building from source
by fsbuild 01 Nov '18

01 Nov '18

External Email - Use Caution Hello Christian, To build Freesurfer e.g., on CentOS7, we use rev 4.8.5 of gcc, g++ and gfortran.  You can tell cmake what compilers to use on the command line with CC=<path>/gcc  CXX=<path>/g++  FC=<path>/gfortran  If you need to specify the gfortran library to use, you can set it with -DGFORTRAN_LIBRARIES=<path to gfortran's lib>/libgfortran.a On Ubuntu 16, the native /usr/bin/gcc is rev 5.4.0, so I installed the gcc 4.9 compilers, and set the above to CC=/usr/bin/gcc-4.9, CXX=/usr/bin/g++-4.9, FC=/usr/bin/gfortran-4.9.  You could try doing the same on Ubuntu 18 as the native gcc may be rev 5.4 or newer, and I don't believe gcc rev 5 will work to build freesurfer. On Oct 31, 2018, at 23:13, Christian Lee Mcdaniel <clm121(a)uga.edu> wrote:        External Email - Use Caution        Hi there,Thank you for directing me to the CMake fswiki. I was able to install the third party packages and execute the command `cmake . -DFS_PACKAGES_DIR="path/to/pkgs"`. However, when I run `make -j4` (after adjusting a few bash files in various libraries to find (#include) other libraries) I am met with numerous similarly-worded link error lines under the step "Linking Fortran executable t4imgs_4dfp" with the general schema:/usr/bin/ld: path/to/fortran-related/filename: relocation R_X86_64_32 against `some.variable.or.file.name` can not be used when making a PIE object; recompile with -fPICwhere path/to/fortran-related/filename is either CMakeFiles/t4imgs_4dfp.dir/some-file.c.o or /usr/lib/gcc/x86_64-linux-gnu/4.8/libgfortran.a(some-function.o)I tried finding a solution online and found the suggestion to add a global -fPIC flag via export CXXFLAGS="$CXXFLAGS -fPIC" (https://stackoverflow.com/questions/13812185/how-to-recompile-with-fpic). I tried this by starting completely over with a new clone of the freesurfer repo and calling cmake . -DFS_PACKAGES_DIR="path/to/pkgs" and the make still failed at the same point. Other suggestions (e.g., to add --enable-shared at ./configure(in that same thread linked above)) suggest adding arguments when configuring, but since this is all part of the make process I'm not sure whether that's possible. Any advice? I did not find any similar issues in the archive. Thank you very much,Christian McDaniel On Oct 31, 2018, at 23:13, Christian Lee Mcdaniel <clm121(a)uga.edu> wrote:        External Email - Use Caution        Hi there,Thank you for directing me to the CMake fswiki. I was able to install the third party packages and execute the command `cmake . -DFS_PACKAGES_DIR="path/to/pkgs"`. However, when I run `make -j4` (after adjusting a few bash files in various libraries to find (#include) other libraries) I am met with numerous similarly-worded link error lines under the step "Linking Fortran executable t4imgs_4dfp" with the general schema:/usr/bin/ld: path/to/fortran-related/filename: relocation R_X86_64_32 against `some..variable.or.file.name` can not be used when making a PIE object; recompile with -fPICwhere path/to/fortran-related/filename is either CMakeFiles/t4imgs_4dfp.dir/some-file.c.o or /usr/lib/gcc/x86_64-linux-gnu/4.8/libgfortran.a(some-function.o)I tried finding a solution online and found the suggestion to add a global -fPIC flag via export CXXFLAGS="$CXXFLAGS -fPIC" (https://stackoverflow.com/questions/13812185/how-to-recompile-with-fpic). I tried this by starting completely over with a new clone of the freesurfer repo and calling cmake . -DFS_PACKAGES_DIR="path/to/pkgs" and the make still failed at the same point. Other suggestions (e.g., to add --enable-shared at ./configure(in that same thread linked above)) suggest adding arguments when configuring, but since this is all part of the make process I'm not sure whether that's possible. Any advice? I did not find any similar issues in the archive. Thank you very much,Christian McDaniel

1 0

Re: [Freesurfer] Third party dependencies directory not found when building from source
by fsbuild 01 Nov '18

01 Nov '18

External Email - Use Caution Hello Christian, Allow me to suggest that you try following the instructions for the Fresrufer cmake build located here, http://freesurfer.net/fswiki/CMake Besides the option to build the 3rd party packages with build_packages.py script, you can also download an archive of the prebuilt packages from http://surfer.nmr.mgh.harvard.edu/pub/data/fspackages/prebuilt/ Another option would be to install the equivalent of the CentOS 7 packages on the Ubuntu system you are using, i.e., you may need to install at least cmake, so you will likely need sudo/admin privileges. Christian Lee Mcdaniel wrote October 30, 2018 at 21:43 via Postbox         External Email - Use Caution         Hello Developers, I am attempting to build Freesurfer from source on my Ubuntu 18.04.1 LTS machine. I have followed the instructions on the freesurfer_linux_developers wiki, and have successfully gotten through steps 1 and 2. However, when trying to install the third party dependencies at step 3 via wget ftp://surfer.nmr.mgh.harvard.edu/pub/dist/fs_supportlibs/prebuilt/centos6_x…, I get the following response: 

3 2

Freesurfer cortical parcellation not covering gray matter in depth
by Yixin Ma 31 Oct '18

31 Oct '18

External Email - Use Caution Hi Bruce, Thanks for your reply to ask me to check lh.pial and rh.pial. I checked and they look good to me as attached picture. Another information is that cortical layers that I generated from the freesurfer recon-all look good as well. So it seems like the problem is only on cortical parcellation? Do you have any other suggestions on what I can check or is there any parameters that I can tune to make cortical parcellation better? Best regards, Yixin [image: Screen Shot 2018-10-27 at 5.28.33 PM.png] [image: Screen Shot 2018-10-30 at 11.44.05 AM.png]

2 1

preproc-sess not passing custom instem to rawfunc2surf-sess
by Jacob Matthews 31 Oct '18

31 Oct '18

External Email - Use Caution Hi FS Team, We've tried running preproc-sess using a custom instem, but it fails at rawfunc2surf-sess. From parsing the code a bit, it appears that while most of the x-sess scripts called in preproc-sess pass their respective outstem's as the instem to the subsequent function, rawfunc2sess is never given an explicit instem and instead builds up its instem based on assumptions from the provided flags. So while our stc-sess outstem according to the log was "fmcpricor.odd" (based on a custom instem "fmcpricor", followed by no mc-sess, and stc-sess), our rawfunc2surf-sess instem defaulted to "fmcpr.odd". Similarly it will also miss some other flags given to preproc-sess. For example, the -nomc flag passed to preproc-sess will not effect rawfunc2surf-sess instem, as it adds "mcpr" to the proposed instem based on the -per-run flag, and does not take the -nomc flag into account. So any preproc-sess calls using -nomc will fail at rawfunc2surf-sess. Am I interpreting this correctly? My assumption is that I will either need to add instem passing to my local preproc-sess file, or otherwise call each of the x-sess scripts directly in my own script rather than using preproc-sess? Thanks, Jacob Matthews MRI Specialist Rotman Research Institute Baycrest Hospital 416-785-2500 ext. 3322

1 0

Options to decrease computation time
by Gregory Book 31 Oct '18

31 Oct '18

External Email - Use Caution I'm curious if there are options to decrease computation time, even at the expense of quality?

3 2

Fwd: Improved segmentation using 3D FLAIR
by Lee Ho Jun 31 Oct '18

31 Oct '18

External Email - Use Caution Hello freesurfer users. I am a radiologist, who is trying to setup a protocol for freesurfer segmentation. I came to know that with 3D T2 weighted images, or 3D FLAIR images, I can get better surface segmentations, and based on search of previous archives of the mailing list, I found that 3D FLAIR images may give better segmentations. My question is that whether applying fat suppression will change the results, since, fat suppression tends to lead to signal dropouts in areas of high susceptibility changes, such as near mastoid air cells. Also, I am a little confused about the bandwidth matched protocols, will "matching" the bandwidths suffice? Best wishes, Ho-Joon Lee.

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