Freesurfer

freesurfer@nmr.mgh.harvard.edu

13 participants
30061 discussions

func2sph-sess target brain setting
by raij＠nmr.mgh.harvard.edu 07 Jun '06

07 Jun '06

Hi, Backround: I am running a fixed effects group analysis with fsaverage (mne305) as the target brain, and merging the activations across subjects on the spherical surface. The individual functional sessions have been co-registered to the individuals' structural brain reconstruction (i.e. the register.dat file inside the bold directory links the functional and structural data within subjects only). Problem: I am not sure how to set the group analysis func2sph-sess and paint-sess "target brain" parameters correctly. Version and environment: I am using the current Martinos Center FreeSurfer standard installation (vs. >3) under nmr-std-env and all the structural reconstructions are new as well. Outline of the analysis (list of commands): preprocessing mkanalysis-sess.new selxavg-sess mkcontrast-sess stxgrinder-sess func2sph-sess <- sphsmooth-sess isxavg-fe-sess stxgrinder-sess paint-sess <- Specific question (with possible answers): *** 1 *** Should I set the target brain already at func2sph-sess: func2sph-sess -analysis XXX -sf sessid -df sesspar -trgsubject fsaverage (if the -trgsubject setting is omitted, then the target appears be the default = ico order 7) *** 2 *** Or is it sufficient (if I ignore the -trgsubject parameter above) to set: paint-sess -analysis XXX -contrast YYY -subject fsaverage -space sph -isxavg fixed -s fegroup -df group.sesspar (here fegroup = the folder where the fixed effects analysis is stored, set at isxavg-fe-sess -s fegroup) *** 3 *** Or must/should I do both 1 and 2 of the above. I am using surf-sess/tksurfer to view the results and set -subject fsaverage even there. I get results that look reasonable either way. I just wonder what is the difference between the methods, and if either method is unambiguously correct. Thanks, Tommi --- Tommi Raij, M.D., Ph.D. MGH/MIT/HMS Athinoula A. Martinos Center for Biomedical Imaging Bldg 149, 13th St Charlestown, MA 02129 U.S.A.

2 3

functional group results on structural data
by Paymon Hosseini 07 Jun '06

07 Jun '06

hi i used the following commands to create a group subjects: func2tal-sess -res 4 -analysis rk_enc2_sm5 -sf sessid -df sesspar isxavg-fe-sess -analysis rk_enc2_sm5 -sf sessid -df sesspar -group Y15ENC2_sm5_grp -space tal stxgrinder-sess -analysis rk_enc2_sm5 -all -space tal -s Y15ENC2_sm5_grp i use this command to view the results: tkmedit-sess -analysis rk_enc2_sm5 -map sig -s Y15ENC2_sm5_grp -space tal -isxavg fixed -c omnibus the output is: -------------------------------------------------------------- tkmedit-sess logfile is /autofs/space/engram_012/users/pman/rk_fmri/code/log/tkmedit-sess.log -------------------------------------------------------------- setenv SUBJECTS_DIR /space/engram/12/users/pman/rk_fmri/STRUCTS INFO: Found session in: ./Y15ENC2_sm5_grp INFO: Y15ENC2_sm5_grp is a group average INFO: Found session in: ./Y15ENC2_sm5_grp INFO: Y15ENC2_sm5_grp is a group average INFO: subject is talairach INFO: subject is talairach INFO: Anatomical directory set to /autofs/space/sake_003/users/inverse/anat_data/talairach/mri/orig ------------------------ SUBJECTS_DIR /space/engram/12/users/pman/rk_fmri/STRUCTS /autofs/space/engram_012/users/pman/rk_fmri/code tkmedit talairach orig -overlay /autofs/space/engram_012/users/pman/rk_fmri/code/Y15ENC2_sm5_grp/bold/rk_enc2_sm5/tal-ffx/omnibus/fsig_000.bfloat -overlay-reg /autofs/space/engram_012/users/pman/rk_fmri/code/Y15ENC2_sm5_grp/bold/rk_enc2_sm5/tal-ffx/register.dat -fthresh 2 -fmid 5 -fslope 1 -aux brain ------------------------ Reading 7 control points... Reading /usr/local/freesurfer/stable/lib/tcl/tkm_common.tcl Reading /usr/local/freesurfer/stable/lib/ the figure looks fine, but, i would like to be able to see the functional maps on an average-subject hi-res structural map. can you please tell me how to do this? i am using "nmrenv" on engram.nmr.mgh. thanks a lot, paymon

2 1

RE: [Freesurfer] Segmentation optimization (fwd)
by Doug Greve 07 Jun '06

07 Jun '06

As far as the orientation goes, mri_convert will assume that the spm .mat file uses an RAS geometry. You should look at the volume in tkmedit and go through each orientation to make sure that each looks correct. ie, for sag the nose should be to the right, for ax the nose should be at the top. For left/right, tkmedit displays in radiological convention, so make sure your fiducial is on the correct side. doug ---------- Forwarded message ---------- Date: Wed, 7 Jun 2006 16:02:06 +0100 From: Joao Pereira <jmsp2(a)wbic.cam.ac.uk> To: Bruce Fischl <fischl(a)nmr.mgh.harvard.edu> Subject: RE: [Freesurfer] Segmentation optimization Here are the acquisistion parameters and some other data as well (images acquired with 1.5T field): Coronal 3D FSPGR IR GENESIS_SIGNA Sequence: GR Seq Variant: SS\SP\SK Flip Angle: 20 Options: FAST_GEMS\VB_GEMS\EDR_GEMS\PFF TR = 13.9 TI = 650 TE = 4.2 NEX = 1 pixel bandwidth = 81.4 Acq. Matrix = 0\256\224\0 Spatial Resolution = 0.982143 220mmx220mmx174mm (256x256x116) voxel size: 0.86x0.86x1.50mm^3 1.164 pixels per mm Also, another question that need to check: I transform the Analyze images so that they are SPM oriented (RAS). When I the convert them with mri_convert (no other options set), this orientation (neurological) is maintained, right? (At least it seems so, from the ijk matrix). Please let me know if this interpretation is correct. Once more, thank you! Joao -----Original Message----- From: freesurfer-bounces(a)nmr.mgh.harvard.edu [mailto:freesurfer-bounces@nmr.mgh.harvard.edu]On Behalf Of Bruce Fischl Sent: 06 June 2006 17:50 To: Joao Pereira Cc: freesurfer(a)nmr.mgh.harvard.edu Subject: Re: [Freesurfer] Segmentation optimization what is the resolution of your images? You might as well send us the complete set of acquisition parameters. Bruce On Tue, 6 Jun 2006, Joao Pereira wrote: > Hello again, > > I've been having some problems with FS segmentation, especially around the > temporal lobes, where there is a clear underestimation of the pial surface > (and manual edits are not being too successful). The problem is clearly > related to the poor resolution of the used images. > > Are there any parameters to change in order to optimise the segmentation > process? > > Also, in order to evaluate the segmentation I want to use the grey matter > segmented volume mask. Is it the ?h.ribbon.mgz files or should I consider > something else? > > Thank you again. > > Joao Pereira > > > _______________________________________________ > Freesurfer mailing list > Freesurfer(a)nmr.mgh.harvard.edu > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer > > > _______________________________________________ Freesurfer mailing list Freesurfer(a)nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

1 0

center of pial surface in world coordinates
by martijn van den heuvel 07 Jun '06

07 Jun '06

Dear freesurfer-users, I want to plot the freesurfer surface rendering (xxx.pial) in the program Neurolens. Loading the image gives no problem and Neurolens gives a beautiful rendering. However, when loading the pial surface in tksurfer I see a difference between the two renderings. The surface in tksurfer is centered to RAS (0,0,0) , like the T1.mgz image (as it supposed to do). Im trying to figure out what this translation is, but I cannot find the answer on the web, hence this email. Does tksurfer loads in a translation file (except for the talaraich.xfm, I got this translation covered), or does it do a general shift to the RAS centre? Generally, Im looking for the translation matrix from RAS to world coordinates, so I can apply this translation to the surface before loading it in Neurolens. I hope someone can help me with this question. Yours sincerely, Martijn van den Heuvel

2 1

FSL output overlay
by Vishwadeep Ahluwalia 07 Jun '06

07 Jun '06

Hi, I'm a relatively new FS user. Ive a single subject functional dataset analyzed in FSL. I ran recon-all autorecon1 ,2 and 3 on the subject's SPGR. Everything worked fine. i ran reg-feat2anat to register my functionals to the anatomical but when i hit compare button in tkregister2 , the alignment is all wrong. I tried changing the dof to 12 and maxangle to 180 but the output again looks the same. the functionals are almost inverted compared to the outline of the anatomical. During the processing i saw a warning saying that the 'neither sform or qform in the nifti-1 are valid 'and another warning saying the ' your volume will probably be incorrectly oriented'. which is evident in the output. I thought FSL and FS both use FLIRT for registration. so why are the results different? Please advise. Regards, Vishwadeep A. Grad. Student Medical Physics PhD Program VCU

1 0

make_average_subject memory issues
by Lars M. Rimol 07 Jun '06

07 Jun '06

Hi, I just ran make_average_subjects with increased swap (from 2G to 8G) on 178 subjects and now it worked, which would seem to indicate that it may have been a memory issue? I did not crash with 3, 10 and 113 subjects with 2G swap. yours, LMR >Hmmm, I've run it on >100 subjects without trouble. I'll try now and see if >I can replicate locally. Is there anything different about these volumes? >Are they all 256^3 and unsigned char? > >cheers, >Bruce >>Could be a memory leak. Can you send us all the text output? We need to know specifically what program it is dying on. >>doug >>>Hi, >>> I am analysing a dataset with 180 subjects. When I run make_average_subjects it crashes, the error message usually >>>being something like this: >>> >>> mghRead: encountered error executing: 'zcat /space/monkeys/1/home/lmr >>>/subjects/VETSA_test/testAVG/tmp/make_average_vol-tmp/orig-19963b.mgz',frame -1 >>> >>> I have tried running it with fewer subjects and it worked with 3, 10 and 73 subjects. Also, it crashes at different points in >>>the processing stream, so once it crashed at subject # 130, which I removed from the list of subjects, and then next time >>>it crashed at # 30. My frist guess was that it is a memory problem. I tried running the top command in a different >>>terminal at the same time as running make_average_subject with 2 and 10 subjects, and it seemed to require up to >>>400m of virtual memory and 10% of memory, but I couldn't see any difference in memory demands between running 2 >>>and 10 subjects. This was running with 4G of physical memory and 2G of swap. >>> >>> Is this a known issue? (Couldn't find anything in the mailing list archive..) >>> >>> Thank you!

2 1

Re: [Freesurfer] errors in library in new version of Freesurfer
by Joongnam Yang 06 Jun '06

06 Jun '06

Thanks for the question. Here is the info that I can provide. (1) When I did 'source' the *.csh, there was no error. here is the output -------- freesurfer-Linux-centos4_x86_64-stable-pub-v3.0.3 -------- Setting up environment for FreeSurfer/FS-FAST (and FSL) FREESURFER_HOME /usr/local/freesurfer FSFAST_HOME /usr/local/freesurfer/fsfast SUBJECTS_DIR /usr/local/freesurfer/subjects FUNCTIONALS_DIR /usr/local/freesurfer/sessions MINC_BIN_DIR /usr/local/freesurfer/mni/bin MINC_LIB_DIR /usr/local/freesurfer/mni/lib PERL5LIB /usr/local/freesurfer/mni/lib/perl5/5.8.5 GSL_DIR /usr/local/freesurfer/lib/gsl QTDIR /usr/lib/qt3 TCLLIBPATH /usr/local/freesurfer/lib/tcltktixblt/lib FSL_DIR /usr/local/freesurfer/fsl (2) when I did, ldd 'which mri_convert', it gave me the following error message. ldd: ./which mri_convert: No such file or directory (3) When I did, however, just 'which mri_convert', it correctly points to its location, which is .........../freesurfer/bin/mri_convert. (4) Freesurfer is running on the latest version of SUSE on a 64-bit machine. I downloaded an appropriate version of Freesurfer. Nam. >>> Nick Schmansky <nicks(a)nmr.mgh.harvard.edu> 06/06/06 4:42 PM >>> What is the output when this is typed: source $FREESURFER_HOME/SetUpFreeSurfer.csh and this: ldd `which mri_convert` and what OS is this installed on? On Tue, 2006-06-06 at 15:24 -0400, Joongnam Yang wrote: > I installed the new version of Freesurfer by downloading from Freesurfer > Wiki website. > > Then, I ran recon-all -subjid subjectname -autorecon1, > and got the following library errors. > Would someone suggest a solution? > > Nam. > > mri_convert: error while loading shared libraries: libgsl.so.0: cannot open > shar ed object file: No such file or directory > mri_convert: error while loading shared libraries: libgsl.so.0: cannot open > shar ed object file: No such file or directory > strings: 'mritotal:': No such file > strings: 'Command': No such file > strings: 'not': No such file > strings: 'found.': No such file > nu_correct: Command not found. > ................. (and much more....) > _______________________________________________ > Freesurfer mailing list > Freesurfer(a)nmr.mgh.harvard.edu > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer > >

1 0

errors in library in new version of Freesurfer
by Joongnam Yang 06 Jun '06

06 Jun '06

I installed the new version of Freesurfer by downloading from Freesurfer Wiki website. Then, I ran recon-all -subjid subjectname -autorecon1, and got the following library errors. Would someone suggest a solution? Nam. mri_convert: error while loading shared libraries: libgsl.so.0: cannot open shar ed object file: No such file or directory mri_convert: error while loading shared libraries: libgsl.so.0: cannot open shar ed object file: No such file or directory strings: 'mritotal:': No such file strings: 'Command': No such file strings: 'not': No such file strings: 'found.': No such file nu_correct: Command not found. ................. (and much more....)

2 1

Re: freesurfer
by Nurunisa Neyzi 06 Jun '06

06 Jun '06

Hello, > I have whole brain T1 maps and the pseudo T1-weighted images that I > created out of them (so they are perfectly aligned). The are all brain > extracted using fsl and > I am doing my own tissue segmentation (again, using fsl), but in both > steps there is much optimization involved, because I want to do it very > carefully avoiding partial volume effects (I am looking at the cortical T1 > values). > I would like to look at the maps in the inflated and flattened spaces. > So I have been learning freesurfer over the past week but my problem is > with the surface reconstruction and that everything is automated. I don't > want freesurfer to do the segmentation for me. I was hoping it would work > well because the data is 1x1x1- but the segmentation is not nearly as > precise as I need it. > Is there a way that I give my already aligned T1 maps and structural > images and tissue segmentation and get out the surfaces with T1 values on > top of it? > > > >

2 1

mri_volcluster example
by Paymon Hosseini 06 Jun '06

06 Jun '06

Hi Can you please post an example of "mri_volcluster" use with fsfast? more specifically, when i run : mri_volcluster \ --in /space/engram/12/users/pman/rk_fmri/Y15_111504/bold/rk_enc_sm5/RNeg_vs_Fix/sig \ --in_type bfloat \ --reg /space/engram/12/users/pman/rk_fmri/Y15_111504/bold/rk_enc_sm5/tal/register.dat \ --thmin 7 \ --minsize 640 \ --allowdiag i get the following error: NOT Adjusting threshold for 1-tailed test INFO: bvolumeRead: min = -16.6743, max = 25.5035 ERROR: Input volume voxel dimensions do not match those in the registration file. If the input volume is in bshort/bfloat format, check that there is an accompanying bhdr file. there are sig_xxx.bfloat and sig_xxx.hdr files in my "/space/engram/12/users/pman/rk_fmri/Y15_111504/bold/rk_enc_sm5/RNeg_vs_Fix/" directory. i am using "nmrenv" on engram.nmr.mgh any advice on how to resolve this? thanks a lot, paymon

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