Freesurfer February 2007

freesurfer@nmr.mgh.harvard.edu

59 participants
80 discussions

Difficulties with segmentation follow-up
by Jeffrey Spielberg 11 May '07

11 May '07

Hi, I sent a message to the forum last June regarding a problem with segmentation of the acc we were running into. We kept having areas of the rostral and caudal acc labeled as corpus callosum or "unlabelled subcortical". I got a reply that this was being worked on and I was wondering if this is now fixed in 3.0.4, and if not if there is any timeline for releasing a fix? Thanks, Jeff

2 3

Legacy flag
by Fornito, Alexander 19 Mar '07

19 Mar '07

Hi, I was just seeking some clarification re: the legacy flag in recon-all. Can I use it on subjects procssed with older versions of Freesurfer to reconstruct surfaces that will be consistent with those generated with older versions (i.e., I can collapse data from the two versions), or is the legacy flag intended to allow new versions of recon-all to be used on old data, effectively 'updating' the surfaces (meaning they should not be analyzed together with data generated using older versions). Thanks, Alex Alex Fornito JN Peters Research Fellow Melbourne Neuropsychiatry Centre and Department of Psychology National Neuroscience Facility The University of Melbourne Levels 2 & 3, Alan Gilbert Building 161 Barry St Carlton South Vic 3053 Australia Ph: +61 3 8344 1861 Fax: +61 3 9348 0469 email: alexander.fornito(a)wh.org.au

4 6

preproc inorm question
by Cameron Ellis 02 Mar '07

02 Mar '07

Hi all, If I were to perform a surface-based analysis, resampling to spherical space, and I did not want to smooth in the volume, would I need to preprocess my data again without smoothing or could I just specify the motion corrected volume in in a new analysis. I think the answer is: I could just specify the motion corrected volume, however, since the global mean for intensity normalization is calculated at the end of preprocessing after smoothing and exists in the fmcsm.meanval file, will there be a problem with the inorm in later stages, will it be performed etc. if I just specify the fmc volume in a new analysis? Thanks, Cameron

2 3

annotations & functional activation
by oxro03＠yahoo.com 01 Mar '07

01 Mar '07

Hi Freesurfers, I have functional activation files (.mgh) which I can overlay onto a subject's surface. Now I am trying to do a functional ROI analysis based on the annotation files, e.g. to get information about the mean activation in a specific area based on FS segmentations. Does anybody have any suggestion on how to do this in a straightforward manner? Thanks a lot. Sandy --------------------------------- Any questions? Get answers on any topic at Yahoo! Answers. Try it now.

2 2

bus error during group analysis
by Leonardo Bonilha 28 Feb '07

28 Feb '07

Hi, I am getting a bus error during a group analysis and I wondered if someone could give me some help. I am running FreSurfer 3.0.5 on an intel Mac (MacBookPro 2.33Ghz, 3Gb, MacOSX 10.4.8). While going through the steps from the web tutorial (with my own data) I am getting the following message: After the command: mri_glmfit --y lh.age.thickness.10.mgh --fsgd age_fsgd.txt doss --glmdir lh.age.glmdir --surf average lh --C age.mat gdfReadHeader: reading age_fsgd.txt INFO: gd2mtx_method is doss Reading source surface /Applications/freesurfer//sub_aging//average/surf/lh.white reading group avg surface area 782 cm^2 from file Reading in average area /Applications/freesurfer//sub_aging//average/surf/lh.white.avg.area.mgh Number of vertices 163842 Number of faces 327680 Total area 67091.234375 AvgVtxArea 0.409487 AvgVtxDist 0.738119 StdVtxDist 0.221868 $Id: mri_glmfit.c,v 1.78.2.7 2007/01/23 21:37:47 greve Exp $ cwd /Applications/freesurfer/sub_aging/stats cmdline mri_glmfit --y lh.age.thickness.10.mgh --fsgd age_fsgd.txt doss --glmdir lh.age.glmdir --surf average lh --C age.mat sysname Darwin hostname Macintosh.local machine i386 user leonardobonilha FixVertexAreaFlag = 1 OneSampleGroupMean 0 y /Applications/freesurfer/sub_aging/stats/lh.age.thickness.10.mgh FSGD age_fsgd.txt glmdir lh.age.glmdir Creating output directory lh.age.glmdir Bus error Interestingly, I ran the same analysis (with the same data) on another machine (this time a PC running Ubuntu) and it went fine, without any problems. Thanks for the support Leo Leonardo Bonilha MD PhD Research Assistant Professor Dep of Neuropsychiatry and Communication Sciences and Disorders University of South Carolina ____________________________________________________________________________________ TV dinner still cooling? Check out "Tonight's Picks" on Yahoo! TV. http://tv.yahoo.com/

2 5

error testing freesurfer installation
by Mishkin DERAKHSHAN 28 Feb '07

28 Feb '07

Hi, I just installed the rh9 version of fressurfer. I am running Ubuntu 6.10. I had the libtiff.so.3 error, which I seemed to have fixed by just making a sym link to libtiff.so.4 ( I hope that is okay). Now I get this error about the aseg.mgz.gz not existing. It looks like it isn't looking in the mri/ dir as it should be. mishkin@thinkpad:/usr/local/freesurfer$ tkmedit bert norm.mgz -segmentation aseg.mgz $FREESURFER_HOME/FreeSurferColorLUT.txt libGL warning: 3D driver claims to not support visual 0x5b Talairach transform /autofs/space/freesurfer/subjects/bert/mri/transforms/talairach.xfm does not exist ... Loading tal xfm file /usr/local/freesurfer/subjects/bert/mri/transforms/talairach.xfm zcat: /usr/local/freesurfer/subjects/bert//aseg.mgz.gz: No such file or directory mghRead(/usr/local/freesurfer/subjects/bert//aseg.mgz, -1): read error Reading /usr/local/freesurfer/lib/tcl/tkm_common.tcl Reading /usr/local/freesurfer/lib/tcl/tkm_wrappers.tcl These are the only things i've added to my .bashrc : export FREESURFER_HOME=/usr/local/freesurfer source $FREESURFER_HOME/SetUpFreeSurfer.sh which produces this output: -------- freesurfer-Linux-rh9-stable-pub-v3.0.5 -------- 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 MNI_DIR /usr/local/freesurfer/mni GSL_DIR /usr/local/freesurfer/lib/gsl FSL_DIR /usr/local/freesurfer/fsl Did I forget to set a variable somewhere? thanks, mishkin

2 1

Output of mri_volcluster
by Dan Dillon 28 Feb '07

28 Feb '07

Dear FreeSurfers, I'm trying to replicate some functional analyses that were originally done in SPM2, and am using mri_volcluster. Here's my command: mri_volcluster --in Old8/bold/taumax30_2_21/tal-ffx/RSF_v_NIF/sig --in_type bfloat --reg Old8/bold/taumax30_2_21/tal-ffx/register.dat --thmin 3 --minsizevox 8. I've noticed that the size (in voxels) and number of clusters changes if I play with the -sign flag. I am confused by this. Here's the output not using the -sign flag: Cluster Size(n) Size(mm^3) TalX TalY TalZ Max 1 28 1792.0 -19.80 -96.70 8.53 5.28907 2 8 512.0 39.60 26.79 -8.07 4.15380 3 9 576.0 99.00 16.69 -54.67 4.09661 4 10 640.0 87.12 63.11 18.95 -3.82155 5 13 832.0 67.32 -42.44 5.81 3.79067 6 11 704.0 -39.60 -16.43 59.78 -3.62503 7 11 704.0 -87.12 -75.98 -43.30 -3.59305 8 73 4672.0 43.56 -22.33 19.54 -3.43893 9 18 1152.0 43.56 -44.81 -41.49 -3.16065 10 8 512.0 -51.48 -32.30 53.20 -3.09981 11 15 960.0 -63.36 -20.49 56.30 -3.00613 12 48 3072.0 -71.28 -7.20 11.41 -3.00547 And here's the output with sign set to positive: Cluster Size(n) Size(mm^3) TalX TalY TalZ Max 1 35 2240.0 -19.80 -96.70 8.53 5.28907 2 21 1344.0 39.60 26.79 -8.07 4.15380 3 13 832.0 99.00 16.69 -54.67 4.09661 4 10 640.0 -71.28 33.20 -35.31 3.91335 5 21 1344.0 67.32 -42.44 5.81 3.79067 6 9 576.0 35.64 -88.58 15.49 3.63267 Why does the size of the first cluster change? Also, where is cluster 4 from the second output in the first output? Is there some re-calculation going on here? Dan Dillon, Ph.D. Post-doctoral Fellow Affective Neuroscience Lab Dept. of Psychology, Harvard University Phone: (617) 496-5222 E-mail: dillon(a)wjh.harvard.edu

2 1

aparc -> voxels program?
by Graham Wideman 28 Feb '07

28 Feb '07

Folks: I'm hoping that there's an aparc --> voxels program. That program might be mri_aparc2aseg, but I wanted to check first if that's the best way to go and whether there are particular options to know about and so on. Basically I want a program that inputs a parcellation, and outputs a label mri where each voxel that's fully or partially between the white and pial surface gets labelled with the parcellation label from the vertex (white or pial) that it's closest to. (A variation that would also work for our purpose would label voxels even beyond that, possibly all voxels). This seems like it's probably already been done, hence my asking. I'm thinking that *might* be what mri_aparc2aseg does... but I'm not clear what the wiki description is trying to say: https://surfer.nmr.mgh.harvard.edu/fswiki/mri_5faparc2aseg?action=highlight… First, it appears from Example 1 that mri_aparc2aseg does something interesting when run with basically no specific command line arguments, so presumably using default input files. But then what is the purpose of --ribbon? Then the description talks about the automatic segmentation volume as though it's an input separate from the ?h.ribbon, yet there are no parameters related to that. And finally the --noribbon flag is mentioned as turning something off, but it's not clear what it's turning off. Bottom line: I'd appreciate some clarification as to which program does what I want, and if it's mri_aparc2aseg, then what options are appropriate to do the simple version of the task where it ignores existing aseg and ribbon. Thanks, Graham

2 3

Re: Files uploaded for you on Martinos Center FileDrop site
by Anil Roy 28 Feb '07

28 Feb '07

Yes, if I just run the raw files, then the brain crashes so I don't get any output. But the background values are high in the orig.mgz file also and the xfm is not accurate. afni however is able to process all these raw files without a problem including the talairach and skull strip. Is there some flag or something that I have to use or is there no way to fix this. Thanks, Anil. On 2/28/07, Bruce Fischl <fischl(a)nmr.mgh.harvard.edu> wrote: > > hmmm, the high background is there in the 001.mgz also, so it's not an > afni thing. Do you see this in all your images? > > On Tue, 20 Feb 2007 > aroy85(a)gmail.com wrote: > > > > > The following files have been uploaded to the Martinos Center FileDrop > site > > by Anil Roy for you to download: > > > > (1) 001.mgz > > > > > > > > These files will expire in 30 days. > > > > To view or download these files, go in your web browser to: > > > > > https://www.nmr.mgh.harvard.edu/facility/filedrop/showgroup/5075/89/4def6e0… > > > > DO NOT SHARE this web address with anyone else. It is your personal > access key to the files. > > > > Overall Comment about files from Anil Roy: > > > > Bruce, Here is the 001.mgz > > > > > > >

2 1

Re: [Freesurfer] Re: Files uploaded for you on Martinos Center FileDrop site
by Anil Roy 27 Feb '07

27 Feb '07

Hi Bruce, I've been trying to improve the talairach xfm and try different things with afni, but so far nothing has worked out. The brain still crashes at the skull strip. I fed in the afni skull strip to resume the pipeline from autorecon2 but the segmentation was really bad, with the white matter surface not accurate. Did you have any suggestions for the background values. Any ideas. Thanks, Anil. On 2/21/07, Bruce Fischl <fischl(a)nmr.mgh.harvard.edu> wrote: > > Hi Anil, > > I got the email, but I'm out of town and won't get to it until next week > Bruce > On > Wed, 21 Feb 2007, Anil Roy wrote: > > > I think my last email didn't get through. Bruce, I've uploaded the > 001.mgz. > > It should arrive in your mailbox. I can upload it again if there was any > > problem. > > > > Thanks, > > > > Anil. > > > > On 2/20/07, Anil Roy <aroy85(a)gmail.com> wrote: > >> > >> Its uploaded > >> > >> On 2/20/07, Bruce Fischl <fischl(a)nmr.mgh.harvard.edu> wrote: > >> > > >> > I need the 001.mgz > >> > On Tue, 20 Feb 2007, Anil Roy wrote: > >> > > >> > > Bruce, I had uploaded the orig.mgz for Nick. I think you should > find > >> > the > >> > > link below. I can also upload the 001.mgz for you. > >> > > There are no dicoms because the scan is from a GE scanner (120 > slices, > >> > > >> > > details are below)and all the files are named as I.??? > >> > > I could also upload these files for you. > >> > > > >> > > Anil. > >> > > > >> > > On 2/20/07, Bruce Fischl < fischl(a)nmr.mgh.harvard.edu > wrote: > >> > >> > >> > >> I think it's something afni is doing, but I'm not sure. Did you > give > >> > us > >> > >> the data before conversion (either the dicoms or the 001.mgz) > >> > >> > >> > >> On Tue, 20 Feb 2007, > >> > >> Anil Roy wrote: > >> > >> > >> > >> > Hey Freesurfers, > >> > >> > > >> > >> > I fed freesurfer the brainmask.mgz generated by AFNI. autorecon2 > >> > and > >> > >> > autorecon3 ran fine with no errors but the resulting surfaces > are > >> > not > >> > >> very > >> > >> > accurate. > >> > >> > > >> > >> > Bruce, you mentioned that my background values were too high. Is > >> > there a > >> > >> way > >> > >> > to reduce them so that freesurfer can do the watershed without > any > >> > >> glitches. > >> > >> > > >> > >> > Thanks, > >> > >> > > >> > >> > Anil. > >> > >> > > >> > >> > On 2/15/07, Bruce Fischl <fischl(a)nmr.mgh.harvard.edu> wrote: > >> > >> >> > >> > >> >> I think it's because your background values are so high. > >> > >> >> On Thu, 15 Feb > >> > >> >> 2007, Anil Roy wrote: > >> > >> >> > >> > >> >> > Thanks a lot Nick. > >> > >> >> > > >> > >> >> > I'll try to convert the AFNI skull strip into brainmask and > nu > >> > and > >> > >> see > >> > >> >> if I > >> > >> >> > can resume the pipeline. > >> > >> >> > > >> > >> >> > I put in the new values for the xfm and ran recon-all again > with > >> > the > >> > >> >> > notalairach flag. > >> > >> >> > > >> > >> >> > recon-all -s Nl-TAl-001 -notalairach -all > >> > >> >> > > >> > >> >> > However, it still quits at the watershed step. I ran another > >> > brain > >> > >> where > >> > >> >> the > >> > >> >> > raw images were dcm and it ran fine so the problem is > definitely > >> > >> >> something > >> > >> >> > specific to this brain. > >> > >> >> > > >> > >> >> > Nick, I know that autorecon1 ran fine for you so I really > have > >> > no > >> > >> clue > >> > >> >> why > >> > >> >> > my machine quits at the watershed step even with the > transform > >> > you > >> > >> gave > >> > >> >> me. > >> > >> >> > > >> > >> >> > Any clues, suggestions? > >> > >> >> > > >> > >> >> > Thanks a lot, > >> > >> >> > > >> > >> >> > Anil. > >> > >> >> > > >> > >> >> > On 2/14/07, Nick Schmansky < nicks(a)nmr.mgh.harvard.edu> > wrote: > >> > >> >> >> > >> > >> >> >> Anil, > >> > >> >> >> > >> > >> >> >> Here is the transform (the contents of the talairach.xfmfile) > >> > that > >> > >> I > >> > >> >> >> generated: > >> > >> >> >> > >> > >> >> >> Transform_Type = Linear; > >> > >> >> >> Linear_Transform = > >> > >> >> >> 9.265516e-01 5.678305e-04 2.317888e-03 3.736965e-01 > >> > >> >> >> -8.173273e-04 8.513330e-01 8.824052e-02 -1.653205e+01 > >> > >> >> >> - 2.688466e-03 -1.064786e-01 9.460568e-01 1.780552e+01 ; > >> > >> >> >> > >> > >> >> >> Remember to use the -notalairach flag with recon-all when > you > >> > create > >> > >> > >> > >> >> the > >> > >> >> >> new talairach.xfm file, so that it doesnt get overwritten. > >> > >> >> >> > >> > >> >> >> 4GB is plenty to run Freesurfer, so it is strange that > >> > mri_watershed > >> > >> is > >> > >> >> >> encountering out-of-memory errors. > >> > >> >> >> > >> > >> >> >> Nick > >> > >> >> >> > >> > >> >> >> > >> > >> >> >> On Wed, 2007-02-14 at 17:20 -0600, Anil Roy wrote: > >> > >> >> >> > Also, I ran it on a Red Hat system and I'm still getting > the > >> > error > >> > >> so > >> > >> >> >> > I don't think it is a machine specific problem. Any > >> > suggestions. > >> > >> >> >> > > >> > >> >> >> > Thanks. > >> > >> >> >> > Anil. > >> > >> >> >> > > >> > >> >> >> > On 2/14/07, Anil Roy < aroy85(a)gmail.com> wrote: > >> > >> >> >> > Here are the details of the acquisitions: > >> > >> >> >> > > >> > >> >> >> > Images acquired on a 3T Signa scanner with a > standard > >> > GE > >> > >> head > >> > >> >> >> > coil. 3D T1-weighted spiral volume scan (120 > slices, > >> > >> >> 1.4875-mm > >> > >> >> >> > thick, TR = 20 msec, TE = 6 msec, 40 degrees flip > >> > angle, > >> > >> 256 > >> > >> >> * > >> > >> >> >> > 192 matrix) > >> > >> >> >> > > >> > >> >> >> > The machine has 4 gigs of RAM. Also, AFNI is able > to > >> > >> >> talairach > >> > >> >> >> > and skull strip. Is there a way I can input the > files > >> > from > >> > >> > >> > >> >> >> > AFNI to Freesurfer to resume the recon-all > pipeline. > >> > >> >> >> > > >> > >> >> >> > Nick, is there a way for me to see your reg so > that I > >> > can > >> > >> be > >> > >> >> >> > sure that its not my created xfm that is causing > >> > this > >> > >> >> >> > problem. > >> > >> >> >> > > >> > >> >> >> > Thanks, > >> > >> >> >> > > >> > >> >> >> > Anil. > >> > >> >> >> > > >> > >> >> >> > > >> > >> >> >> > On 2/14/07, Bruce Fischl < > fischl(a)nmr.mgh.harvard.edu> > >> > > >> > >> wrote: > >> > >> >> >> > why is the background so high? Also, what > >> > field > >> > >> >> >> > strength was this at? There > >> > >> >> >> > looks to be dielectric-like center > >> > brightening. > >> > >> The > >> > >> >> >> > contrast isn't terrible > >> > >> >> >> > though. > >> > >> >> >> > > >> > >> >> >> > On Wed, 14 Feb 2007, Nick > >> > >> >> >> > Schmansky wrote: > >> > >> >> >> > > >> > >> >> >> > > Bruce, > >> > >> >> >> > > > >> > >> >> >> > > the orig.mgz can be downloaded via the > >> > filedrop, > >> > >> >> see > >> > >> >> >> > link in email > >> > >> >> >> > > below. > >> > >> >> >> > > > >> > >> >> >> > > or i have copied it to: > >> > >> >> >> > > > >> > >> >> >> > > >> > >> >> >> > /autofs/space/minerva_001/users/nicks/subjects/anil/mri > >> > >> >> >> > > > >> > >> >> >> > > N. > >> > >> >> >> > > > >> > >> >> >> > > > >> > >> >> >> > > On Wed, 2007-02-14 at 17:02 -0500, Bruce > >> > Fischl > >> > >> >> >> > wrote: > >> > >> >> >> > >> put the .mgz somewhere and I'll take a > >> > look. > >> > >> And > >> > >> >> >> > make sure it's the > >> > >> >> >> > >> orig.mgz, which is un-normalized > >> > >> >> >> > >> > >> > >> >> >> > >> > >> > >> >> >> > >> On Wed, 14 Feb 2007, Nick Schmansky > wrote: > >> > > >> > >> >> >> > >> > >> > >> >> >> > >>> Anil, > >> > >> >> >> > >>> > >> > >> >> >> > >>> The image has almost no > contrast. I've > >> > >> attached > >> > >> >> a > >> > >> >> >> > TIFF of one slice. > >> > >> >> >> > >>> How was this acquired? Bruce, any > idea > >> > why it > >> > >> > >> > >> >> >> > would be so white? > >> > >> >> >> > >>> > >> > >> >> >> > >>> Nick > >> > >> >> >> > >>> > >> > >> >> >> > >>> > >> > >> >> >> > >>> On Wed, 2007-02-14 at 16:10 -0500, > >> > >> >> >> > aroy85(a)gmail.com wrote: > >> > >> >> >> > >>>> The following files have been > uploaded > >> > to the > >> > >> >> >> > Martinos Center FileDrop site > >> > >> >> >> > >>>> by Anil Roy for you to download: > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> (1) orig.mgz > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> These files will expire in 30 days. > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> To view or download these files, go > in > >> > your > >> > >> web > >> > >> >> >> > browser to: > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> > >> > >> >> >> > > >> > >> >> >> > >> > >> >> > >> > >> > >> > https://www.nmr.mgh.harvard.edu/facility/filedrop/showgroup/5010/426/7389dc… > >> > > >> > >> > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> DO NOT SHARE this web address with > >> > anyone > >> > >> >> >> > else. It is your personal access key to > the > >> > >> files. > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> Overall Comment about files from Anil > >> > Roy: > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> Hi, > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> Attached is the orig.mgz file for the > >> > subject > >> > >> > >> > >> >> >> > that quits at the watershed step. > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> Thanks, > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> Anil. > >> > >> >> >> > >>>> > >> > >> >> >> > >>>> > >> > >> >> >> > >>> > >> > >> >> >> > >> > >> > >> >> >> > >> > >> > >> >> >> > > > >> > >> >> >> > > > >> > >> >> >> > > > >> > >> >> >> > > >> > >> >> >> > > >> > >> >> >> > > >> > >> >> >> > _______________________________________________ > >> > >> >> >> > Freesurfer mailing list > >> > >> >> >> > Freesurfer(a)nmr.mgh.harvard.edu > >> > >> >> >> > > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer > >> > >> >> >> > >> > >> >> >> > >> > >> >> > > >> > >> >> > >> > >> > > >> > >> > >> > > > >> > > >> > >> > > >

2 2

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Freesurfer February 2007