Freesurfer

freesurfer@nmr.mgh.harvard.edu

19 participants
30088 discussions

gamma.mgh
by Rees, Elin 16 Oct '13

16 Oct '13

Sorry, I mean the gamma.mgh maps for the cross-sectional differences appear to be blank as no overlay is overlaid on the fsaverage when I do the same command as with the longitudinal gamma.mgh. The commands I am running for the cross-sectional group comparison are: mri_glmfit --glmdir lh_groupdiff.dir --fsgd fsgd.fsgd doss --y lh.thickness10.mgh --C groupdiff.mtx --surf fsaverage lh --cortex tksurfer fsaverage lh inflated -overlay /lh_groupdiff.dir/groupdiff/gamma.mgh Thanks Message: 8 Date: Wed, 16 Oct 2013 11:47:59 -0400 From: Douglas N Greve <greve(a)nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] gamma.mgh To: freesurfer(a)nmr.mgh.harvard.edu Message-ID: <525EB52F.5000006(a)nmr.mgh.harvard.edu> Content-Type: text/plain; charset=ISO-8859-1; format=flowed what do you mean they are "coming up blank" On 10/16/2013 05:15 AM, Rees, Elin wrote: > Dear FreeSurfers, > > I've been looking at between-group (controls and disease) differences > in cortical thickness cross-sectionally and longitudinally (PC1). I've > outputted sig.mgh maps and effect size maps (gamma.mgh) but the > cross-sectional gamma.mgh maps are coming up blank. Could you explain > why this is (and why it works for the longitudinal data) and how I can > create a cross-sectional effect size map? > > As the scans I am an analysing are from four different scanners I am > also looking at scanner differences in the controls only. No > significant differences are coming up but the effect size maps are > showing both positive and negative differences across the whole brain > and the gammavar.mgh maps are completely yellow (5 on the colour bar). > How do I interpret this? > > Any help would be much appreciated. Thanks, > Elin > > > > > > > _______________________________________________ > Freesurfer mailing list > Freesurfer(a)nmr.mgh.harvard.edu > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer -- Douglas N. Greve, Ph.D. MGH-NMR Center greve(a)nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422 Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 www.nmr.mgh.harvard.edu/facility/filedrop/index.html Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ ------------------------------ Message: 9 Date: Wed, 16 Oct 2013 11:51:41 -0400 From: Douglas N Greve <greve(a)nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] DOSS FS 5.3 To: amirhossein manzouri <a.h.manzouri(a)gmail.com> Cc: free surfer <freesurfer(a)nmr.mgh.harvard.edu> Message-ID: <525EB60D.1010808(a)nmr.mgh.harvard.edu> Content-Type: text/plain; charset=ISO-8859-1; format=flowed From the command line you will need to run mri_glmfit. The mri_glmfit.log file in the QDEC output folder will have a sample mri_glmfit command line. You will need to modify this by adding "doss" after the FSGD file. You will also need to create new contrast matrices to match the DOSS model. Search for fsgd on our wiki to get examples doug On 10/16/2013 03:19 AM, amirhossein manzouri wrote: > Dear Experts, > Hi, > I have already done my data analysis having group as fixed factor and > age as nuisance factor in QDEC FS 5.3. I also need to run it with > DOSS. Would you please advise how to run the same analysis with > DOSS from terminal and if it is possible to look at the results in > QDEC Display again. > > -- > Best regards, > Amirhossein Manzouri > > -- Douglas N. Greve, Ph.D. MGH-NMR Center greve(a)nmr.mgh.harvard.edu Phone Number: 617-724-2358 Fax: 617-726-7422 Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting FileDrop: https://gate.nmr.mgh.harvard.edu/filedrop2 www.nmr.mgh.harvard.edu/facility/filedrop/index.html Outgoing: ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/ ------------------------------ _______________________________________________ Freesurfer mailing list Freesurfer(a)nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer End of Freesurfer Digest, Vol 116, Issue 29 *******************************************

2 1

Confusion about dods
by Jürgen Hänggi 16 Oct '13

16 Oct '13

Dear FS experts I compared two groups wrt cortical thickness using total grey matter volume as a covariate. The results obtained without the covariate (first row in the attached figure) and with the covariate modeled across both groups (doss option, second row in the figure) differed only in quantity (clusters are smaller with the covariate). However, when using the covariate with dods, the results changed completely (third row in figure) compared to doss and compared without the covariate. Then I tested (using dods) the covariates itself (using 0 0 1 -1) and get almost identical clusters (although in the opposite direction, forth row in the figure) as found in the group comparison with the covariate (dods, third row). In my opinion, this is puzzling! Any idea? Thanks in advance Regards Jürgen ---------------------------------------------------------------------------- Jürgen Hänggi, Ph.D. Division Neuropsychology Institute of Psychology University of Zurich Binzmuehlestrasse 14, PO Box 25 8050 Zurich, Switzerland 0041 44 635 73 97 (phone office) 0041 76 445 86 84 (phone mobile) 0041 44 635 74 09 (fax office) BIN 4.D.04 (office room number) j.haenggi[at]psychologie.uzh.ch (email) http://www.psychologie.uzh.ch/neuropsy/ (website) http://www.juergenhaenggi.ch (private website) This e-mail (and any attachment/s) contains confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorised copying, disclosure or distribution of the material in this e-mail is strictly forbidden. ----------------------------------------------------------------------------

1 0

gamma.mgh
by Rees, Elin 16 Oct '13

16 Oct '13

Dear FreeSurfers, I've been looking at between-group (controls and disease) differences in cortical thickness cross-sectionally and longitudinally (PC1). I've outputted sig.mgh maps and effect size maps (gamma.mgh) but the cross-sectional gamma.mgh maps are coming up blank. Could you explain why this is (and why it works for the longitudinal data) and how I can create a cross-sectional effect size map? As the scans I am an analysing are from four different scanners I am also looking at scanner differences in the controls only. No significant differences are coming up but the effect size maps are showing both positive and negative differences across the whole brain and the gammavar.mgh maps are completely yellow (5 on the colour bar). How do I interpret this? Any help would be much appreciated. Thanks, Elin

2 1

LGI radius
by Jon Wieser 15 Oct '13

15 Oct '13

Hi freesurfer experts is it possbile to change the radius for the LGI computation. ( it's normally set at 25 mm) we waned to try computing the LGI with a smaller radius. I looked the in the "compute_lgi.m" file, and it does not use radius, (even though radius is an input parameter for it) Thanks Jon -- Jon Wieser Research Specialist UW-Milwaukee Psychology Department, Pearse Hall Rm 366 2441 East Hartford Ave Milwaukee, WI 53211 Phone: 414-229-7145 Fax: 414-229-5219

2 3

Fwd: talairach registration troubles
by Robyn Migliorini 15 Oct '13

15 Oct '13

Dear freesurfer folks, My colleagues and I have been running into some troubles with our talairach registration. These are the steps that we run: 1. cd SUBJECT_DIR: 2. recon-all -all -subjid $SUBJECT 3. cd $SUBJECT/mri 4. mri_nu_correct.mni --i orig.mgz --o orig.mgz --n 3 5. recon-all -s $SUBJECT -talairach -clean-tal -tal-check 6. cd ../../ 7. recon-all -all -subjid $SUBJECT recon -all -version yields this: $Id: recon-all,v 1.379.2.17 2011/05/20 22:48:18 nicks Exp $ Then we have been checking our talairach registration by following these instructions: https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/Talairach When we go to visually examine the talairach transformed volume (*file --> load transform--> load transform for main volume --> talairach.xfm*), we've found that many of the images look chunked, jagged, or striated, like pieces of a puzzle fit together poorly rather than a smooth whole brain (see attached images for examples). When we look at the talairach overlay using *tkregister2 --mgz --s subjectID --fstal --surf orig *on these same brains, they don't look particularly misaligned. We have tried doing some minor moving and stretching of the talairach volume to better fit our subject volume when it seems necessary, but when we compare the edited vs unedited talairach, the one we fixed still has the striation problem, though sometimes less so. What might be causing this? What might be some remedies? We are also particularly concerned because we know that the alignment of the corpus callosum is important for the talairach registration progress, and the population that we work with (children with fetal alcohol exposure) has know callosal abnormalities (e.g., some deviations in shape, placement, volume), although we are seeing problems even in ones that look perfectly aligned. Thank you very much for any advice you can offer! Best, Robyn Migliorini Center for Behavioral Teratology San Diego State University / University of California San Diego Joint Doctoral Program in Clinical Psychology

1 0

labels coordinates from atlases
by maria estanqueiro 15 Oct '13

15 Oct '13

Dear all, I need the center coordinates of the anatomical labels from the desikan and destrieux atlases. I saw a similar post a while ago in which it was recommended the following: #1 Convert the annotation to a set of labels. This is for the Desikan atlas, but you can spec the destrieux instead mri_annotation2label --subject fsaverage --hemi lh --outdir ~/tmp/aparc # For each label, run mri_surfcluster to get the centroid mri_surfcluster --in ~/subjects/fsaverage/surf/lh.thickness --clabel lh.pericalcarine.label --sum sum.lh.pericalcarine --centroid --thmin 0 --hemi lh --subject fsaverage I did the last step step for 2 different labels, one in the temporal and the other in the occipital cortex. However, the coordinates that were generated in the sum files were the same. Does anyone know why? Is this the right procedure for what I want to do? Thanks!

2 1

Qdec taking a very long time
by Mrk Drakesmith 15 Oct '13

15 Oct '13

Hi FreeSurfers I am quite new to freesurfer. I am trying to run a GLM analysis on some FreeSurfer data. I am analysing volume (surface based) with a smoothing of 5mm, using one discrete factor and one covriate on 252 subjects. I started the analysis on Friday 4th October. Its now been 8 days and the process is still running. The terminal is spitting out long columns of numbers so it looks like it is running, but I am supprised that it takes this long. Doing similar analysis in FSL and SPM takes a few minutes to about an hour. I am assuming therefore that something has gone wrong. If anyone can help me get to the bottom of this that woudl be very helpful. Many thanks Mark

3 3

Re: [Freesurfer] QA in Tracula
by Anastasia Yendiki 15 Oct '13

15 Oct '13

Can you please send a screenshot of the tract that you get (after rerunning with reinit=1) in the volume? Does it go off the white matter or the brain mask at any part? On Tue, 15 Oct 2013, Garikoitz Lerma-Usabiaga wrote: > Thanks Anastasia!I've tried to mend just the S_23 subject, just the rh.sflp (first > image). > > After the -prior and -path processes, I still get exactly the same tract. Should I > change another option? for example ncpts? > > The rest of the tracts look OK, it is failing just this one, it is not a registration > problem. > > Thanks again for your help! > Gari > > > On Mon, Oct 14, 2013 at 10:45 PM, Anastasia Yendiki <ayendiki(a)nmr.mgh.harvard.edu> > wrote: > > Hi Gari - What you're describing is something that may happen if the > initialization of the tract fails. You can repeat with a different > initialization by adding "set reinit = 1" and rerunning the -prior and > -path parts. You should do this only for the subject and tract that has > this issue, so set pathlist to that tract only. > > If a subject has several tracts that went wrong like that, it maybe that > something else is wrong, either with the aparc+aseg from the freesurfer > recon, or with its registration to DWI space. So I'd check on that. > > Since FA is not in units of length/surface/volume, there's no need to > normalize it with another measure in such units. The "weighted averages" > in the stats file are essentially the mean values in a probabilistic > sense. That is, the value of the FA at each voxel is mutiplied with the > probability at that voxel (from path.pd.nii.gz). > > Hope this helps, > a.y > > On Sat, 12 Oct 2013, Garikoitz Lerma-Usabiaga wrote: > > Dear Anastasia,I am visually inspecting the output of Tracula, > and checking > the stats on pathstats.overall. > > For example, in the following image (S_23_...) it can be seen > that the > rh.slfp is missing (there is just a very small piece of it). > The rest of the > tracts seem to be ok. > In the stats the four xx_Avg_Center values are a NaN, and the > value of the > volume and Len_xx are the same, excepting Len_Center, which is > zero. > Although there are more statistics, I understand we should > convert the whole > line to NaN and not include the data of this specific > subject/tract in the > statistics, right? (and should I script it to make it happen > always? - I > mean, always that there is Len_Center=0, make everything NaN - > ). > Inline image 5 > I did freeview dpath/rh.slfp_PP_avg33_mni_bbr/path.pd.nii.gz as > recommended > in a the email list, and the path looks exactly like in this image. > > > > In the next image (S_13_...) there are more problems. > -- rh.slft missing (not showing in the image). Len_Center=0 again. > -- fmajor almost not there, but Len_Center=1 > -- lh.ccg much shorter than the rh.ccg > -- very strange looking rh.cab > > Inline image 6 > Should I consider all those tracts out of the statistics? > Is there any way to mend those? (if yes, where can I find documentation?) > > And last question: I understand that I could control the volume of the > tracts with ICV as it is done for other volumes, but, should I control for > the FA values? What exactly is FA_Avg_Weighted? (sorry, couldn't find any > documentation on that while searching). > > > Thank you very much again and sorry for long post! > > Gari > > PS note that I plotted the hippocampus as well because I am analyzing it > in > the same study. > > > > > > > > > > > > The information in this e-mail is intended only for the person to whom it is > addressed. If you believe this e-mail was sent to you in error and the e-mail > contains patient information, please contact the Partners Compliance HelpLine at > http://www.partners.org/complianceline . If the e-mail was sent to you in error > but does not contain patient information, please contact the sender and properly > dispose of the e-mail. > > > >

1 0

Re: [Freesurfer] bvecs and bvals don't have the same number of entries
by Anastasia Yendiki 15 Oct '13

15 Oct '13

Hi Laura - This means that there is only 1 line in each of the files, so the end-of-line characters aren't recognized. I'd make sure that the file is saved as plain text. a.y On Tue, 15 Oct 2013, Laura Christine Anderson wrote: > When I run "wc -l" on the bvec and bval files, I get "1" for both. Running > just "wc" gives "1 279 3312" for bvec and "1 93 456" for bval > > Laura C. Anderson, B.A. > Graduate Student > Clinical & Developmental Psychology > University of Maryland > College Park, MD 20742 > BPS 0112 > > > > > On 10/14/13 5:18 PM, "Anastasia Yendiki" <ayendiki(a)nmr.mgh.harvard.edu> > wrote: > >> >> Hi Laura - I don't think the error "bvecs and bvals don't have the same >> number of entries" should have to do with interspersed zero b-values. >> When >> you run "wc -l" on the bvecs and bvals files, what do you get? >> >> a.y >> >> On Mon, 14 Oct 2013, Laura Christine Anderson wrote: >> >>> Hi Anastasia, the bval/bvec files were created on a Mac and not with >>> excel. I am not seeing the ^M character when I view the files in unix >>> with >>> the more command or in gedit perhaps when I sent you the files, >>> something >>> became corrupted? I am wondering if my issue has something to do with >>> the >>> fact that there are 3 sets of 30 rows with a b0 image before each of >>> the 3 >>> sets. In the examples I have seen, the b0 images all seem to be at the >>> beginning of the bvec and bval files. I am attaching a screenshot of the >>> bvec and bval files as I see them in case that is helpful. If you would >>> rather me send the actual files again, what is the best way to do that >>> without the formatting getting messed up? Before, I was using a free >>> fileshare program. >>> >>> Thank you! >>> Laura >>> >>> Laura C. Anderson, B.A. >>> Graduate Student >>> Clinical & Developmental Psychology >>> University of Maryland >>> College Park, MD 20742 >>> BPS 0112 >>> >>> >>> >>> >>> On 10/11/13 2:38 PM, "Anastasia Yendiki" <ayendiki(a)nmr.mgh.harvard.edu> >>> wrote: >>> >>>> >>>> Hi Laura - Were these bval/bvec files created on windows or with excel >>>> by >>>> any chance? When you view them in unix (with the more command for >>>> example) >>>> you'll see an "^M" character instead of every end-of-line. Can you try >>>> editing these files in a unix editor to fix that? Let me know if this >>>> solves the problem. >>>> >>>> a.y >>>> >>>> On Fri, 11 Oct 2013, Laura Christine Anderson wrote: >>>> >>>>> Hi Anastasia, >>>>> Thanks so much for your help! Attached are bvec, bval and trac-all.log >>>>> files. >>>>> >>>>> I have been stuck at one of the preprocessing steps (trac all) for >>>>> some >>>>> time and keep getting the >>>>> error: bvecs and bvals don't have the same number of entries. I saw a >>>>> couple blogs online about >>>>> this, but did not see any resolution. I am wondering if it is because >>>>> the B0 images are >>>>> interspersed? (see attached bvec file). If you have any insight, I >>>>> would love some feedback >>>>> whenever you get the chance. >>>>> >>>>> Laura >>>>> >>>>> >>>> >>>> >>>> The information in this e-mail is intended only for the person to whom >>>> it >>>> is >>>> addressed. If you believe this e-mail was sent to you in error and the >>>> e-mail >>>> contains patient information, please contact the Partners Compliance >>>> HelpLine at >>>> http://www.partners.org/complianceline . If the e-mail was sent to you >>>> in >>>> error >>>> but does not contain patient information, please contact the sender and >>>> properly >>>> dispose of the e-mail. >>> >>> > > > >

1 0

Group Analysis with fsgd
by Worker, Amanda 15 Oct '13

15 Oct '13

Dear all, I am trying to run a group analysis through the terminal using an .fsgd file, with no additional covariates. GroupDescriptorFile 1 Title PDPLUS2 Class PD Class MSA Class PSP Input PDPLUS007/ PD Input PDPLUS005/ MSA Input PDPLUS004/ PSP etc... When I try to run a command similar to the example fsgd analysis on the freesurfer wiki: mri_glmfit \ --glmdir pdplus \ --y y.mgh \ --fsgd g3v0.fsgd \ --C PD-MSA.mtx \ --C PD-PSP.mtx \ --C MSA-PSP.mtx \ --C group.effect.mtx \ --C PD+MSA-vs-ad.mtx I have received the following error message: Creating output directory pdplus2 Loading y from /home/k1204763/images/PDPLUS/y.mgh mghRead(/home/k1204763/images/PDPLUS/y.mgh, -1): could not open file ERROR: loading y /home/k1204763/images/PDPLUS/y.mgh I am not too sure whether y.mgh actually exists, is there a way that I can create this file for the three groups? Any other thoughts on this error would very much be appreciated. Thanks in advance, Amanda

1 0

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