Freesurfer

freesurfer@nmr.mgh.harvard.edu

11 participants
30080 discussions

Re: [Freesurfer] Checking segmentations with aseg.mgz
by Reema Jayakar 25 Jul '16

25 Jul '16

Hi Dr. Fischl, Thank you for clarifying that about norm.mgz for checking segmentations. Good to know that brain and brainmask wash out low contrast borders. Cheers, Reema -- *Reema Jayakar, M.A.* Doctoral Candidate - Clinical Neuropsychology Health Resources and Services Administration (HRSA) Fellow Department of Psychology Georgia State University Email: rjayakar1(a)student.gsu.edu On Mon, Jul 18, 2016 at 4:42 PM, Reema Jayakar <reemajayakar(a)gmail.com> wrote: > Hello FS Listserv members, > > With regard to checking amygdala segmentation, is it appropriate to > overlay the aseg.mgz file on the brainmask.mgz? Or should I be overlaying > aseg.mgz on norm.mgz? I am aware of the general rule of thumb that > editing the subcortical segmentations generally is not recommended. > > Also, where can I find documentation on the Freeview options "Resample to > standard RAS space" and > > "Apply registration file". I am trying to understand whether or not it is necessary for me to use any of these options for what I am doing with the subcortical segmentations. > > > Thanks in advance for your help! > > Reema > > > > -- > *Reema Jayakar, M.A.* > Doctoral Candidate - Clinical Neuropsychology > Health Resources and Services Administration (HRSA) Fellow > Department of Psychology > Georgia State University > Email: rjayakar1(a)student.gsu.edu > >

1 0

mri_label2label volume registration problem
by Agoston Mihalik 25 Jul '16

25 Jul '16

Hi Freesurfer Developers, I would like to ask for your expertise in the following problem I encountered: I want to create subject specific labels based on a label I created in fsaverage. The transformation needs to be based on Talairach intermediate registration, since my label was drawn based on Talairach coordinates. I ran the following command: mri_label2label --srclabel test.label --srcsubject fsaverage --trglabel lh.test2.label --trgsubject subj222 --regmethod volume The problem is that the output label is scattered and all around the place when I look at it in tksurfer - see the image attached (lh.regmethod_volume.tif). When I did the registration based on surface, it worked fine (see the image also attached - lh.regmethod_volume.tif), however, it resulted in 0-4 mm misalignment when I looked at the Talairach coordinates of the label borders. I guess it is due to the surface registration that optimises the surfaces to be aligned, not the Talairach coordinates. Please find some information attached in case it is any help - content of files, output etc. Here is my freesurfer version: freesurfer-Darwin-lion-stable-pub-v5.3.0 Your help would be really appreciated. Thanks and best, Agoston

2 1

ERROR: cannot find volume matching
by Erik Jahner 25 Jul '16

25 Jul '16

Dear Free Surfer experts I am attempting to run a functional connectivity analysis on some resting state data. However, I keep getting stuck in this final step of the online Functional Connectivity Pathway. I have made a change based on discussion boards, but it do not help. CHANGE: in the mkanalysis step, I changed the "-taskreg L_Posteriorcingulate.dat 1" to "-taskreg L_Posteriorcingulate 1². This did not solve the problem. My mkanalysis lines are as follows: mkanalysis-sess -analysis fc.lpccseed.surf.lh -surface fsaverage lh -fwhm 5 -notask -taskreg L_Posteriorcingulate 1 -nuisreg vcsf.dat 5 -nuisreg wm.dat 5 -mcextreg -polyfit 5 -nskip 4 -fsd rest -TR TR2 -hpf .01 -force -per-run mkanalysis-sess -analysis fc.lpccseed.surf.rh -surface fsaverage rh -fwhm 5 -notask -taskreg L_Posteriorcingulate 1 -nuisreg vcsf.dat 5 -nuisreg wm.dat 5 -mcextreg -polyfit 5 -nskip 4 -fsd rest -TR TR2 -hpf .01 -force -per-run mkanalysis-sess -analysis fc.lpccseed.surf.mni305 -mni305 -fwhm 5 -notask -taskreg L_Posteriorcingulate 1 -nuisreg vcsf.dat 5 -nuisreg wm.dat 5 -mcextreg -polyfit 5 -nskip 4 -fsd rest -TR TR2 -hpf .01 -force -per-run I then run the following: selxavg3-sess -s 401 -a fc.lpccseed.surf.lh -overwrite and the error that is produced is as follows: (what does "cannot find volume ² mean? and how might I fix this) /Applications/freesurfer/LagStructure/413 ------------------------- $Id: fast_selxavg3.m,v 1.100.2.2 2012/11/30 18:40:38 greve Exp $ /Applications/freesurfer/fsfast/toolbox/fast_selxavg3.m /Applications/freesurfer/fsfast/toolbox/fast_ldanaflac.m /Applications/freesurfer/matlab/MRIread.m ------------------------- outtop = /Applications/freesurfer/LagStructure Extension format = nii.gz ERROR: cannot find volume matching /Applications/freesurfer/LagStructure/413/rest/001/L_Posteriorcingulate ERROR: loading nonpar reg /Applications/freesurfer/LagStructure/413/rest/001/L_Posteriorcingulate ------------------------------------------ ERROR: fast_selxavg3() failed\n

2 1

LME for surface area
by Clara Kühn 25 Jul '16

25 Jul '16

Dear FreeSurfer Experts, I'm following the LME tutorial to analyze my data. The example that is given in the tutorial is for a left hemispheric cortical thickness analysis. I would also like to look at surface area. Would I follow the same steps? At the point in the tutorial that uses the lme_mass_fit_Rgw command I have to specify a Distance type (euc or surf). How would I do that regarding surface area? Would I have to skip this step or is there a similar command or does it just work with different arguments? Unfortunately, surface area isn't mentioned in LME tutorial. Thanks for your help! Clara -- Clara Kühn, Phd Student Max-Planck-Institute for Human Cognitive and Brain Science Department of Neuropsychology Stephanstrasse 1A 04103 Leipzig, Germany Phone: +49 341 - 9940 2271 Fax: +49 341 - 9940 2260 Web: www.cbs.mpg.de E-Mail: ckuehn(a)cbs.mpg.de

1 0

Post Doc Position advert!
by Muthuraman Muthuraman 25 Jul '16

25 Jul '16

*Post-Doc* *Neuroscientist with focus on multimodal brain imaging* The *medical faculty of the University of Mainz*, Germany, invites applications for a Post Doc Position for a functional translational neuroscience (FTN) scholarship two years position to work on multi-modal imaging, neurostimulation (TMS) and EEG. Additionally, the new position requires experience with structural and functional MRI data analysis, computational modeling or magnetic diffusion imaging. The candidate should have a PhD or equivalent degree preferably in neuroscience, biomedical physics, psychology, psychiatry, or biology and should be able to supervise experiments and data analyses in programming platforms. The working language at the institute is English. The location for this research will mainly be the workgroups “Section of movement disorders, neurostimulation and neuroimaging” of Prof. Sergiu Groppa at the Department of Neurology and “Biomedical statistics and multimodal signal processing unit” of Prof. Muthuraman Muthuraman, both at the Focus Program Translational Neurosciences (http://www.ftn.uni-mainz.de/) and Neuroimaging Centre Mainz (http://www.ftn.nic.uni-mainz.de/) and within the CRC “Neurobiology of resilience to stress-related mental dysfunction” and SFB TR-128 “Initiating/effector versus regulatory mechanisms in Multiple Sclerosis” (www.sfbtr128.de) Expected close collaborations with national and international partners. The application should include a statement of research interests and reasons for applying to the project, curriculum vitae (max. 4 pages) composed according to good scientific practice. The position will be open until filled. To apply for the position, please send the above documents to *Prof.* *Sergiu Groppa and Prof. Muthuraman Muthuraman*, Department of Neurology, Section of Movement Disorders and Neurostimulation, University of Mainz, Langenbeckstrasse 1, 55131 Mainz, Germany, or by Email to *segroppa(at)uni-mainz.de <http://uni-mainz.de> & mmuthura(at)uni-mainz.de <http://uni-mainz.de>*.

1 0

Incongruity between by-hand correlations and qdec correlations
by Jeffrey Crawford 24 Jul '16

24 Jul '16

Hello, I am a new user of freesurfer and I have not found any pages on the wiki that address my issue. I am running a correlational analysis between regions of the brain and cognitive scores for controls and patients with schizophrenia. After preprocessing all of the MPRAGESs, I generated stats tables for all of the structures in the brain for all of my sample on freesurfer. I then took the raw stats values and did massive correlation matrix on spss to see what regions of the brain most correlated with my factor scores. After this, I isolated the most significant values to see if I could determine any sort of pattern from the data and used this for reference as I continued the analysis on qdec. >From qdec, I tried to reproduce my correlations from the raw stats table data and none of the correlations that I produced were viewable on the brains in Freesurfer. I have included a selection of pictures to illustrate my process, so that you all better understand my predicament. (Picture1) These are three simple graphs I created in excel from the data I was using on spss. The graph on the top left are my factor scores for all of my patients. The graph on the bottom left are the areas of the cuneus for all of my patients. The graph on the right is the correlation between those two values, which produces a correlation of -.55 at a .002 significance. (Picture 2 and 3) I then replicated the two graphs from my excel sheet on qdec. I generated the stats tables for my sample again, then used the area of the cuneus produced by qdec. My factor score (picture 2) and the area of the cuneus (picture 3) match my two original graphs exactly. This leads me to believe that qdec is using the exact same values that I was using in spss. (Picture 4) These are my design parameters that I inputted. I am measuring area, FWHM of 10, left hemisphere, no discrete factor (since I only want the correlation run on the 31 patients I selected initially), my factor score is my continuous variable, no nuisance factor, and am using DODS. (Picture 5) This is the brain that was produced. The marker is placed where the cuneus is located and there is no significant negative correlation in that location. (Picture 6) This is the graph produced by qdec for the correlation between my factor score and the area of the cuneus. It looks nothing like the correlation graph I produced in picture 1, so I am confused. >From talking to my colleague who has used freesurfer before, there should be a consistency between correlations drawn from the raw stats table values and the correlations presented by freesurfer, but I just can't find it. I have feeling that, if I am doing something wrong (the probable case), it is happening at picture 4 where I am inputting the sample information into the analysis. If not, then does freesurfer do any sort of transformation or thresholding that I am not aware of that would change the presentation of the raw results from the stats table? This is of great concern to me, because a large part of my intended analysis will be on subcorical and white matter values, which in my understanding, are only accessible from the raw stats table. I apologize for my ignorance and hope that someone out there can shed some light on what is going wrong. If you need me to send my stats table or my factors scores, then I can. Many thanks, Jeff

1 0

question about save_mgh and mri_label2vol
by Jinsong Tang 24 Jul '16

24 Jul '16

I have two questions" 1. the type of aparc+aseg.mgz file is INT(1). after save_mgh(v, 'aparc+aseg82regions.mgz',M,mr); I find the type of aparc+aseg82regions.mgz is FLOAT (3). How to save aparc+aseg82regions.mgz file in INT ? 2. I want to convert the aparc+aseg82regions.mgz to .nii file (in native space) I use the following command lines: mri_label2vol --seg aparc+aseg82regions.mgz --temp rawavg.mgz --o aparc+aseg82regions-rawavg.mgz --regheader aparc+aseg82regions.mgz mri_convert -odt int aparc+aseg82regions-rawavg.mgz aparc+aseg82regions-rawavg.nii after these, I found the dimension is changed from 256*256*256 to 176*204*200. Volume information for aparc+aseg82regions.mgz type: MGH dimensions: 256 x 256 x 256 voxel sizes: 1.0000, 1.0000, 1.0000 type: FLOAT (3) fov: 256.000 dof: 1 Volume information for aparc+aseg82regions-rawavg.mgz type: MGH dimensions: 176 x 204 x 200 voxel sizes: 1.0000, 1.0000, 1.0000 type: INT (1) fov: 204.000 dof: 0 Volume information for aparc+aseg82regions-rawavg.nii type: nii dimensions: 176 x 204 x 200 voxel sizes: 1.0000, 1.0000, 1.0000 type: INT (1) fov: 176.000 dof: 0 I guess I must have done something wrong. But I can not figure out. Best, Jinsong

2 2

question about save_mgh and mri_label2vol
by Jinsong Tang 24 Jul '16

24 Jul '16

Dear all, I have two questions. 1. the type of aparc+aseg.mgz file is INT(1). after save_mgh(v, 'aparc+aseg82regions.mgz',M,mr); I find the type of aparc+aseg82regions.mgz is FLOAT (3). How to save aparc+aseg82regions.mgz file in INT ? 2. I want to convert the aparc+aseg82regions.mgz to .nii file (in native space) I use the following command lines: mri_label2vol --seg aparc+aseg82regions.mgz --temp rawavg.mgz --o aparc+aseg82regions-rawavg.mgz --regheader aparc+aseg82regions.mgz mri_convert -odt int aparc+aseg82regions-rawavg.mgz aparc+aseg82regions-rawavg.nii after these, I found the dimension is changed from 256*256*256 to 176*204*200. Volume information for aparc+aseg82regions.mgz type: MGH dimensions: 256 x 256 x 256 voxel sizes: 1.0000, 1.0000, 1.0000 type: FLOAT (3) fov: 256.000 dof: 1 Volume information for aparc+aseg82regions-rawavg.mgz type: MGH dimensions: 176 x 204 x 200 voxel sizes: 1.0000, 1.0000, 1.0000 type: INT (1) fov: 204.000 dof: 0 Volume information for aparc+aseg82regions-rawavg.nii type: nii dimensions: 176 x 204 x 200 voxel sizes: 1.0000, 1.0000, 1.0000 type: INT (1) fov: 176.000 dof: 0 I guess I must have done something wrong. But I can not figure out. Thanks a lot! Best, Jinsong

1 0

mri_glmfit --mask ERROR: dimension mismatch 1 between y and mask
by Mihaela Stefan 22 Jul '16

22 Jul '16

Hello freesurfers, I would like to use mri_glmfit with --mask but I get this error: dimension mismatch 1 between y and mask. I created a binary mask from 22 aparc labels (using mri_binarize) and I would like to run a surface-based analysis only on those regions. The command I use is: mri_glmfit --y lh.g2v0.thickness.15.mgh --fsgd g2v0.fsgd dods --C lh.group.diff.mtx --surf fsaverage lh --mask lh22mask.mgz --glmdir lh.group_diff_g2v0.glmdir The input file was generated with this command: mris_preproc --fsgd g2v0.fsgd --cache-in thickness.fwhm15.fsaverage --target fsaverage --hemi lh --out lh.g2v0.thickness.15.mgh >From the error, it seems that I am not using the right input file. What kind of input file should be used with a mask? I also thought to use --label but I am not sure how to combine my 22 labels in one single label. mris_label2annot can combine them but the output is an annotation file not a label. As a note, we will want to do FDR correction, not MC. Thanks! Mihaela

2 8

Fwd: Help with strange 'deprecated' errors with recon-all -i
by Cat Hubbard 22 Jul '16

22 Jul '16

Hi, I am trying to import and recon-all -i my subject files and am getting these strange errors that I have never encountered before. Iteration 1 Tue Jul 19 19:21:54 EDT 2016 nu_correct -clobber ./tmp.mri_nu_correct.mni.83022/nu0.mnc ./tmp.mri_nu_correct.mni.83022/nu1.mnc -tmpdir ./tmp.mri_nu_correct.mni.83022/0/ defined(@array) is deprecated at /Applications/freesurfer/mni/bin/nu_estimate_np_and_em line 165. Use of ?PATTERN? without explicit operator is deprecated at /Applications/freesurfer/mni/bin/sharpen_volume line 153. I am trying to reconcile, having looked at various posts I see some mention of this maybe having to do with the version of perl I am using. The version I am currently running is perl 5, version 16, subversion 2 (v5.16.2). I also found other threads with mention of a similar error, and a possible fix but unsure how to implement. Can anyone please help me with this? Thanks!! https://mail.nmr.mgh.harvard.edu/pipermail//freesurfer/2013-May/029906.html https://github.com/BIC-MNI/N3/commit/f5a9d0b07369039ac2d1b87e3eb7d73f6fbde7…

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