Freesurfer

freesurfer@nmr.mgh.harvard.edu

3 participants
30135 discussions

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

Re: [Freesurfer] Combined parcellation of the left and right hemispheres (Bruce Fischl)
by varshas＠cns.iisc.ernet.in 22 Jul '16

22 Jul '16

> Message: 17 > Date: Thu, 21 Jul 2016 09:31:57 -0400 (EDT) > From: Bruce Fischl <fischl(a)nmr.mgh.harvard.edu> > Subject: Re: [Freesurfer] Combined parcellation of the left and right > hemispheres > To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu> > Message-ID: > <alpine.LRH.2.20.1607210931470.14081(a)door.nmr.mgh.harvard.edu> > Content-Type: text/plain; charset=US-ASCII; format=flowed > > Hi Varsha > > try using mri_aparc2aseg Hi Bruce, I tried using mri_aparc2aseg. While it gives me one volume file as I wanted, the indices on the volume don't seem to correspond to the look-up table that I created. I created the look-up table from the annotation files for the left and right hemispheres, appropriately changing the indices to make it a continuous list of ROIs. Regards, Varsha > cheers > Bruce > On Thu, 21 Jul 2016, varshas(a)cns.iisc.ernet.in > wrote: > >> >> Hi, >> >> I've been working on getting a finer parcellation of the >> Desikan-Killiany >> atlas using mris_divide_parcellation. I was able to generate two >> annotation files; one each for the left and right hemisphere. I want to >> create a volume similar to aparc+aseg.mgz, containing the labels and >> indices of the ROIs from both hemispheres. I tried using mri_label2vol. >> However, this gives me two separate volumes for the left and right >> hemispheres. Is there a way to combine these volumes into one? >> Alternatively, can one generate such a volume form an annotation file >> that >> has information of both hemispheres? >> >> Any help is greatly appreciated! >> >> Regards, >> Varsha >> >> >> >> >> >> _______________________________________________ >> Freesurfer mailing list >> Freesurfer(a)nmr.mgh.harvard.edu >> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer >> >> >> > > > ------------------------------ > > Message: 18 > Date: Thu, 21 Jul 2016 09:33:29 -0400 (EDT) > From: Bruce Fischl <fischl(a)nmr.mgh.harvard.edu> > Subject: Re: [Freesurfer] Freesurfer build stamps do not match > To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu> > Message-ID: > <alpine.LRH.2.20.1607210932310.14081(a)door.nmr.mgh.harvard.edu> > Content-Type: text/plain; charset=US-ASCII; format=flowed > > Hi Subin > > it looks more likely that you just don't have write access to your > SUBJECTS_DIR. Can you check that with ls -l? Or maybe you just haven't set > it and it still points to the one we distribute? Typically you want to set > the SUBJECTS_DIR env variable to some other directory where you will store > all the subjects in a study. > > cheers > Bruce > > > On Thu, 21 Jul 2016, Lee > Subin Kristine wrote: > >> >> Hi FreeSurfer Team, >> >> >> After I have downloaded the FreeSurfer dev version, I get the following >> error >> whenever I use the recon-all comand: >> >> >> mkdir: cannot create directory '/usr/local/freesurfer/subjects/NYS': >> Permission >> denied >> mkdir: cannot create directory '/usr/local/freesurfer/subjects/NYS': >> Permission >> denied >> mkdir: cannot create directory '/usr/local/freesurfer/subjects/NYS': >> Permission >> denied >> cp: cannot create regular file >> '/usr/local/freesurfer/subjects/NYS/scripts/build-stamp.txt': No such >> file or >> directory >> cat: /usr/local/freesurfer/subjects/NYS/scripts/build-stamp.txt: No such >> file or >> directory >> INFO: FreeSurfer build stamps do not match >> Subject Stamp: >> Current Stamp: freesurfer-Linux-centos6_x86_64-dev-20160611-876a0e6 >> /usr/local/freesurfer/subjects/NYS/scripts/patchdir.txt: No such file or >> directory. >> >> recon-all used to work fine when I used the stable 5.3 version. >> I think it may have something to do with the fact that I previously ran >> recon-all >> in v.5.3, and so the current subject directory's stamp is still 5.3 (or >> blank..) >> and it does not match the current stamp of the dev version? I am not >> sure.. >> Could anyone tell me how I can fix this problem? >> >> Thanks in advance, >> Subin >> >> >> > > > ------------------------------ > > Message: 19 > Date: Thu, 21 Jul 2016 09:35:35 -0400 (EDT) > From: Bruce Fischl <fischl(a)nmr.mgh.harvard.edu> > Subject: Re: [Freesurfer] Subtracting overlays > To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu> > Message-ID: > <alpine.LRH.2.20.1607210933500.14081(a)door.nmr.mgh.harvard.edu> > Content-Type: text/plain; charset="iso-8859-15" > > Hi Tobias > > yes, mris_calc should do it. If you give the output extension .mgz it will > save in that format instead of curv I believe (and will work fine as an > overlay) > > cheers > Bruce > > > On Thu, 21 Jul 2016, Granberg, Erik Tobias wrote: > >> Hi,? >> Is there a way to subtract one surface overlay from another (for example >> in >> fsaverage) and still having the output being an .mgh overlay file?? >> I saw that mis_calc does a similar procedure but automatically outputs >> the result >> as out.mgz.? >> >> Thank you for the help!? >> >> Kind regards, >> Tobias >> ? >> _________________________________________? >> Tobias Granberg, MD, PhD? >> >> Post-doctoral research fellow >> A.A. Martinos Center for Biomedical Imaging? >> Massachusetts General Hospital | Harvard Medical School >> >> Resident in Radiology? >> Karolinska Institutet | Department of Clinical Science, Intervention and >> Technology >> Karolinska University Hospital | Department of Radiology >> >> > > ------------------------------ > > Message: 20 > Date: Thu, 21 Jul 2016 13:42:48 +0000 > From: "Granberg, Erik Tobias" <EGRANBERG(a)mgh.harvard.edu> > Subject: Re: [Freesurfer] Subtracting overlays > To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu> > Message-ID: <9294313C-F63A-420D-BF9C-D568A9A7B21D(a)mgh.harvard.edu> > Content-Type: text/plain; charset="us-ascii" > > Hi Bruce, > > Great, thanks for the quick reply! That worked beautifully. I was > apparently using the wrong output option. > Much appreciated! > > Kind regards, > Tobias > > _________________________________________ > Tobias Granberg, MD, PhD > > Post-doctoral research fellow > A.A. Martinos Center for Biomedical Imaging > Massachusetts General Hospital | Harvard Medical School > > Resident in Radiology > Karolinska Institutet | Department of Clinical Science, Intervention and > Technology > Karolinska University Hospital | Department of Radiology > > > On 21 jul 2016, at 09:35, Bruce Fischl > <fischl(a)nmr.mgh.harvard.edu<mailto:fischl@nmr.mgh.harvard.edu>> wrote: > > Hi Tobias > > yes, mris_calc should do it. If you give the output extension .mgz it will > save in that format instead of curv I believe (and will work fine as an > overlay) > > cheers > Bruce > > > On Thu, 21 Jul 2016, Granberg, Erik Tobias wrote: > > Hi, > Is there a way to subtract one surface overlay from another (for example > in > fsaverage) and still having the output being an .mgh overlay file? > I saw that mis_calc does a similar procedure but automatically outputs the > result > as out.mgz. > Thank you for the help! > Kind regards, > Tobias > > _________________________________________ > Tobias Granberg, MD, PhD > Post-doctoral research fellow > A.A. Martinos Center for Biomedical Imaging > Massachusetts General Hospital | Harvard Medical School > Resident in Radiology > Karolinska Institutet | Department of Clinical Science, Intervention and > Technology > Karolinska University Hospital | Department of Radiology > _______________________________________________ > Freesurfer mailing list > Freesurfer(a)nmr.mgh.harvard.edu<mailto:Freesurfer@nmr.mgh.harvard.edu> > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer > >

2 3

Superimposing average hippocampal colume onto a betted brain
by Dr Sampada Sinha 22 Jul '16

22 Jul '16

Dear freesurfer experts, Thank you for guiding me on how to create average volume of all subjects of depression patients. I later binarized the average volume generated to extract the amygdala and hippocampus volume(Right and left). Please find attached. How do I superimpose the extracted amygdala and hippocampus on to the betted image using freesurfer commands? I am trying to reproduce the following image as an example. (Can I use vol2vol command?) Greatly appreciate all your help. Thanks and regards, Sampada -- Sampada Pre-doctoral student Department of Geriatric Mental Health (DGMH) King George Medical University Lucknow-226003 India

1 0

PhD position in structural imaging in First Episodes of Non-Affective Psychosis
by Diana Tordesillas-Gutiérrez 22 Jul '16

22 Jul '16

"PhD position in structural imaging in First Episodes of Non-Affective Psychosis" A four years funded full-time PhD position will be available at the Neuroimaging Unit – Health Research Institute Valdecilla (IDIVAL) headed by Professor Benedicto Crespo- Facorro from October 2016. We are seeking for a highly motivated individual with experience in brain imaging and computational/statistical background to take a leading role in structural neuroimaging analysis. The candidate should hold a degree in psychology, biology, neuroscience or a related field. This work will be co-directed by Prof. B Crespo-Facorro (Psychiatrist) and Dr. D Tordesillas-Gutiérrez (Physicist). Salary is according to Spanish standards. Interested candidates should send their CV along with a letter of motivation and their detailed marks during their Master (or latest achieved degree) to Dr. Diana Tordesillas-Gutiérrez, e-mail: diana.tordesillas(a)humv.es ------------------------------------------------------------------------------------------------------------------------------ Diana Tordesillas Gutiérrez, PhD Neuroimaging Unit IDIVAL IDIVAL Building Avda. Cardenal Herrera Oria s/n 39011 Santander Spain Tel. +34 942202677 diana.tordesillas(a)humv.es diana.tordesillas(a)gmail.com ------------------------------------------------------------------------------------------------------------------------------

1 0

aparc+aseg.mgz for network analysis
by Jinsong Tang 21 Jul '16

21 Jul '16

Hello freesurfers, I want to select the subcortical structures (thalamus, pallidum, caudate, putamen, accumbens, hippocampus, and amygdala) and the 68 distinct brain regions as nodes of the individual brain networks. The labels are [9,11:13,17:18,26,48,50:54,58, 1002:1035, 2002:2035]. However, in the aparc+aseg.mgz files, there are also some other labels. Is there any way to delete all other unselected labels? Thank you so much! Best, Jinsong

4 5

question about image size
by Dorsa Haji Ghaffari 21 Jul '16

21 Jul '16

Hi, I have segmented 288 slices of 512x512 dicom MRIs but when I create a mask of the left thalamus I get 256 slices of 256x256 images. Is there any way to get the same number of slices with the same size as I had before? Thank you Dorsa

2 2

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