Freesurfer

freesurfer@nmr.mgh.harvard.edu

5 participants
30076 discussions

PhD Position in Neuroimaging and Deep Learning at Karolinska Institute, Stockholm, Sweden
by Joana Braga Pereira 23 Feb '22

23 Feb '22

External Email - Use Caution Dear all, We are looking for a PhD student to work in an interdisciplinary and internationally competitive environment at Karolinska Institute, one of the top medical universities in the world and the home for the Nobel Prize in Medicine. We combine imaging analyses, statistical tools, machine learning algorithms, biofluid assays and genetics to study brain aging and the earliest signs of Alzheimer’s and Parkinson’s diseases. This project will be carried out under the supervision of Associate Professor Joana B. Pereira and it will involve applying methods based on deep neural networks to identify the anatomical and functional brain connections that are most vulnerable to Alzheimer’s disease and assess whether they can predict cognitive decline. The student will also have the opportunity to learn and apply computational models to identify the mechanisms responsible for the damage of brain connections in AD. Finally, the methods developed in this project in the second version of our software (BRAPH; http://secure-web.cisco.com/1Q2rgK7b8V70npiFIHUI8zfRjlE6qZiFHoTMy0cbxynHT1P…) which will soon be released. The ideal candidate should have the following qualities: - Background in computational science, neuropsychology, physics, statistics, medicine or a related field. - Experience in statistical techniques, programming, or scripting in MATLAB, Python and/or R. - Experience or interest in structural MRI, DTI, fMRI or PET imaging analyses. - Clinical or research experience with aging and/or neurodegenerative disorders are a merit. - Previous scientific publications are also a merit. The PhD studies will be performed for 4 years. For further information about my group and how to apply please go to: https://secure-web.cisco.com/1ajXNJ3Ewvrf48vViFvbFKr5slri9XmrpQLG9K2_LASnQX… The application deadline is 15/03/2022. If you have any questions please contact me at jbragapereira(a)gmail.com or joana.pereira(a)ki.se. Best wishes, Joana

1 0

Incorrect SI/AP swap from bbregister
by Tisdall, Dylan 23 Feb '22

23 Feb '22

External Email - Use Caution Thanks for the help with my previous question on formatting the bbregister command line. Now that I've got it running, I'm seeing that 1 of my two subjects ends up with a 90 degree rotation (swapping SI and AP axes) when I try to register my 7T T1 or T2 scan to the surfaces from my 3T T1 scan. I've so far tried to analyze similar datasets from two subjects, and I'm only having this SI/AP swap in one of them (the second subject works great once I apply an initial transform to get it close) What steps should I take to try to understand what's going wrong? I've loaded the moving image into freeview, applying the initial transform as part of the load, and the result seems to align pretty closely with the surfaces when I load those as well... so I'd expect bbregister should align these well. Thanks again, Dylan

1 1

cortical measurments
by REEM ABU BAKR BAHATHIQ 23 Feb '22

23 Feb '22

External Email - Use Caution Hello Freesurfer experts How can I get a text file of the cortical measurements of a group of images after they have been processed with recon -all?? Best, Reem

2 1

convert .mgz to .nii
by REEM ABU BAKR BAHATHIQ 23 Feb '22

23 Feb '22

External Email - Use Caution Hello FreeSurfer experts, I want to run a classification task using deep learning on data processed by Freesurfer recon-all and I have some queries. I will use brainmask .mgz First of all, based on my simple knowledge, I cannot feed my model with this output directly. .mgz files must be converted to a .nii file for use in analytics. I found that this is done by writing mri_convert any_volume.mgz any_volume.nii.gz But how can I convert the files of all participants (more than 1000) into .nii files instead of going one by one? Are there any examples or tutorials provided to demonstrate the use of Freesurfer output in machine learning applications? My greetings,

2 3

Re: [Freesurfer] GLM analysis - volume and surface area
by Laura Willers de Souza 22 Feb '22

22 Feb '22

External Email - Use Caution Thanks so much for the explanation! There is something different in the analysis of volume and surface area in relation to the analysis of cortical thickness? Like some more command, or some other detail that must be changed (besides, of course, the flag that specifies the measure to be analyzed) - I'm already using permutation in the correction of multiple comparisons. Best, Laura Willers de Souza Master Student in Biochemistry Zimmer Lab - Federal University of Rio Grande do Sul - Porto Alegre/RS - Brazil

2 2

Annotation File to ROI Volume Estimate
by Jacoby, John 22 Feb '22

22 Feb '22

Hi all, We were sent a FreeSurfer annotation file that denotes some ROIs we're interested in. What's the proper way to map this annotation to the cortical ribbon volume and get an estimate of that volume for each ROI? I've already registered the annotations to the subject from fsaverage space using mri_surf2surf. I'm thinking something like: mri_label2vol --annot {registered_annotation_file} --temp {$SUBJECT/mri/orig.mgz} --o {output_volume_segmentation_file} --subject {$SUBJECT} --hemi {lh,rh} --identity --proj frac 0 1 0.01 And then run mri_segstats on the output_volume_segmentation_file This seems to do what I need, but I want to confirm that I'm not missing import details. I'm also not entirely sure what the 'delta' parameter of the --proj flag is doing. I gather that it's some kind of step size, but will it have a big impact on the resulting ribbon volume? Thanks for any help! Best, John

2 1

Re: [Freesurfer] Subject ID error with bbregister
by Tisdall, Dylan 22 Feb '22

22 Feb '22

External Email - Use Caution Hi Doug and Matt, I'm pretty confident my $SUBJECTS_DIR is correct, its the same one I used for recon-all on this subject and the files seem to be in the right place; wee below: [mtisdall@ftdc-gpu01 126346]$ echo $SUBJECTS_DIR /home/mtisdall/Research/7TAnalysis/subjects [mtisdall@ftdc-gpu01 126346]$ ls $SUBJECTS_DIR/126346 label mri scripts stats surf tmp touch trash I don't see a blank subject line in the lta file. I tried to make the initial transform manually in Freeview by just loading the two volumes, using `Transform Volume` until they were close, and then saving the resulting transform. I then re-loaded my 7T data, using the option in FreeView to apply the transform on load, and it looks right. See the file contents below: [mtisdall@scisub 126346]$ cat 7Tto3Tinit.lta type = 1 # LINEAR_RAS_TO_RAS nxforms = 1 mean = 0.0000 0.0000 0.0000 sigma = 1.0000 1 4 4 9.961947798728943e-01 8.715574443340302e-02 4.656613983300417e-10 -1.702789425849915e+00 -8.682409673929214e-02 9.924039244651794e-01 -8.715574443340302e-02 -7.269048690795898e+00 -7.596124429255724e-03 8.682409673929214e-02 9.961947798728943e-01 1.319813346862793e+01 0.000000000000000e+00 0.000000000000000e+00 0.000000000000000e+00 1.000000000000000e+00 src volume info valid = 1 # volume info valid filename = /home/mtisdall/Research/7TAnalysis/126346/niis/3T_MPRAGE.nii volume = 208 240 256 voxelsize = 1.000000000000000e+00 1.000000000000000e+00 1.000000000000000e+00 xras = 1.000000000000000e+00 0.000000000000000e+00 0.000000000000000e+00 yras = 0.000000000000000e+00 1.000000000000000e+00 0.000000000000000e+00 zras = 0.000000000000000e+00 0.000000000000000e+00 1.000000000000000e+00 cras = 5.000000000000000e-01 1.455932617187500e+01 -1.323728942871094e+01 dst volume info valid = 1 # volume info valid filename = /home/mtisdall/Research/7TAnalysis/126346/niis/3T_MPRAGE.nii volume = 208 240 256 voxelsize = 1.000000000000000e+00 1.000000000000000e+00 1.000000000000000e+00 xras = 1.000000000000000e+00 0.000000000000000e+00 0.000000000000000e+00 yras = 0.000000000000000e+00 1.000000000000000e+00 0.000000000000000e+00 zras = 0.000000000000000e+00 0.000000000000000e+00 1.000000000000000e+00 cras = 5.000000000000000e-01 1.455932617187500e+01 -1.323728942871094e+01 fscale 0.100000 Note that, when I move the --s flag to the end of the command line, it runs slightly further, but ultimately fails searching for `subject-unknown` within my subjects directory: [mtisdall@scisub 126346]$ bbregister --mov mgz/7T_MEMPRAGE.mgz --init-reg 7Tto3Tinit.lta --reg 7Tto3T.lta --o 7TMEMPRAGE_warped.mgz --t1 --s 126346 ... Loading mov error: mghRead(/home/mtisdall/Research/7TAnalysis/subjects/subject-unknown/mri/orig.mgz, -1): could not open file Any other suggestions on what I'm doing wrong here? Thanks again, Dylan > Date: Mon, 21 Feb 2022 16:58:22 -0500 > From: "Douglas N. Greve" <dgreve(a)mgh.harvard.edu> > Subject: Re: [Freesurfer] Subject ID error with bbregister > To: freesurfer(a)nmr.mgh.harvard.edu > Message-ID: <6f0565c8-c4f3-8289-d41b-0558a1fb69f9(a)mgh.harvard.edu> > Content-Type: text/plain; charset=UTF-8; format=flowed > > Is there a blank subject line in the lta file? If so, it is probably > trying to get the subject from there. You can try putting the --s after > the --lta > > On 2/21/2022 3:43 PM, Tisdall, Dylan wrote: >> External Email - Use Caution >> >> Hi all, >> >> Quick question on a puzzling error with bbregister version 7.1.1. When I run: >> >> bbregister --s 126346 --mov mgz/7T_MEMPRAGE --init-reg 7Tto3Tinit.lta --reg 7Tto3T.lta --o 7TMEMPRAGE_warped.mgz --t1 >> >> I get the error: >> >> ERROR: must spec a subject id >> >> which I thought I'd done with the --s flag. What am I missing? >> >> Even more odd, if I just move the to the end of the command, bbregister will run, but eventually fails with: >> >> error: mghRead(/hom/mtisdall/Research/7TAnalysis/subject//subject-unknown/mri/orig.mgz, -1): could not open file >> >> Is there something obvious I'm doing wrong with these two command lines that cause these behaviors? >> >> >> Thanks, >> Dylan >> >> >> >> >> >> _______________________________________________ >> Freesurfer mailing list >> Freesurfer(a)nmr.mgh.harvard.edu >> https://secure-web.cisco.com/12zXL_eM4ShAKwlFflhqDPeh22njZyomMazJzeiBX6fxvF… >

2 1

Re: [Freesurfer] autorecon2-wm not reflecting edits in white matter segmentation (7.1.1, 7.2.0)
by Fischl, Bruce 21 Feb '22

21 Feb '22

Hi Edan It’s hard for me to see what’s going on in those images. What is the intensity of the white matter in the middle of the occipital strands in the brain.mgz and the brain.finalsurfs.mgz? If it is significantly lower than 110 you probably want some control points in the occipital lobe. Cheers Bruce From: Edan Daniel <edand(a)princeton.edu> Sent: Friday, February 18, 2022 12:15 AM To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu>; Fischl, Bruce <BFISCHL(a)mgh.harvard.edu> Subject: Re: [Freesurfer] autorecon2-wm not reflecting edits in white matter segmentation (7.1.1, 7.2.0) External Email - Use Caution Hi Bruce! We are focusing on the occipital lobe where the white matter boundary is often too far into the gray matter, and we need those voxels to record visual signals. Attaching a few images for reference - notice how much 'thicker' the wm segmentation is in the white.preaparc vs the lh.orig.nofix. wm edits of this sort were accepted when we were using the same pipeline with v6 (separate datasets), but now with v7 our edits get 're-corrected' and overwritten. We'd be happy to provide more information or images. Many thanks!!! Edan @ PrincetonBrainDevLab On Thu, Feb 17, 2022 at 7:01 PM Fischl, Bruce <BFISCHL(a)mgh.harvard.edu<mailto:BFISCHL@mgh.harvard.edu>> wrote: Hi Edan Sometimes control points can be more effective than editing the wm, particularly if the intensities in that region are not what we expect (like if the wm and gm are darker than what we try to set them to). Can you post an image or two? Cheers Bruce From: freesurfer-bounces(a)nmr.mgh.harvard.edu<mailto:freesurfer-bounces@nmr.mgh.harvard.edu> <freesurfer-bounces(a)nmr.mgh.harvard.edu<mailto:freesurfer-bounces@nmr.mgh.harvard.edu>> On Behalf Of Edan Daniel Sent: Thursday, February 17, 2022 6:51 PM To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu<mailto:freesurfer@nmr.mgh.harvard.edu>> Subject: Re: [Freesurfer] autorecon2-wm not reflecting edits in white matter segmentation (7.1.1, 7.2.0) External Email - Use Caution * Different data, same pipeline. In v6 our changes were included in the final white surface segmentation, whereas in v7 they are overwritten during the placement 'refining'. * ?h.orig.nofix and ?h.orig DO reflect edits (?h.orig to some extent). white.preaparc does NOT. * We noticed that the placement is very stubborn and we were not able to overcome this. Methods like manually editing the wm surface worked well with v6 and do not work now with v7. We have also tried editing the filled.mgz with v7.2 following the new tutorial, and that did not work either. Is there any way we can overcome this stubbornness? Many thanks! Edan @ Princeton BrainDevLab On Thu, Feb 17, 2022 at 6:07 PM Douglas N. Greve <dgreve(a)mgh.harvard.edu<mailto:dgreve@mgh.harvard.edu>> wrote: The edits you are making only change the initial surface; the surface placement then refines the placement. Sometimes the surface placement can be quite stubborn. If this is the case, the ?h.orig.nofix and ?h.orig will reflect your edits (nofix exactly; ?h.orig to some extent). Also check the the white.preaparc is where the problem starts. When you say that v6 did better, do you mean on this same data or on different data? The surface placement has not changed that much since v6. On 2/14/2022 12:50 PM, Edan Daniel wrote: External Email - Use Caution We have been having issues with recreating the final surfaces after editing the white matter segmentation, which used to work smoothly in version 6. After running a subject through recon-all (version 7.1.1 *without* the parallel flag) and then editing the white matter, we run the following command (w v7.1.1): recon-all -autorecon2-wm -autorecon3 -subjid ' freesurferfoldername ' -cc-crs ' num2str(corpus_point) The ribbon file and surfaces do update *very slightly* but it’s largely ignoring our (sometimes large) edits. This pipeline worked well for us in version 6- do you have any idea what might be going on? Would love to hear your thoughts about this. Many thanks!! Edan, Braindevlab@Princeton _______________________________________________ Freesurfer mailing list Freesurfer(a)nmr.mgh.harvard.edu<mailto:Freesurfer@nmr.mgh.harvard.edu> MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer<https://secure-web.cisco.com/1BhcjfvnDAXHKeDR6R-iodPMLY3xi6xzTIkDcLTzdfO3AL…> _______________________________________________ Freesurfer mailing list Freesurfer(a)nmr.mgh.harvard.edu<mailto:Freesurfer@nmr.mgh.harvard.edu> MailScanner has detected a possible fraud attempt from "secure-web.cisco.com" claiming to be https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer<https://secure-web.cisco.com/1BhcjfvnDAXHKeDR6R-iodPMLY3xi6xzTIkDcLTzdfO3AL…> _______________________________________________ Freesurfer mailing list Freesurfer(a)nmr.mgh.harvard.edu<mailto:Freesurfer@nmr.mgh.harvard.edu> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer<https://secure-web.cisco.com/12WGi9fwMNRKoc7g0UsFngq4QzxkPkp8uUBbbPuKzCH_q-…>

3 7

Contrasts 3 groups 4 covariates
by fernanda rohrsetzer 21 Feb '22

21 Feb '22

External Email - Use Caution Hi guys, I'm a beginner in Freesurfer and I want to perform an analysis of the difference between 3 groups in cortical thickness,volume and area having age, mean thickness and etiv as covariate. my design matrix for 3 groups comparison as "1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 (LR-HR) 1 0 -1 0 0 0 0 0 0 0 0 0 0 0 0 (LR-MDD) 0 1 -1 0 0 0 0 0 0 0 0 0 0 0 0 (HR-MDD) If possible, I would like to know how my contrasts change when I wish to test the change in thickness with age, after removing the group effects. Another question I have is how can I create contrasts to control etiv between groups, like 2x2 comparison. I'm sorry if I was not clear... thanks Fernanda my FSGD file is GroupDescriptorFile 1 Title MyStudy Class LR Class HR Class MDD Variables age meanThick totalarea eTIV Input Sub001 LR 14.3 1350519.71 1954527.71 1552459.71 Input Sub002 MDD 15.7 1558515.86 1689767.98 1809809.87

2 1

GLM analysis - volume and surface area
by Laura Willers de Souza 21 Feb '22

21 Feb '22

External Email - Use Caution Hello FreeSurfer Developers, I would like to know if it is correct to perform volume and surface area analysis using freesurfer's GLM!? In several mailing list questions I have already seen instructions for performing these types of analysis. But I understand that freesurfer does vertex-wise analysis in the 2D surface space. So, it is not clear to me how it is possible to evaluate a 3D measurement, as volume, with a 2D method!? I've also read that surface area analysis via vertex-wise analysis doesn't make much sense. The best would be to do ROI-based analysis. Can you explain that to me? Thanks in advance! Laura Willers de Souza Master Student in Biochemistry Zimmer Lab - Federal University of Rio Grande do Sul - Porto Alegre/RS - Brazil

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