Freesurfer

freesurfer@nmr.mgh.harvard.edu

18 participants
30131 discussions

missing values, aseg
by Eszter Boros 12 Dec '22

12 Dec '22

External Email - Use Caution Dear FreeSurfer Team, I am writing because I was extracting the aseg values using Freesurfer 7.1.1 and much to my surprise, these values were 'empty', or zero: 5th-Ventricle WM-hypointensities Left-WM-hypointensities Right-WM-hypointensities non-WM-hypointensities Left-non-WM-hypointensities Right-non-WM-hypointensities What could be the reason behind this? How can I get these values? I used the following command: asegstats2table --subjects subject1 subject2 (... etc) --common-segs --meas volume --stats=aseg.stats --table=aseg.volume.txt Thank you for your help in advance! And let me know if you need any more information. All the best, Eszter Boros

2 1

Fusion of PET and T1 weighted
by Amar Shukla 12 Dec '22

12 Dec '22

External Email - Use Caution Dear Freesurfer Community. I want to do image fusion (PET+T1) = T' weighted image. Then I want to go for the feature extraction from this T1 modality which I will get from the fusion.So how i can go for this. Questions are: 1) Do I have to use PET Surfer for this? Does Pet Surfer support the fusion of the two different modality. 2) I am in the perception that I want PET+ T = T' , or does it justify the multimodality approach for the detection of disease. 3) If i extract features from the different modality and then i combine it. Can we say that this is the multimodality approach of Image Fusion. Please advise me and provide the approach or the tool which will help me for the multimodality. -- With Regards, Amar Shukla Assistant Professor, M tech AI ANN(2014-2016), Centre Of Information Technology, UPES Contact :+91 897 977 8016

2 5

FreeSurfer equivalent to fslsplit?
by Wighton, Paul 12 Dec '22

12 Dec '22

Hey FreeSurfer aficionados, Is there a FreeSurfer equivalent to fslsplit? If I have a 4d nifti, is there a way to split that into a set of 3d niftis from the command line? Thanks, -Paul

3 3

outputs for recon-all of T1-weighted and intersubject registration
by Zeng, Qi 12 Dec '22

12 Dec '22

External Email - Use Caution Hi Freesurfer experts, If I just want to normalize, register, and segment the T1-weighted images to the GM, WM, and CSF tissues, can certain steps be skipped from recon-all, such as the inflation process? What is the output for the segmentation of GM, WM, and CSF tissues? is aseg.mgz or aparc+aseg.mgz better? Following that, if I want to apply an affine transformation for all subjects to a common template, is the following command line correct? > mri_convert aparc+aseg.mgz aparc+aseg.nii.gz > bet aparc+aseg.nii.gz my_structural.nii.gz > flirt -ref ${FSLDIR}/data/standard/MNI152_T1_2mm_brain -in my_structural.nii.gz -omat my_affine_transf.mat > fnirt --in=aparc+aseg.nii.gz --aff=my_affine_transf.mat --cout=my_nonlinear_transf --config=T1_2_MNI152_2mm > applywarp --ref=${FSLDIR}/data/standard/MNI152_T1_2mm --in=aparc+aseg.nii.gz --warp=my_nonlinear_transf --out=my_warped_structural Best, QZ

2 1

Are inflated surfaces vertices transformed in Freeview?
by Sarah Lavelle 12 Dec '22

12 Dec '22

External Email - Use Caution Hi Doug, Thank you for your help! Yes, I have followed the points using vertex indices but there were some that didn't match up. It could be because of the 1-based indexing. Where would I change this? I am using freesurfer_read_surface.m to read in the vertices and faces : function [vertices, faces] = freesurfer_read_surf(fname) % freesurfer_read_surf - FreeSurfer I/O function to read a surface file % % [vertices, faces] = freesurfer_read_surf(fname) % % Reads the vertex coordinates (mm) and face lists from a surface file. % Copyright (C) 2000 Darren L. Weber % History: 08/2000, Darren.Weber_at_radiology.ucsf.edu % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % There used to be a line here where ver was set to the version of % the file. Without it, "ver" reverts to a matlab defined structure % and generates an error. %fprintf('FREESURFER_READ_SURF [v %s]\n',ver(11:15)); if(nargin < 1) help freesurfer_read_surf; return; end %QUAD_FILE_MAGIC_NUMBER = (-1 & 0x00ffffff) ; %NEW_QUAD_FILE_MAGIC_NUMBER = (-3 & 0x00ffffff) ; TRIANGLE_FILE_MAGIC_NUMBER = 16777214; QUAD_FILE_MAGIC_NUMBER = 16777215; % open it as a big-endian file fid = fopen(fname, 'rb', 'b'); if (fid < 0), str = sprintf('could not open surface file %s.', fname); error(str); end fprintf('...reading surface file: %s\n', fname); tic; magic = freesurfer_fread3(fid); if (magic == QUAD_FILE_MAGIC_NUMBER), Nvertices = freesurfer_fread3(fid); Nfaces = freesurfer_fread3(fid); fprintf('...reading %d quad file vertices\n',Nvertices); vertices = fread(fid, Nvertices*3, 'int16') ./ 100 ; if (nargout > 1), fprintf('...reading %d quad file faces (please wait)\n',Nfaces); faces = zeros(Nfaces,4); for iface = 1:Nfaces, for n=1:4, faces(iface,n) = freesurfer_fread3(fid) ; end if(~rem(iface, 10000)), fprintf(' %7.0f',iface); end if(~rem(iface,100000)), fprintf('\n'); end end end elseif (magic == TRIANGLE_FILE_MAGIC_NUMBER), fprintf('...reading triangle file\n'); tline = fgets(fid); % read creation date text line tline = fgets(fid); % read info text line Nvertices = fread(fid, 1, 'int32'); % number of vertices Nfaces = fread(fid, 1, 'int32'); % number of faces % vertices are read in column format and reshaped below vertices = fread(fid, Nvertices*3, 'float32'); % faces are read in column format and reshaped faces = fread(fid, Nfaces*3, 'int32'); faces = reshape(faces, 3, Nfaces)'; else str = sprintf('unknown magic number in surface file %s.', fname); error(str); end vertices = reshape(vertices, 3, Nvertices)'; fclose(fid); fprintf('...adding 1 to face indices for matlab compatibility.\n'); faces = faces + 1; t=toc; fprintf('...done (%6.2f sec)\n\n',t); return ii) Also, I want to transform a section of the inflated surface into a *2D flatmap*, and import it into *MATLAB* while *maintaining the same VertexIDs* so I know which vertex on the flatmap corresponds to a vertex on a pial surface. Do you know how I can do this? I have created multiple flatmaps in Freeview by : 1. Creating a patch on lh.inflated in Freeview and saving it 2. Flattening the patch using mris_flatten() in terminal 3. Loading lh.inflated in Freeview and then loading the flattened patch which then creates a 2D flatmap However, I cannot figure out how to load this 2D flatmap into MATLAB. I've tried saving the result in (3) as a surface and then loading it into matlab with freesurfer_read_surf.m. This takes a long time and the result looks a bit distorted. It also does not conserve the VertexIDs. Do you have any recommendations? Many thanks again for your help, Sarah

2 1

Pial surface does not follow the gray boundary
by 白晨 12 Dec '22

12 Dec '22

External Email - Use Caution Hello FreeSurfer Developers, I'm attempting to analyze the T1 imaging of children aged 2–7 years, a problem with the segmentation occurred,the pial surface of many subjects did not follow the gray matter boundary, and large area of gray matter in right hemisphere was excluded from pial surface (as shown in pictures), we used control points to extend white matter in order to extend pial surface, but can't fully resolve the problem, Is there any other solution? I will be very grateful if you can give meadvice. Also, I've attached the files of quality checking(Internal surface and external surface) and recon-all.log of a subject in case it's of any use. My setup is the following: FreeSurfer version: freesurfer-Linux-centos6_x86_64-stable-pub-v6.0.0-2beb96c platform: Ubuntu 20.04.4 LTS uname -a: Linux shenkang-B560M-AORUS-ELITE 5.15.0-53-generic #59~20.04.1-Ubuntu SMP Thu Oct 20 15:10:22 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux (T1 voxel size is 0.5mm*1mm*1mm) Any help would be greatly appreciated! With kind regards, Chen Bai

2 1

Creating and analyzing combined ROIs, combining different FreeSurfer versions
by Trolle, Carl 12 Dec '22

12 Dec '22

Dear FreeSurfer team, 1. We're trying to focus our cortical analysis following suggestions here https://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg23939.html. The FS ROI tutorial page mentions that labels should be drawn or registered on the fsaverage template; could we then break the fsaverage aparc_annot atlas and combine the fsaverage labels of interest into our new label or should we break each individual subjects aparc_annot atlas and then create the combined ROI from the subject specific labels? I'm assuming that we either way will have to resample the combined label to each subject using mri_label2label? 2. When running the mri_glmfit-sim analysis later on, the thread above mentions not using the --cache option. Can we use "--mczsim 3 pos" instead for a one-sided test? 1. We've had some success in correcting incorrect parcellations using the mris_register and mri_ca_label functions. As we had some issues with mris_register in 7.1.1, we used the developer version for these steps, but 7.1.1 when re-running recon-all as this is the version we used for our study. Is it ok to combine versions like this? Thank you for your time! Kind regards, Carl Trolle Research fellow Spaulding Neuroimaging Lab Platform: linux-centos7_x86_64-7.1.1-20200723-8b40551 FreeSurfer version: 7.1.1

2 1

ERROR: fast_mat2vol
by stdp82＠virgilio.it 12 Dec '22

12 Dec '22

External Email - Use Caution I would extract the rs-fMRI timeseries from Yeo’s Network components. I’m running: preproc-sess -s ${1} -fsd rest -sliceorder up -surface self lhrh -fwhm 0 -per-runfcseed-config -wm -fcname wm.dat -fsd rest -pca -cfg wm.config fcseed-sess -s ${1} -cfg wm.configfcseed-config -vcsf -fcname vcsf.dat -fsd rest -pca -cfg vcsf.configfcseed-sess -s ${1} -cfg vcsf.config mkanalysis-sess -analysis fc.noseed.surf.lh -surface self lh -fwhm 0 -notask -nuisreg vcsf.dat 5 -nuisreg wm.dat 5 -polyfit 5 -nskip 4 -mcextreg -fsd rest -TR 1.67 -per-run -hpf 0.01 -lpf 0.1 -nowhiten mkanalysis-sess -analysis fc.noseed.surf.rh -surface self rh -fwhm 0 -notask -nuisreg vcsf.dat 5 -nuisreg wm.dat 5 -polyfit 5 -nskip 4 -mcextreg -fsd rest -TR 1.67 -per-run -hpf 0.01 -lpf 0.1 -nowhiten The line below produced this error: selxavg3-sess -s ${1} -a fc.noseed.surf.lh -no-con-ok -svres Saving rho1 ERROR: fast_mat2vol: szvol inconsistent with mat2d nv = 151008, size(mat,2) = 142954 run 2 t= 1.6reading data ... 0.518631Temporally filteringSaving rho1 ERROR: fast_mat2vol: szvol inconsistent with mat2d nv = 151008, size(mat,2) = 142954Found 0 voxels with corrected AR1 > 0.90Not WhiteningFound 8054 zero-valued voxelsComputing contrasts The res folder in fc.noseed.surf.lh or fc.noseed.surf.rh contains res-001.nii.gz 158KB res-002.nii.gz 158KB that cannot be opended. Stefano

2 1

pial surface error: ?h.pial surface excludes part of gray matter/brainmask
by Alessia Sepe 12 Dec '22

12 Dec '22

External Email - Use Caution Hi FreeSurfer Developers, I have a major problem with the reconstruction of the pial surface. The issue is the opposite to the one relate to the brainmask: the reconstruction excludes part of the gray matter. In attachment an example of the output after running -autorecon2-wm -noaseg, and the recon-all.log if needed. Does anyone have any thoughts on how to fix it? FreeSurfer version: v5.3.0 up to 'recon-all -s <subjid> -autorecon2-inflate1 -noaseg -cc-crs <x> <y> <z> -pons-crs <x> <y> <z>', v6.0.0 for manual editing on the wm and 'recon-all -s <subjid> -autorecon2-wm -noaseg -cc-crs <x> <y> <z> -pons-crs <x> <y> <z>'. Thank you. Best wishes, Alessia Sepe MSc | PhD Fellow KU Leuven Laboratory for Neuro- and Psychophysiology Department of Medicine ON II Herestraat 49, 3000 Leuven

2 1

Effect size
by Lyna Mariam El Haffaf 12 Dec '22

12 Dec '22

External Email - Use Caution Hi! I'm taking the liberty of writing another email on this subject, as I didn't get a reply to my previous email. My colleagues and I performed a general linear model with FreeSurfer mri_glmfit for whole brain analysis in intergroup comparison (g2v1) with a Monte-Carlo correction. I'd like to get the effect size of the analysis and I was wondering how could I get the eta squared or R2? I also saw that the gamma.mgh file was the contrast effect size. Could this be used as an effect size for this type of analysis? I believe an equation such as : pseudo_rsquared = (1 - model.llf / model.llnull) could be helpful but I wonder how to include this in the mri_glmfit script. Thank you in advance! Lyna El Haffaf University of Montreal

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