Freesurfer

freesurfer@nmr.mgh.harvard.edu

17 participants
30099 discussions

Best Way To Edit?
by Lindsay Gubernick 22 Aug '14

22 Aug '14

Hello, I was just looking for some suggestions for the best way to manually edit an image such as the one attached. As you can see, there is both white and gray matter excluded from the pial surface on the medial aspect of the right hemisphere. Is this best corrected by adding voxels to the brainmask.mgz, adding a new control set, both, or something else entirely? Thank you very much for the help. Lindsay Gubernick [image: Inline image 1].

1 1

Left and Right Vessels
by Mohammed Goryawala 22 Aug '14

22 Aug '14

Hello Freesurfer experts, In a study I found significant differences in the volume of the left and right vessels of the freesurfer atlas. I wanted to know what regions are included in that ROI and if there are any known functions of these regions. Thank you Mohammed Goryawala, PhD Miller School of Medicine, University of Miami

3 4

mean cortical thickness on N ROIs
by elisa veronese 21 Aug '14

21 Aug '14

Dear FreeSurfers, suppose I should estimate the mean cortical thickness on a number N of ROIs (for instance, the left inferior temporal, the left middletemporal, and the left superior temporal). How would you suggest me to proceed? - Just performing a simple average of the 3 cortical thickness values obtained with the aparcstats2table? - Computing a weighted average of the 3 cortical thickness values obtained with the aparcstats2table? In this case, what would you use as weight? The area of the 3 regions? Thank you. Best, Elisa -- Elisa Veronese - PhD *Research Unit on Brain Imaging and Neuropsychology (RUBIN)* *Inter-University Center for Behavioral Neurosciences (ICBN)University of Udine and University of Verona, ItalyICBN Website: http://icbn.uniud.it/tiki-index.php <http://icbn.uniud.it/tiki-index.php>*

2 1

Fwd: Re: Custom aseg Color Lookup Table
by C.P. Frost 21 Aug '14

21 Aug '14

Thanks for the reply, Doug. The main problem is that areas not included in my custom table appear colored in freeview. These areas' numbers correctly correspond to their indices on the default FreeSurferColorLUT, although they don't appear labeled in the info panel, nor are they listed to the left of the main display. For instance, the left cerebral cortex (not on my list) is colored, and clicking on it yields an index of 3 in the info panel, but no label. The good news is so far as I can tell, every area I specified does show up colored appropriately and has a label associated, but I have no idea where the ones I DIDN'T specify are coming from. -------- Original Message -------- Subject: Re: [Freesurfer] Custom aseg Color Lookup Table Date: Wed, 20 Aug 2014 11:12:56 -0400 From: Douglas N Greve <greve(a)nmr.mgh.harvard.edu> To: freesurfer(a)nmr.mgh.harvard.edu Try loading aseg.mgz as the aux volume so you can see the number at a particular voxel. If the color is not correct, see what the number is and what it corresponds to in the LUT doug On 08/19/2014 05:57 PM, C.P. Frost wrote: > Hi all, > > I'm trying to use Freeview to display subcortical regions in aseg.mgz > with a custom color lookup table. RGB colors will indicate on a jet > scheme how well that region's volume correlates to performance on a > behavioral measure. In other words, an R^2 coefficient of -1 is dark > blue, 0 is green, and 1 is dark red. I adapted the approach for aseg > structures from this method: > http://brainder.org/2011/07/05/freesurfer-brains-in-arbitrary-colours/ > > When I try right now, by e.g. loading > subjects_dir/average/mri/aseg.mgz, and choosing Color map > Lookup > Table, Lookup Table > Load lookup table... and selecting my custom > table file, it *almost* works. The file is copied below (it has no > headers). > > 4 Left-Lateral-Ventricle 92 255 163 0 > 5 Left-Inf-Lat-Vent 224 255 31 0 > 7 Left-Cerebellum-White-Matter 112 255 143 0 > 8 Left-Cerebellum-Cortex 133 255 122 0 > 10 Left-Thalamus-Proper 153 255 102 0 > 11 Left-Caudate 102 255 153 0 > 12 Left-Putamen 61 255 194 0 > 13 Left-Pallidum 102 255 153 0 > 14 3rd-Ventricle 0 224 255 0 > 15 4th-Ventricle 61 255 194 0 > 16 Brain-Stem 71 255 184 0 > 17 Left-Hippocampus 102 255 153 0 > 18 Left-Amygdala 41 255 214 0 > 26 Left-Accumbens-area 92 255 163 0 > 28 Left-VentralDC 122 255 133 0 > 30 Left-vessel 102 255 153 0 > 31 Left-choroid-plexus 102 255 153 0 > 43 Right-Lateral-Ventricle 173 255 82 0 > 44 Right-Inf-Lat-Vent 122 255 133 0 > 41 Right-Cerebellum-White-Matter 82 255 173 0 > 42 Right-Cerebellum-Cortex 122 255 133 0 > 49 Right-Thalamus-Proper 173 255 82 0 > 50 Right-Caudate 143 255 112 0 > 51 Right-Putamen 92 255 163 0 > 52 Right-Pallidum 41 255 214 0 > 53 Right-Hippocampus 133 255 122 0 > 54 Right-Amygdala 20 255 235 0 > 58 Right-Accumbens-area 82 255 173 0 > 60 Right-VentralDC 122 255 133 0 > 62 Right-vessel 122 255 133 0 > 63 Right-choroid-plexus 143 255 112 0 > 72 5th-Ventricle 92 255 163 0 > 85 Optic-Chiasm 112 255 143 0 > 251 CC_Posterior 71 255 184 0 > 252 CC_Mid_Posterior 31 255 224 0 > 253 CC_Central 10 255 245 0 > > The index numbers were chosen to match those found in the default aseg > LUT, as listed on the FS wiki: FSTutorial > AnatomicalROI > > FreeSurferColorLUT. Some regions appear correctly, but others either > a) display with incorrect labels, particularly > Right-Cerebellum-White-Matter or Right-Cerebellum-Cortex on *cortical* > white matter and ribbon, or b) display with no label in the info > panel, which makes it unclear whether it's working. > > > Notes: > - Originally indices were simply 1 to n, but that caused a wider array > of problems, and I inferred (perhaps incorrectly) freeview was > matching regions with indices rather than label names. > -This particular map is mostly green since most correlations are > small, which is fine; others will have a broader range of values. > -I'm not sure what the effect is of pressing the "Show existing labels > only" toggle, which eliminates a couple areas from the label list, but > it's unclear how it chooses them. > > > Can anyone offer advice on how to ensure the correct colors are > displayed in the correct regions? (Or an easier way to do this?) > > Cheers, > C.P. Frost > UW-Madison > > > _______________________________________________ > 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/

2 1

head surface
by Meryem Ayse Yucel 21 Aug '14

21 Aug '14

Hi freesurfer community, I am trying to get faces and vertices of the head surface mesh in Matlab. freesurfer_read_surf works perfectly when I extract faces and vertices for the pial surface (from lh.pial), but I get an "unknown magic number in surface file" error when I use the same function for lh.seghead (which I assume is the head surface?). Any help on what I am doing wrong? Thanks! Meryem

2 1

GLM: Error Matric-Scale
by Robby De Pauw 21 Aug '14

21 Aug '14

Hi there FreeSurfer When trying the GLM-analysis I keep getting the following error. Could you help me with fixing it? ERROR: matrix is ill-conditioned or badly scaled, condno = 1e+08 -------------------------------- Possible problem with experimental design: Check for duplicate entries and/or lack of range of continuous variables within a class. If you seek help with this problem, make sure to send: 1. Your command line: mri_glmfit --y lh.whiplash.thickness.10B.mgh --fsgd whiplash.fsgd dods --C group_diff.mtx --surf fsaverage lh --cortex --glmdir lh.whiplash.glmdir 2. The FSGD file (if using one) 3. And the design matrix above Thank you in advance, Robby Robby De Pauw, dra. Ghent University Department of Physiotherapy and Rehabilitation Sciences 3B3 De Pintelaan 185 B-9000 Ghent Robby.DePauw(a)Ugent.be

2 2

FLAIR Registration Problems
by Zamroziewicz, Marta Karolina 21 Aug '14

21 Aug '14

Hello, This email is related to the archived “bbregister and FLAIR” email chain posted by Kayle Sawyer on 5/23/2014 (http://www.mail-archive.com/freesurfer%40nmr.mgh.harvard.edu/msg36117.html) We have also been registering with FLAIR scans, and have run into a problem with approximately half of our subjects. Recon-all completes successfully but a bad registration step leaves our MPRAGE and FLAIR scans misaligned. I’ve attached two screenshots, the first of which shows a bad registration and the second of which shows a good registration. Per your previous suggestion, we ran bbregister using the following flag: bbregister --s SLOTS007 --mov flair.mgz --init-header --reg reg.test.dat --t2 --lta reg.test.lta This did result in a good registration, and we are hoping to incorporate this into our processing pipeline to ensure that we can use the FLAIR for refining the pill surface. The problem that we’ve run into is that we aren’t sure at which point this bbregister step occurs within the recon-all process flow. We were thinking there are a couple of possibilities: 1. manually create the transformation matrix for the FLAIR using the command mentioned above, and then re-run whichever steps necessary in recon-all to include this manually created xform; 2. apply the manually created xform to the mri/orig/FLAIRraw.mgz image to create a new, properly aligned mri/FLAIR.mgz, and then re-run recon-all with whatever steps/flags we need to include this FLAIR image to refine the pial surface. Unfortunately we're not exactly sure which flags to include in recon-all in order to most efficiently re-process our data with the manually created FLAIR transformation, and we were hoping you might offer some advice on this point. Please let me know if I can provide any additional information. Thanks for your time and assistance! [cid:D837E275-197C-46BB-A1D4-B7559DF58D48@beckman.uiuc.edu][cid:D62F65DB-AF9A-4ADC-A596-79E0821DE24D@beckman.uiuc.edu] Marta Zamroziewicz MD/PhD Candidate Decision Neuroscience Laboratory University of Illinois Urbana-Champaign mzamro2(a)illinois.edu<mailto:mzamro2@illinois.edu> 815-980-3882

2 5

FreeSurfer platform/performance questions
by Lazarevich, Alexander 21 Aug '14

21 Aug '14

Hello FreeSurfer mailing list, First off, I am not a scientist and I have no experience running FreeSurfer. I have installed it, and seen it run, but that's it. I work for the IT department in the Beckman Institute, part of the University of Illinois. We support a computer lab here called the "Visualization Laboratory" - aka Vislab. (http://itg.beckman.illinois.edu/visualization_laboratory/) The Vislab lab has dozens of very nice workstations in it. (16 x 3GHz Xeon cores, 128GB of mem, NVIDIA Quadro K6000 video, etc.) In short, the Vislab supports research which needs high end computing. (less power than supercomputers, but much better than normal desktops and much easier to access/use than supercomputers) The Vislab has hundreds of active users, students, faculty, postdocs, etc., all of whom do some kind of scientific research. One of the packages we have recently installed upon request is FreeSurfer. We have installed freesurfer-Linux-centos6_x86_64-stable-pub-v5.3.0.tar.gz<ftp://surfer.nmr.mgh.harvard.edu/pub/dist/freesurfer/5.3.0/freesurfer-Linux…> on our Red Hat Enterprise Linux 6 workstations, as well as on our Windows 7 Enterprise 64bit workstations. We have also installed Oracle's VirtualBox on a Windows 7 Enterprise 64bit machine, and run the VM freesurfer-Virtualbox-linux-x86-stable-pub-v5.3-full.vdi.gz<ftp://surfer.nmr.mgh.harvard.edu/pub/dist/freesurfer/5.3.0/freesurfer-Virtu…> inside that virtual environment. The researchers who requested FreeSurfer are all part of the same group, and they are trying to complete a very large research project which requires FreeSurfer to process large amounts of data. They have normally been running FreeSurfer on the VM in the XUbuntu VirtualBox environment on their own average desktop PC's. However, they are concerned about their processing time with FreeSurfer, so they are hoping to take advantage of the Vislab computing resources to decrease the processing time in FreeSurfer. We are running into the following problems/concerns: 1) The FreeSurfer VM (Xubuntu) that runs in VirtualBox is a 32bit host. Is there a 64bit VM available? That 32bit host can only access 4GB of memory, whereas they would like to be able to access more of the workstations 128GB of memory. 2) The users have expressed concern about switching platforms to Red Hat Enterprise Linux6 with the FreeSurfer freesurfer-Linux-centos6_x86_64-stable-pub-v5.3.0.tar.gz<ftp://surfer.nmr.mgh.harvard.edu/pub/dist/freesurfer/5.3.0/freesurfer-Linux…> installed. They have told us that FreeSurfer on XUbuntu might give different results than if it were run on Red Hat Enterprise Linux 6, given the same dataset. They are concerned about switching platforms in the middle of the research study, and that it might contaminate results. Is this possible? Or should the same version of FreeSurfer provide the same results, even with a different platform? 3) Does anyone else on this list run very large FreeSurfer processes on high end 64bit computers? If a high-end processing method works well for you, I would love to hear about it so I can share it with the researchers here. Thanks for your time! ____________ Alex Lazarevich 5321 Beckman Institute - 217-244-1565 Information Technology Services - http://its.beckman.illinois.edu<http://its.beckman.illinois.edu/> Beckman Institute - http://beckman.illinois.edu<http://beckman.illinois.edu/> University of Illinois - http://illinois.edu<http://illinois.edu/>

3 6

no improvement for recon-all result after edits
by Jidan Zhong 21 Aug '14

21 Aug '14

Hi Freesurfer experts, I was doing some manual editing on our subjects which are patients with lesions. I managed to correct most of them, however, 4 subjects have the same problem that I can't fix no matter what edits I tried. So, the 4 subjects, they all have the same problem, which is: frontal part of superior temporal gyrus region is not segmented as white matter, thus they are just "missing" from the final surfaces and final segmentations. But when just do eye check, there is nothing wrong with this part, even the intensity looks the same of other white matter parts.Its just cut suddenly. When I went back to those subjects, first I tried with control points. - It didnt help. I confirmed that white matter is present connected to the main brain white matter volume. -- I reran from recon-all autorecon2-wm, didnt help. Then I checked the aseg.mgz, that part is segmented as white matter correctly. I even modified the aseg.mgz (although not suggested), it still didn't work It is just weird that the surface was just cut from the middle of STG. So, is there any other way I can tell FreeSurfer that this is really STG and please keep it? Thanks!! -- Regards, Jidan

3 3

Creating lower resolution surface models
by Nicholas Heugel 21 Aug '14

21 Aug '14

I am running analysis where I am taking fMRI data using Freesurfer to project it onto the cortical surface and load it into Brainstorm to use as a constraint on the source reconstruction. I am able to do all the steps just fine, the problem I am encountering is that Freesurfer creates a surface with 200,000+ vertices, for the analysis I am running I downsample it in brainstorm to 15,000 vertices to reduce run time. However, when the label file for Freesurfeer gets down sampled it significantly changes how the scouts look. Is there a way to have Freesurfer create a surface with 15000 vertices and then project the volumetric fMRI onto that surface so I don't have to downsample the data? I know there is the function fmris_decimate, which should allow me to decimate targeted files. But I am not sure if I can then use mdi_vol2surf to put the fMRI data onto the decimated surface files, or if I need to have the entire surface file generated from the beginning with only 15,000 vertices. Can I use the recon-all function to define the number of vertices I want, or is that a variable I can't adjust? Any help in the matter would be appreciated. Nicholas

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