Freesurfer May 2017

freesurfer@nmr.mgh.harvard.edu

145 participants
219 discussions

Re: [Freesurfer] Worse determination of ?h.white with v6.0 in comparison to v5.3 - worse GM/WM contrast
by Antonin Skoch 02 May '17

02 May '17

Doug, I tried to replicate again v5.3 results of subject 1. I checked wm.mgz, orig.mgz, ?h.orig, ?h.pial, ?h.white and found everything identical (by using mri_diff and mris_diff) except FLAIR.mgz and ?h.pial (after FLAIRpial refinement). The difference was small, I have traced it to the level of bbregister, namely to flirt.fsl. I cannot trace the exact version of flirt used in the original data, neither cannot check contents of xyztrans.sch, but I suspect that this is the reason of difference. My current version of flirt reads: FLIRT version 6.0. I have uploaded the replicated subject 1 as file subj_1_v5.3_replicate_2.tar.gz to your server (fist upload attempt of file subj_1_v5.3_replicate_2.tar.gz was aborted). I have added also current xyztrans.sch. Antonin Doug, did you process from scratch or did you retain edits to brainmask.mgz and wm.mgz?The -cubic should not be necessary, since recon-all (which I primarily obtained from https://surfer.nmr.mgh.harvard.edu/pub/dist/freesurfer/5.3.0-patch/ has UseCubic=1 ). I could try to replicate v5.3 again on my side. Antonin It is subject #1, trying to replicate in our version of 5.3 using the scripts/recon-all.local-copy ./recon-all.local-copy -all -FLAIRpial -s 1.dng.v53.local-cubic -cubicThe results were close, but they were not exact On 5/1/17 4:13 PM, Antonin Skoch wrote: Dear Doug, that is strange. What precisely you cannot replicate? The results with v5.3 or with v6.0? What subject from the group I uploaded you have tried? I could try to run the comparison again but I have seen the results already in many subjects and the difference between -cubic and no cubic seems profound and systematic in favor of v5.3 in the aspect of wm.mgz leak ouside brain and GM/WM contrast. Concerning the expert options file: Despite using flag -xopts-use in recon-all in v5.3, I did not use any expert option file in the subject I have uploaded. The reason of using -xopts-use was that I was processing the subjects in batch where some of them had expert-options file with entry bbregister -init-header due to the fact that init-fsl failed in these subjects. I wanted to make my life easier by processing all of them by identical recon-all command line parameters and added -xopts-use to all subjects (even in the subjects without expert option file). I supposed that this could not do any harm. Antonin I can't seem to replicate your results locally, even with the recon-all you used. The one thing I'm missing is the expert options file. Can you send that to me? On 04/24/2017 12:49 PM, Antonin Skoch wrote: > Dear Doug, > > the subject with leak of white surface outside brain (my first post with > screenshots) is subject 1. Slice number (coronal) around 100. > The subject in second post (with text below) is subject 2, slice number > (coronal) also 100. > > I have processed the subjects with v 6.0 (in fact dev version from feb 2017, > but this is irrelevant) with -cubic -no-mprage. It looks much like v5.3, i.e. > the wm.mgz and surfaces are much better, but v5.3 looks still better, at > least for subject 1. > I have uploaded them as file v6.0_cubic_no_mprage.tar.gz to your ftp site. > > The optical difference in norm.mgz/brain.mgz between v5.3 and v6.0 with > -cubic is very minor, but still there is some other thing which renders > wm.mgz worse than with v5.3 for subject 1. > The -cubic has profound effect, the images seem much smooth with lose of > contrast without using -cubic. > > Regards, > > Antonin > > > And what slice number? > On 04/24/2017 11:16 AM, Douglas N Greve wrote: > > Anonin, of the three subjects you sent, which one is shown in these > > pictures? > > > > > > On 04/19/2017 05:23 PM, Antonin Skoch wrote: > >> Dear experts, > >> > >> I am sending just one more example to illustrate issue with white > >> surface estimation in v6.0. See the attached screenshots: In v6.0 > >> there seems to be insufficient contrast in brain.finalsurfs.mgz, so > >> the white surface is leaking at three spots dramatically outwards > >> towards pial surface. The white surface in v5.3 looks much more > >> anatomically relevant in the same spot. > >> > >> Could you please comment on how to avoid such issues in v.6.0? > >> > >> Regards, > >> > >> Antonin Skoch

1 0

Re: [Freesurfer] Skull stripping
by Antonin Skoch 02 May '17

02 May '17

Dear John, you can prepare expert options file with the entry: mri_gcut -T <threshold value> and run recon-all -skullstrip -clean-bm -gcut -subjid <subject name> -expert expert_opts_file.txt The mri_gcut help page reads: -T <value> set threshold to value (%) of WM intensity, the value should be >0 and <1; larger values would correspond to cleaner skull-strip but higher chance of brain erosion. Default is set conservatively at 0.40, which provide approx. the same negligible level of brain erosion as 'mri_watershed'. Antonin Skoch Dear FS experts, I want to inquire if there are any threshold for the flag "gcut " in the command:recon-all -skullstrip -clean-bm -gcut -subjid <subject name> If this not available are there any tools that can help to control how much dura will be cut out Thanks in advance! John

1 0

[FreeSurfer] cortical parcellation
by VA Research 01 May '17

01 May '17

Does freesurfer output individual roi volume masks for each of the cortical parcellations in aparc+aseg? thanks Joseph Veliz VA Hospital West LA

3 3

Re: [Freesurfer] Worse determination of ?h.white with v6.0 in comparison to v5.3 - worse GM/WM contrast
by Antonin Skoch 01 May '17

01 May '17

Dear Doug, that is strange. What precisely you cannot replicate? The results with v5.3 or with v6.0? What subject from the group I uploaded you have tried? I could try to run the comparison again but I have seen the results already in many subjects and the difference between -cubic and no cubic seems profound and systematic in favor of v5.3 in the aspect of wm.mgz leak ouside brain and GM/WM contrast. Concerning the expert options file: Despite using flag -xopts-use in recon-all in v5.3, I did not use any expert option file in the subject I have uploaded. The reason of using -xopts-use was that I was processing the subjects in batch where some of them had expert-options file with entry bbregister -init-header due to the fact that init-fsl failed in these subjects. I wanted to make my life easier by processing all of them by identical recon-all command line parameters and added -xopts-use to all subjects (even in the subjects without expert option file). I supposed that this could not do any harm. Antonin I can't seem to replicate your results locally, even with the recon-all you used. The one thing I'm missing is the expert options file. Can you send that to me? On 04/24/2017 12:49 PM, Antonin Skoch wrote: > Dear Doug, > > the subject with leak of white surface outside brain (my first post with > screenshots) is subject 1. Slice number (coronal) around 100. > The subject in second post (with text below) is subject 2, slice number > (coronal) also 100. > > I have processed the subjects with v 6.0 (in fact dev version from feb 2017, > but this is irrelevant) with -cubic -no-mprage. It looks much like v5.3, i.e. > the wm.mgz and surfaces are much better, but v5.3 looks still better, at > least for subject 1. > I have uploaded them as file v6.0_cubic_no_mprage.tar.gz to your ftp site. > > The optical difference in norm.mgz/brain.mgz between v5.3 and v6.0 with > -cubic is very minor, but still there is some other thing which renders > wm.mgz worse than with v5.3 for subject 1. > The -cubic has profound effect, the images seem much smooth with lose of > contrast without using -cubic. > > Regards, > > Antonin > > > And what slice number? > On 04/24/2017 11:16 AM, Douglas N Greve wrote: > > Anonin, of the three subjects you sent, which one is shown in these > > pictures? > > > > > > On 04/19/2017 05:23 PM, Antonin Skoch wrote: > >> Dear experts, > >> > >> I am sending just one more example to illustrate issue with white > >> surface estimation in v6.0. See the attached screenshots: In v6.0 > >> there seems to be insufficient contrast in brain.finalsurfs.mgz, so > >> the white surface is leaking at three spots dramatically outwards > >> towards pial surface. The white surface in v5.3 looks much more > >> anatomically relevant in the same spot. > >> > >> Could you please comment on how to avoid such issues in v.6.0? > >> > >> Regards, > >> > >> Antonin Skoch

2 1

R: R: R: Re: R: Re: Hipposubfields in FreeSuerferColorLUT.txt
by stdp82＠virgilio.it 01 May '17

01 May '17

Dear list and Eugenio, I have run for each subject mri_label2vol --label SUBJ/label/rh.perirhinal_exvivo.thresh.label --regheader SUBJ/mri/norm.mgz --fill-ribbon --subject SUBJ --hemi rh --o SUBJ/mri/rh.perirhinal.nii.gz next, to use FS-FAST on perihrinal.nii.gz, I have run fcseed-config -seg SUBJ/mri/lh.perirhinal.nii.gz -fcname L_Perirhinal.dat -fsd rest -mean -cfg mean.L_Perirhinal.config This error occurs: ERROR: no segmentations IDs (-segid) or ROI (-roi) specified What's should I do? Thanks Stefano On Mon, 06 Feb 2017 13:59:38 -0800 Do you mean to use them as seeds in FSFAST? You'll need to map them into the conformed space using mri_label2vol with the --regheader option (see the --help); view the volume to make sure it looks ok. Then run fcseed-config specifying the volume with -seg and the code with segid ----Messaggio originale---- Da: stdp82(a)virgilio.it Data: 6-feb-2017 22.27 A: "Freesurfer support list"<freesurfer(a)nmr.mgh.harvard.edu> Ogg: [Freesurfer] R: Re: R: Re: Hipposubfields in FreeSuerferColorLUT.txt Thank you very much.Could you provide me any suggestion to extract CA1 left or CA1 right?And how can I use them in FAST? By codes?Regards,Stefano ----Messaggio originale---- Da: "Iglesias Gonzalez, Eugenio" <e.iglesias(a)ucl.ac.uk> Data: 6-feb-2017 17.42 A: "stdp82(a)virgilio.it"<stdp82(a)virgilio.it>, "Freesurfer support list"<freesurfer(a)nmr.mgh.harvard.edu> Ogg: Re: [Freesurfer] R: Re: Hipposubfields in FreeSuerferColorLUT.txt You would extract CA1 left from lh.hippoSfLabels-T1.v10.mgz and CA1 right from rh.hippoSfLabels-T1.v10.mgz (assuming you’ve used T1s to segment the subfields). Cheers, /E Juan Eugenio Iglesias ERC Senior Research Fellow Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/ On 6 Feb 2017, at 16:17, stdp82(a)virgilio.it wrote: Thanks Eugenio for your response. My apologies but this matter is unclear to me. If I would use results for FS-FAST how I can identify for example CA1 right from CA1 left? Additionally, for enthorinal which is the correct code in FAST? Stefano ----Messaggio originale---- Da: "Iglesias Gonzalez, Eugenio" <e.iglesias(a)ucl.ac.uk> Data: 6-feb-2017 15.54 A: "stdp82(a)virgilio.it"<stdp82(a)virgilio.it>, "Freesurfer support list"<freesurfer(a)nmr.mgh.harvard.edu> Ogg: Re: [Freesurfer] Hipposubfields in FreeSuerferColorLUT.txt Ciao Stefano, The results for the left and right hippocampi are in different files, so there’s no ambiguity despite the fact that they indeed share the same label codes. Cheers, /Eugenio Juan Eugenio Iglesias ERC Senior Research Fellow Translational Imaging Group University College London http://www.jeiglesias.com http://cmictig.cs.ucl.ac.uk/ On 6 Feb 2017, at 13:18, stdp82(a)virgilio.it wrote: Hi list, within FreeSuerferColorLUT.txt, do the numbers between 200-226 descrived the hippo-subfield outputs? Additionally, how can I obtain separate right and left masks? Thanks Stefano _______________________________________________ Freesurfer mailing list Freesurfer(a)nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail. _______________________________________________ Freesurfer mailing list Freesurfer(a)nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.

1 0

Re: [Freesurfer] Using multiple MEMPRAGE scans as input to recon-all: which echos to use for dura correction?
by Jared Saletin 01 May '17

01 May '17

Thanks Bruce! That’s really good to know. So therefore, no need to have a “compromise” set of echo images for dura correction if I’m not feeding both inputs to recon-all? Would it make sense to run both days separately and just average thickness values between ROIs to reach a single value for each subject? I guess this question is slightly complicated by having functional data on both days as well. I was debating which structural scan (post-freesurfer) to use as my registration base for the functional data. It seems I have two options: use the scan collected the same day as the functional — or use a single structural for both days (hence the impulse to feed both scans to a single recon-all. I suppose a third option is just pick the scan with best quality and go with that for everything. Do you have a gut feeling on which is the better strategy? Best, Jared > Hi Jared > > it's really an empirical question. With high quality data we have found > that adding scans doesn't help CNR very much (it reduced the N but the > blurring induced by differential distortion also reduces the C). If your > scans are noisy it might be worth it, but it really depends on the > acquisition > > cheers > Bruce > > > On Fri, 28 Apr 2017, Jared > Saletin wrote: > > > Hi Freesurfer Experts! > > > > I have two scans for each subject in a new experiment—taken on subsequent days, each using the MEMPRAGE protocol. > > > > I’ve successfully implemented DURA correction using the 4 echos when I’ve reconstructed each scan separately. > > > > However, since I have 2 scans for each subject, I was thinking rather than reconstruct the brain twice and average the thickness measurements, it would be better to feed both MEMPRAGE scans to recon-all. > > > > My question is if I do this, which echos should I use for Dura correction: day 1 or day 2 or is there a way to combine them into an average? > > > > Is this (feeding both days’ scans to recon-all) a reasonable strategy, or am I better to reconstruct the data separately? > > > > Thanks so much! > > Jared > > _______________________________________________ > > Freesurfer mailing list > > Freesurfer at nmr.mgh.harvard.edu <https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer> > > https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer <https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer> > > — Jared M. Saletin, PhD Assistant Professor of Psychiatry and Human Behavior Alpert Medical School of Brown University E.P. Bradley Sleep Research Lab 300 Duncan Drive Providence, RI 02906 Email: jared_saletin(a)brown.edu Phone: 401-421-9440 Fax: 401-453-3578

2 1

Neuroscience of Language Postdoctoral Position at the University of California, San Francisco (UCSF) in California, USA
by Roeland Hancock 01 May '17

01 May '17

Neuroscience of Language Postdoctoral Position at the University of California, San Francisco (UCSF) in California, USA The Hoeft Neuroscience Lab (http://brainLENS.org PI: Fumiko Hoeft MD PhD) at the UCSF Dept of Psychiatry and Weill Institute for Neurosciences is looking for a full-time postdoctoral fellow for the ‘Intergenerational Neuroimaging of Language and Reading Networks’ project, which uses a ‘natural’ cross-fostering family design to partially dissociate the effects of genetics, prenatal, and postnatal environment on the human neural systems that underlie language processing. Related articles on this approach in different domains can be found here: Yamagata et al. J Neurosci 2016 (http://goo.gl/vMK8iy) Ho et al. Trends in Neurosci 2016 (http://goo.gl/SyXLcK) and Scientific American (http://goo.gl/YTiH6D) The successful candidate must have a PhD in Cognitive Psychology, Cognitive Neuroscience, Cognitive Science, Linguistics or a closely related field, experience in neuroimaging, excellent foundational skills in computation, programming and mathematics/statistics, and exceptional organizational skills. Candidates with experience in neurolinguistics or behavioral genetics are particularly encouraged to apply. Compensation is in accordance with the NIH pay scale for postdoctoral fellows. Appointments are made for 1 year terms with the possibility of renewal. Postdoctoral fellows are represented by UC Postdoc Union and enjoy competitive benefit packages. The Hoeft Neuroscience Lab is an interdisciplinary research environment that applies cutting-edge neuroimaging techniques, genetic analysis, and computational approaches to examine language development processes, the etiology of neurodevelopmental disorders and the impact of diverse linguistic environments (e.g. bilingual education) on development. The successful candidate will have the opportunity to be involved in other research projects in the Laboratory, outreach and community engagement projects, and access to a wealth of archival developmental and longitudinal neuroimaging data on language and literacy. The position can begin immediately. Applications should be emailed to brainlens(a)ucsf.edu with “[Postdoc job]” and the applicant's full name in the subject of the email. Applications must include: 1. A current CV 2. A one page statement of research interests related to the proposed project 3. A one page cover letter describing qualifications for the position with specific examples, career goals and why you feel you are a good fit for the lab. 4. The names of three references. Qualified candidates will be asked to have letters of reference forwarded. Do not submit reference letters with the application. The University of California, San Francisco is a premier biomedical research institution, ranked second in the world for Neuroscience and Behavior by US News. UCSF has outstanding resources for cognitive neuroscience research, including a research-dedicated Siemens Prisma 3T MRI scanner in addition to shared 3T and 7T GE scanners, MEG and TMS equipment. The laboratory enjoys close collaborations with other research groups in the Bay Area, and national and international collaborations. The San Francisco Bay Area is a vibrant, multi-cultural city, ranked one of the most walkable cities in the US. In addition to the area’s outstanding cultural attractions and cuisine, the unique beauty of redwood forests, coastal cliffs, wine country and natural preserves is less than an hour from the city.

1 0

Re: [Freesurfer] Worse determination of ?h.white with v6.0 in comparison to v5.3 - worse GM/WM contrast
by Antonin Skoch 01 May '17

01 May '17

Dear Doug, the subject with leak of white surface outside brain (my first post with screenshots) is subject 1. Slice number (coronal) around 100. The subject in second post (with text below) is subject 2, slice number (coronal) also 100. I have processed the subjects with v 6.0 (in fact dev version from feb 2017, but this is irrelevant) with -cubic -no-mprage. It looks much like v5.3, i.e. the wm.mgz and surfaces are much better, but v5.3 looks still better, at least for subject 1. I have uploaded them as file v6.0_cubic_no_mprage.tar.gz to your ftp site. The optical difference in norm.mgz/brain.mgz between v5.3 and v6.0 with -cubic is very minor, but still there is some other thing which renders wm.mgz worse than with v5.3 for subject 1. The -cubic has profound effect, the images seem much smooth with lose of contrast without using -cubic. Regards, Antonin And what slice number? On 04/24/2017 11:16 AM, Douglas N Greve wrote: > Anonin, of the three subjects you sent, which one is shown in these > pictures? > > > On 04/19/2017 05:23 PM, Antonin Skoch wrote: >> Dear experts, >> >> I am sending just one more example to illustrate issue with white >> surface estimation in v6.0. See the attached screenshots: In v6.0 >> there seems to be insufficient contrast in brain.finalsurfs.mgz, so >> the white surface is leaking at three spots dramatically outwards >> towards pial surface. The white surface in v5.3 looks much more >> anatomically relevant in the same spot. >> >> Could you please comment on how to avoid such issues in v.6.0? >> >> Regards, >> >> Antonin Skoch

2 1

Hippo-subfiled Need Help
by Ramesh Babu 01 May '17

01 May '17

Dear Experts, I ave done hippocampal subfield segmentation by using 1x1x1 t1 scan. Please see the attached image and give your suggestions. It took nearly 7 hrs 10 min. to complete hippo-subfield segmentation for one scan by using the command recon-all -i grp205_t1.nii -s grp205 -sd grp205 -all Started at Wed Apr 26 11:03:34 IST 2017 Ended at Wed Apr 26 17:17:35 IST 2017 #@#%# recon-all-run-time-hours 6.234 recon-all -s grp205 finished without error at Wed Apr 26 17:17:35 IST 2017 done Then for hipposubfield, recon-all -s grp205 -sd grp205 -hippocampal-subfields-T1 I want to run for 40 scans grp-1 20 scans and grp-2 20 scans. How to modify in the command. Since it is taking long time I want to run 10 scans n give rest again 10 scans. Can I do so? I am15.6 GiB ram, Intel® Core™ i7-4790 CPU @ 3.60GHz × 8 processor, GeForce GT 610/PCIe/SSE2 graphic card.It is octa core processor. Can I run 8 analysis simultaneously (in 8 separate terminal?). Your suggestion will be helpful to me to proceed further. Thanks RB

2 3

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 May 2017