Freesurfer May 2014

freesurfer@nmr.mgh.harvard.edu

139 participants
206 discussions

brain.finalsurfs.mgz editing
by Khadka, Sankalpa 02 May '14

02 May '14

Dear freesurfer community, My question is directed to those who have experience with editing brain.finalsurfs.mgz. According to the tutorials, after making edits on brainmask.finalsurfs.mgz (which is normally done when the pial surface is not accurate), we can run, recon-all -autorecon-pial -subjid subj" However, I found an anomaly when I looked at the steps that -autorecon-pial flag goes through. The very first step of autorecon-pial flag is "mri_mask" command. This takes brain.mgz and brainmask.mgz as inputs and gives brain.finalsurfs.mgz as output. Now, if we have made edits on brain.finalsurfs.mgz, it will be overwritten by the new one and hence our edits won't take an effect. I ran few experiments too. First, I made edits on brain.finalsurfs.mgz and ran recon-all with autorecon-pial flag. I ended up with a new brain.finalsurfs.mgz and my edits did not make any changes. Second, I made edits on brain.finalsurfs.mgz and ran with white, smooth2, inflate2, autorecon3 flag (basically starting from mri_make_surfaces command). On doing so, the edits made desirable changes and brain.finalsurfs.mgz remained intact. Third, I made changes on brainmask.mgz and brain.mgz (inputs for mri_mask). It also works as expected. Finally, I also followed the recommended steps from the tutorial by creating brainmask.finalsurfs.manedit.mgz, however, the edits from "manedit" were not carried on to brainmask.finalsurfs.mgz Could anyone please shed some light on this? What would be the best strategy to follow? Thank you, Sankalpa

2 1

brain.finalsurfs.manedit.mgz file edits & longitudinal processing
by Douglas Merkitch 02 May '14

02 May '14

Dear Freesurfer experts, I have a question regarding longitudinal edits. Specifically the brain.finalsurfs.manedit.mgz file and the longitudinal processing stream in Freesufer 5.1 on a Mac (freesurfer-i686-apple-darwin9.8.0-stable5-20110525). The problem that I am having is that the edits saved in the brain.finalsurfs.manedit.mgz file do not carry over when processing longitudinally. Previous posts have addressed this topic, but have not (to my knowledge) solved the issue. Please see the threads from the Freesurfer archive below for a brief context: 1. http://www.mail-archive.com/freesurfer%40nmr.mgh.harvard.edu/msg27695.html 2. http://www.mail-archive.com/freesurfer%40nmr.mgh.harvard.edu/msg33471.html Please let me know if more information is needed to understand the issue. Any insight/ideas regarding this issue would be greatly appreciated! Thanks, Doug Doug Merkitch Neurological Sciences Rush University Medical Center Phone: (312) 563-3853<tel:%28312%29%20563-3853> Fax: (312) 563-4660<tel:%28312%29%20563-4660> Email: Douglas_Merkitch(a)rush.edu<mailto:Douglas_Merkitch@rush.edu>

2 1

FSGD and contrast files for group analysis
by Kenza Drissi 02 May '14

02 May '14

Dear Marie,I want to run a group analysis on a set of subjects, correcting for gender(Male/Female), diagnois regarding epilepsy (no focus/focus in rh/focus inlh/focus in both h), age and IQ. Should my FSGD file look somewhat like thisbelow:GroupDescriptorFile 2Class NoFocus_MaleClass NoFocus_FemaleClass RhFocus_MaleClass LhFocus_MaleClass BhFocus_MaleClass RhFocus_FemaleClass LhFocus_FemaleClass BhFocus_FemaleVariables age IQIf so, considering that I want to investigate the effect of epilepsy on corticalgyrification, what should my contrast matrix look like?The hypothesis is that the gyrification index is diminished for epilepticpatients, probably more when there is bilateral focus.I know that we need as many number as regressors, so I expect there to be 10numbers, is it so?Thank you so very much for clarifying this point,Kenza

2 1

effect size for conjunction analyses?
by Caspar M. Schwiedrzik 02 May '14

02 May '14

Hi! I am doing some conjunction analyses (Nichols et al.) on my fMRI data, and I was wondering whether anyone can make a recommendation how to display effect sizes for this kind of analyses? Would you show the effect size estimate corresponding to the p-value that is chosen, an average, or anything else? Thanks, Caspar

2 1

MICCAI PC workshop -- May 16 (fwd)
by Bruce Fischl 02 May '14

02 May '14

FYI ---------- Forwarded message ---------- Date: Fri, 2 May 2014 11:28:11 -0400 From: Polina Golland <polina(a)csail.mit.edu> To: Bruce Fischl <fischl(a)nmr.mgh.harvard.edu> Subject: MICCAI PC workshop -- May 16 Hi Bruce, Can you please forward this around? Thanks, Polina. MICCAI PC Workshop May 16, 2014, 1-6pm, CSAIL Kiva and Star Conference rooms (32-D463, 32-G449). MICCAI program committee members will present their work in areas of Medical Image Computing and Computer Assisted Interventions. This is a great opportunity to hear about current research of 20 MICCAI groups from all around the world. Open to public; anyone interested in the topics of MICCAI is welcome to attend. Tentative Program can be found here: http://miccai2014.org/pc_workshop.html.

1 0

per voxel regressor in LME toolbox?
by Marnie Shaw 02 May '14

02 May '14

Dear Freesurfers, Is there a way to include a per voxel regressor in an LME model within the matlab LME toolbox? Just like you can do with mri_glmfit? Cheers, Marnie

3 4

Multi-echo FLASH
by Jon Houck 02 May '14

02 May '14

Hi all, I'm building a scan protocol for what is primarily a MEG study, which needs T1 information for co-registration and T1/T2 information to construct an accurate BEM from the inner skull surface. It seems like the two multi-echo FLASH sequences at 5 and 30 degrees would give me all of the information that I need without an MPRAGE, but I'm not sure how well the two MEFs work in the standard Freesurfer pipeline. Would it be reasonable to acquire no MPRAGE but two MEF sequences at 5 and 30 degrees, combine them following the method from https://surfer.nmr.mgh.harvard.edu/fswiki/MEF, and send the resulting T1 to recon-all? With my scan time constraints, the alternative would be to acquire a multi-echo MPRAGE and a single 5-degree MEF. Thanks for any help on what is probably a naïve question. Jon Houck

2 2

Re: [Freesurfer] using multiple cores
by Lukas.Scheef＠ukb.uni-bonn.de 02 May '14

02 May '14

Hi Bruce! Thanks for the fast response. I used the current stable version 5.3.0.1 (freesurfer-Darwin-lion-stable-pub-v5.3.0.1). So I will try as you have suggested ... Best wishes, Lukas > Bruce Fischl Fri, 02 May 2014 05:47:26 -0700 > > Hi Luke > > what version of FS are you running? We see significant speedups. Try > running top in a unix terminal and seeing if the cpu usage of > mri_ca_register (during autorecon2, the longest single step) is morethan 100% > > cheers > Bruce > > On Fri, 2 May 2014, lukas.sch...(a)ukb.uni-bonn.de wrote: > > > > Hi folks! > > > > I try to use -openmp option on an iMAC with 4 cores. Unfortunately, I do not > > see any accelartion when processing a single data set with or without the > > flag -openmp flag set. > > > > recon-all -s ernie_1 -all > > > > leads to ~6,4h processing time. > > > > Unfortunately > > recon-all -s ernie_2 -all -openmp 4 > > > > did not change anything. Do I miss anything? Do I have to prepare my system > > in a special way? Or run any scripts in advance? > > > > (System: iMac, 32GbRam,Intel i5, 4 cores, Mac Os 10.9, latest FS version) > > > > Best wishes, > > > > Luke

1 0

using multiple cores
by Lukas.Scheef＠ukb.uni-bonn.de 02 May '14

02 May '14

Hi folks! I try to use -openmp option on an iMAC with 4 cores. Unfortunately, I do not see any accelartion when processing a single data set with or without the flag -openmp flag set. recon-all -s ernie_1 -all leads to ~6,4h processing time. Unfortunately recon-all -s ernie_2 -all -openmp 4 did not change anything. Do I miss anything? Do I have to prepare my system in a special way? Or run any scripts in advance? (System: iMac, 32GbRam,Intel i5, 4 cores, Mac Os 10.9, latest FS version) Best wishes, Luke

2 1

Hippocampal subfields on 1.5 Tesla
by Alan Francis 01 May '14

01 May '14

Hi Bruce and FreeSurfers: I have received a manuscript to review for possible publication. The authors have used the subfields algorithm on 1.5T scans and obtained a parcellation with values. They have drawn some major conclusions on the basis of the findings. My understanding is that this method can only be done on 3T. Is the 1.5T results valid? Please advice. thanks, Alan Francis

3 6

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 2014