Freesurfer

freesurfer@nmr.mgh.harvard.edu

16 participants
30085 discussions

mris_volmask ERROR
by Guang Zeng 08 Dec '09

08 Dec '09

Hi, there, I submit several subjects to FS piepline using our Sun Grid Engine System, all of them failed at the same place which is the mris_volmask step in autorecon3. It gave me an error like: ERROR: In /usr/pubsw/packages/vtk/5.2.0/src/Common/vtkMultiThreader.cxx, line 372 vtkMultiThreader (0x33d05a90): Unable to create a thread. pthread_create() returned 12 I knew this error is caused by some VTK function FS used to allocate memory. So what I did is I use recon-all -s subjectid , and the process will finish. What happen is the same subjects have been run on our local machines, and finished without any problems. However, when I looked at the wmparc.stats file of the same subject, the results are quite different. For example, ctx_isthumus cingluate and ctx_posterior singulate value are not zeros for my local machine results, while ctx_isthumus cingluate and ctx_posterior singulate value are zeros for my SGE machine results. And this problem have to all these subjects. Another funny thing is if I look at the ?h.aparc.stats file, the results from SGE and local machine are the same, while when I load the results to tksurfer, they are quite different. So my question is: Since I need use SGE to run multiple subjects again, next time when I met this mris_volmask ERROR, I need rerun the process from the beginning, not from autorecon3, right? Thank! Guang _________________________________________________________________ Chat with Messenger straight from your Hotmail inbox. http://www.microsoft.com/windows/windowslive/hotmail_bl1/hotmail_bl1.aspx?o…

2 1

optseq post-hoc sorting
by Kathrine Shepherd 08 Dec '09

08 Dec '09

Hello-- I'm designing an event-related task that does not have predefined conditions. instead, I plan to do some post-hoc sorting based on performance and stimuli characteristics. I expect that after sorting, I will have 5 conditions- but it could be more or less. Is it advisable to determine the stim presentation schedules in optseq as if I had already specified predefined conditions? Or is it optimal to just completely randomize the jitter without making any assumptions about the response function? Also is there a tool available to estimate the stat power of my schedules after I've defined them? Many thanks for your advice, -Kathrine

4 4

fixed versus random effects
by Kathrin-Sophie Herbst 08 Dec '09

08 Dec '09

Hi, for my second-level statistics I have been running a fixed-effects analysis as well as a random-effects-a. with mri_glmfit (11 subjects). It turns out that the fixed-effect-analyis shows hardly any effect while in the random-effects-analysis there are results (at a low threshold: 1.6 or 2). Vision versus rest looks fine in both and gets even stronger in the fixed effects. Is there a statistical reason for stonger outcomes in random-effects for results around the threshold or is this a hint for an error in my analysis? Thanks, Kathrin my command was: mri_glmfit --y rh.ces.nii --osgm --glmdir rh.ffx.osgm --nii --mask ../rh.mask.nii --yffxvar rh.cesvar.nii --ffxdofdat ../ ffxdof.dat --surf fsaverage rh my computer: snapjudge my path: autofs/space/bomba_020/users/ksherbst/data/brightghosts_loc/ loc_group_11 and I am running freesurfer stable 4

2 1

Registration for Upcoming Freesurfer Course now open, Feb. 1 & 2
by Allison Stevens 08 Dec '09

08 Dec '09

Hello, The Freesurfer group will be offering a two-day Freesurfer Tutorial and Workshop on February 1st and 2nd at the Martinos Center in Charlestown, MA. A description of the February course and a list of the topics covered can be found on the Course wiki page here: http://surfer.nmr.mgh.harvard.edu/fswiki/CourseDescription On that website, you will find a link to register for the course. There is only room for 28 total participants and registration is on a first-come, first-serve basis. Shortly after all spots are filled, information will be sent out about hotels and parking. If you would like to arrange those details now, you can find all the information that will be sent out on the Course wiki page: http://surfer.nmr.mgh.harvard.edu/fswiki/CourseDescription THINGS TO KNOW BEFORE YOU REGISTER: 1. The registration AND payment deadline is January 8, 2010. 2. fMRI analysis is not a component that is taught in this course. 3. Payment can only be done in the form of a check or if you are from within the Partners network, a fund transfer can be arranged. Credit Cards are *not* accepted. 4. There are no refunds if you cannot attend the course. You can put the tuition payment towards the next course or send someone else in your spot. 5. All questions about payment should go to Stacey Ladieu (sladieu(a)nmr.mgh.harvard.edu) Before e-mailing Stacey, please read the Course wiki page (listed above) and the registration page, carefully. Most questions participants have are answered there. 6. If the course is full, please register to be placed on the waiting list *only* if you think you could attend the course on short notice if a spot were to open up. Regards, The FreeSurfer Team

1 0

centOS vs. ubuntu
by dahlia＠nmr.mgh.harvard.edu 07 Dec '09

07 Dec '09

Hi, Will Freesurfer and MNE work on CentOS 5 (I see in the Download page that CentOS 4 is supported, but I understand that most workstations at the center are CentOS 5)? Also, since I need to install the OS myself, and I understand that ubuntu is easier in that respect, I'm wondering whether these two packages will work under ubuntu, and what other advantages/disadvantages I should expect. Thanks for any advice!!! Dahlia.

2 6

a question about the mris_volmask ERROR
by Guang Zeng 07 Dec '09

07 Dec '09

Hi, there, I have question about the mris_volmask ERROR. I ran multiple subjects on our SUN Grid Engine system. sometimes some subjects are sent to the same cluster node, and limited memory is assigned to each subject. Because of this, I often met an ERROR at the mris_volmask --label_left_white 2 --label_left_ribbon 3 --label_right_white 41 --label_right_ribbon 42 --save_ribbon --save_distance sub1 step in autorecon3., which gave me an error like: ERROR: In /usr/pubsw/packages/vtk/5.2.0/src/Common/vtkMultiThreader.cxx, line 372 vtkMultiThreader (0x33d05a90): Unable to create a thread. pthread_create() returned 12 I knew this error is caused by some VTK function FS used to allocate memory. What I did is everytime when I met this kind of error, I use recon-all -s subj -autorecon3 -no-isrunning , and the process will finish. However, I found if I did something like this, it will cause problems at the wm parcellation step. For example, I have the same subject which is sent to SGE, fails at mris_volmask, and then I call autorecon3 to finish the process, and I go to wmparc.stats, lotf of rois such as isthumus ctx-cingluate, ctx-posterior singulate are 0. While I ran the same subject at a local machine, and finished without any problems, I looked at isthumus cingluate and posterior singulate, they are not zeros. So, does it mean that I need start from beginning when I met the errors cause by mris_volmask? Thanks a lot! guang _________________________________________________________________ Chat with Messenger straight from your Hotmail inbox. http://www.microsoft.com/windows/windowslive/hotmail_bl1/hotmail_bl1.aspx?o…

1 0

recon-all exited with errors
by Dana W. Moore 07 Dec '09

07 Dec '09

Hi everyone, I had a recon-all exit with errors, but I can't find a specific error message that explains what the problem is. This is v4.5.0. The last several lines of the script are pasted below: Any ideas as to what went wrong? Thanks, Dana gca peak Unknown = 0.94427 ( 0) gca peak Left_Inf_Lat_Vent = 0.22551 (39) gca peak CSF = 0.23520 (22) gca peak Left_Accumbens_area = 0.50011 (60) gca peak Left_VentralDC = 0.15663 (92) gca peak Left_undetermined = 1.00000 (35) gca peak Left_vessel = 0.65148 (62) gca peak Left_choroid_plexus = 0.09084 (48) gca peak Right_Inf_Lat_Vent = 0.30937 (35) gca peak Right_Hippocampus = 0.24721 (68) gca peak Right_Amygdala = 0.32542 (80) gca peak Right_Accumbens_area = 0.34035 (62) gca peak Right_vessel = 0.83418 (60) gca peak Right_choroid_plexus = 0.10189 (48) gca peak Fifth_Ventricle = 0.45329 (25) gca peak WM_hypointensities = 0.14318 (83) gca peak non_WM_hypointensities = 0.09346 (54) gca peak Optic_Chiasm = 0.34849 (76) mri_ca_label GCA sequential renormalization: label 28 not consistently computed. not using caudate to estimate GM means setting label Left_Hippocampus based on Right_Hippocampus = 1.00 x + 0 setting label Left_Amygdala based on Right_Amygdala = 1.00 x + 0 setting label Right_VentralDC based on Left_VentralDC = 1.15 x + 0 estimating mean gm scale to be 1.00 x + 0.0 estimating mean wm scale to be 1.00 x + 0.0 estimating mean csf scale to be 0.97 x + 0.0 Right_Pallidum too bright - rescaling by 0.949 (from 1.050) to 101.6 (was 107.1) saving intensity scales to aseg.auto_noCCseg.label_intensities.txt Linux compute-3-14.v4linux 2.6.18-53.el5 #1 SMP Wed Oct 10 16:34:19 EDT 2007 x86_64 x86_64 x86_64 GNU/Linux recon-all exited with ERRORS at Fri Dec 4 22:20:24 EST 2009 Dana W. Moore, Ph.D. Neuropsychology Fellow Cornell Neuropsychology Service Weill Medical College of Cornell University New York Presbyterian Hospital Department of Neurology & Neuroscience 428 East 72nd Street, Suite 500 New York, NY 10021 Phone: 212-746-2441 Fax: 212-746-5584 Email: dwm2003(a)med.cornell.edu

1 0

GM-wm error
by Ritobrato Datta 07 Dec '09

07 Dec '09

Hello All, I ran recon-all for a subject and when I did the following to check the gray matter and white matter segmentation I found that it couldnt do it. tkmedit subjid brainmask.mgz lh.white -aux T1.mgz -aux-surface rh.white Please see attached fig. All slices are like this one. Can anyone suggest how do I fix it ? Waiting for your reply Thank you Ri

3 2

Longitudinal processing with variable time points?
by Dankner, Nathan (NIH/NIMH) [F] 07 Dec '09

07 Dec '09

Hello all, I am hoping to do some longitudinal processing on a dataset. However, the intervals between time points for each subject are not uniform. Can freesurfer's longitudinal processing stream handle this, or do all of the time points need to be equal?

2 1

Re: [Freesurfer] "dummy variable" in mri-glmfit
by Stefan Ehrlich 05 Dec '09

05 Dec '09

Hi Doug and Nick, thanks for your quick answer. I think that the assumption of linearity can often be made in genetics and in large GWAS studies the coding is done as suggested by Nick. However, Nick's solution does not suit this particular case since one of the homozygous groups has less than 10 subjects which, in my opinion, might skew our results. Therefore we would like to lump homozygotes and the carriers together and have 2 groups (instead of 3). Unfortunately I and Stefan Brauns do not understand your reply, Doug. We were suggesting to include the binary variable for genotype as a "variable" ( covariate) instead of a "class" (factor) in the FSGD. This would enable us to get around the problem of "small cell sizes". Why do you say this is not possible with an FSGD? We would just specify Variables age SNP and SNP would be 0 and 1 instead of 0, 1 and 2 (as suggested by Nick) This is how the header would look like: GroupDescriptorFile 1 Title rs8216888_status MeasurementName thickness Class SCZMALEMGH Class SCZFEMALEMGH Class HCMALEMGH Class HCFEMALEMGH Class SCZMALEIA Class SCZFEMALEIA Class HCMALEIA Class HCFEMALEIA Class SCZMALEUMN Class SCZFEMALEUMN Class HCMALEUMN Class HCFEMALEUMN Class SCZMALEUNM Class SCZFEMALEUNM Class HCMALEUNM Class HCFEMALEUNM Variables age SNP In contrast, if we would include it as a class - we would have 32 instead of 16 classes and then "Variables age" The question is if it violates some assumptions if we specify a "variables" (covariate) which is binary. Many thanks, Stefan Message: 6 Date: Wed, 02 Dec 2009 13:18:20 -0500 From: Douglas N Greve <greve(a)nmr.mgh.harvard.edu> Subject: Re: [Freesurfer] "dummy variable" in mri-glmfit To: Stefan Brauns <stefan.brauns(a)googlemail.com> Cc: freesurfer <freesurfer(a)nmr.mgh.harvard.edu> Message-ID: <4B16AF6C.9080103(a)nmr.mgh.harvard.edu> Content-Type: text/plain; charset=UTF-8; format=flowed Do you mean having just another column in your design matrix with 0s and 1s? You can do this, but not with an FSGD. You'll have to supply your own matrix. An easy way to do this would be to run mri_glmfit with and FSGD without the genotype. This will create a matrix Xg.dat in the output dir, then just modify that matrix and pass it to a new call to mri_glmfit doug Stefan Brauns wrote: > Hi there, > > we would like to test the effect of a binary variable (genotype > = carrier vs. homozygous) on cortical thicknes in mri-glmfit. Since we > are also controlling for gender and aquisition site (4 sites) we > already have 16 groups. In order to control for age as a covariate we > need at least 2 subjects per group to be able to estimate an age slope. > > If we include the aforementioned binary variable (genotype) as a > factor (two different "groups"), we would have 32 groups and > unfortunately not enough subjects per group. > > Is it possible to include binary variables ("dummy variable" coded as > 0 and 1) such as genotype or gender as covariates (slope), in order to > reduce the number of groups and examine the effect on thickness? In > simple regression this would not affect the results - what would we > expect here? > > Many thanks, > > Stefan >

3 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