Freesurfer December 2006

freesurfer@nmr.mgh.harvard.edu

49 participants
59 discussions

unpacksdcmdir & multiple volume protocols
by Sebastian Moeller 21 Dec '06

21 Dec '06

Hi Doug, hi List, I ran into slight problems with unpacksdcmdir. I acquired some combined proton density and T2 volumes with one sequence (plus a bunch of additional slices from Inline/generic tab on out allegra menue), which resulted in mri_convert choling at the import of the second part of the series. In other words the proton density volume was imported fine, but the T2 and the MIP images did not make it into the target directory. Now I sit on a bunch of these things, where I am really interested in the T2s, but all freesurfer imports are the PDs (whichg are nice to have). So is this a known prblem with work arounds or is this a case of "don't do that then"? On a related note, I tried mri_convert to convert bfloat stacks into 3d analyze (from 1 stack per slice containing all time points, to one stack per timepoint containing all slices) while using --in_orientation (to correct for our inability to select the proper patient orientation); and lo, all that was converted was the first timepoint (I did not specify any specific frame, and was therefore expecting all time points to appear). Interestingly, doing the format conversion first and the "--in_orientation" thing later on the .img works like a charm, I am now puzzled whether this is the right behaviour. I am more than willing to test anything out or send test images if that helps... Anyway, thanks for the nice tools and the great support Ahoi Sebastian -- Sebastian Moeller Tel.: 04 21 - 2 18 - 78 38 oder 96 91 Fax.: 04 21 - 2 18 - 90 04 GSM: 01 62 - 3 25 45 59 moeller(a)brain.uni-bremen.de AG Kreiter / FB 2 Institut fuer Hirnforschung III Abteilung Theoretische Neurobiologie Universitaet Bremen Biogarten Hochschulring 16a Postfach 33 04 40 28359 Bremen

1 0

mri_annotation2label
by Ran Tao 20 Dec '06

20 Dec '06

Hi All, When I ran mri_annotation2label --subject UNC_4 --hemi rh --labelbase ./labels/aparc-rh --table $FREESURFER_HOME/surface_labels.txt Some labels are always skipped such as 6,7... I also tried on tutorial data, the same thing happened. Any ideas? Is surface_labels.txt a complete label list I should use? I also want to add some own labels(not extract them from annot), and cannot figure it out. How can I do this? Thanks in advance, Ran

1 0

surfaces
by Daniel G. Wakeman 20 Dec '06

20 Dec '06

Hi freesurfers, I have a question. I deleted some of the optic nerve on a participant and ran source /usr/local/freesurfer/nmr-std-env cd /space/doodles/3/users/dwakeman/pilot/again setenv SUBJECTS_DIR `pwd` recon-all -autorecon2-wm -autorecon3 -subjid p001 -mail dwakeman When the surfaces were regenerated the optic nerve was still present in the surfaces. Can someone please take a look at this and tell me what I'm doing wrong? This has repeated on several other participants; however, I need a couple more edits on those participants. Thanks, Dan

1 0

Can Tkmedit use a certain parcellated cortical area?
by Bum Seok Jeong 20 Dec '06

20 Dec '06

Hi! I'd like to see a certain parcellated cortical area, for example, right inferior frontal gyrus (Rt IFG).., using Tkmedit. After that, I'd like to use a certain cortical area, Rt IFG, as ROI which could be used in measuring functional activity, fa value so on. Is it possible? Thank you for advance. bsjeong

1 0

faster mri_parse_sdcmdir?
by Keith Schneider 20 Dec '06

20 Dec '06

I've been using mri_parse_sdcmdir for a long time to convert my dicom files to nifti or analyze. It takes 10 minutes or more on my G5 Mac for an entire data session. Some other Mac tools like Osirix and Neurolens can read in all of the dicom files and parse them in a few seconds. It seems like mri_parse_sdcmdir spends most of its time parsing the dicom files and reading the headers. Is there some way to speed this up? I'm not complaining, just inquiring. There are some programmers here that may be able to help. The source for Neurolens is available, for instance. I found another program on the web, dinifti that seemed like it was going to be pretty fast, but then it died with a segmentation fault. keith

2 1

NIfTI format conversion
by Kirk, Gregory 19 Dec '06

19 Dec '06

Hi All , I am using mri_convert to convert Siemens 30 direction diffusion data into NIfTII for proc. with fsl. mri_convert does it well except that it is a bit problematic that it comes out in neurological convention. Also the tools is fsl do not seem capable to change the orientation to radiological convention. I want to use a structural cooregistered to the diffusion b0 to navigate for placing seeds for stochastic tract processing. mri_convert takes my Siemens structurals and puts them in NIfTI files that are oriented completely wrong, upside down etcetera. Obviously mri_convert knows how to read Siemens structurals ... it's does it all the time for freefurfer processing. So I take analyze 7.5 format structurals and it comes out in radiological convention and fsl tools are incapable to change these into neurological convention ! so a bit of a sticky wicket, of course it would be nice if fsl would support it's own tools for data inport, but as far as I know they do not. NIfTI code from COX et. al. is well.... as he puts it himself Klingon software, Klingon's that fell on their head anyway. it is not organized well enough to spend the time to see if anything usable is there. I guess the q-form or s-form are not set such that the fsl tools can work with 7.5 or mri_convert produced files. any way to get correctly oriented NIfTI structurals or to reorient the diffusion data to radiological. cheerio Greg.

3 3

MRInormalize: could not find any valid peaks
by Juergen Haenggi, Psychiatrische Uni-Klinik 19 Dec '06

19 Dec '06

Dear Freesurfer users While running recon-all, the following error occured: #-------------------------------------------- #@# Intensity Normalization Sat Dec 16 13:10:39 CET 2006 /scratch/bli/AN1792/subjects/Patient02_2002/mri mri_normalize -g 1 nu.mgz T1.mgz using max gradient = 1.000 reading from nu.mgz... normalizing image... talairach transform 0.796 -0.256 -0.057 -8195.100; 0.086 0.398 -0.583 12578.642; 0.224 0.596 0.440 21656.441; 0.000 0.000 0.000 1.000; INFO: Modifying talairach volume c_(r,a,s) based on average_305 MRInormalize: could not find any valid peaks mri_normalize: normalization failed Linux node0214 2.6.15.6-id-x86_64-k8-smp #1 SMP Wed Mar 22 12:35:20 CET 2006 x86_64 x86_64 x86_64 GNU/Linux recon-all exited with ERRORS at Sat Dec 16 13:10:41 CET 2006 If I try to check the talairach.xfm with tkregister2, I see only a little part of the brain in the left upper corner of the graphic window. Because the DICOM files used were converted from ANALYSE files, I guess that there is a problem with the coordinate origin of the DICOMs. Could it be that this is the cause of my problem? The transformation matrix seems also not OK. Because I have about 250 brains which is the fastest way to fix that problem? Thanks in advance Best regards Juergen ----------------------------------------------- Juergen Haenggi, Ph.D. student Psychiatric University Hospital Division of Psychiatry Research University of Zurich, Switzerland PO Box 1931 Lenggstrasse 31, 8032 Zurich 0041 44 384 26 10 (office phone) 0041 76 445 86 84 (mobile phone) 0041 44 384 26 86 (fax) H 115 (office room number) juergen.haenggi(a)bli.unizh.ch (division email) http://www.dpr.unizh.ch/ (division website) http://www.juergenhaenggi.ch (private website) -----------------------------------------------

2 1

glibc
by skenny＠uchicago.edu 18 Dec '06

18 Dec '06

hi all, i'm trying to do a freesurfer install on a cluster that is using glibc-2.2.5 and the latest freesurfer requires glibc- 2.3 libraries. i'm wondering if the source is available so that i can compile for use with glibc-2.2.5, or alternatively if someone might suggest another workaround for this (?) thanks sarah

1 0

Scan quality & FS
by Derin Cobia 14 Dec '06

14 Dec '06

Some of the scans I work with are of the same subject, but differ only in the length of time they were in the scanner. Obviously the scans of longer length (everything else being equal) will yield better quality; however, I'm interested in a kind of "bang-for-your-buck" approach. Is there a point where some amount of diminishing returns occurs with respect to FS processing and scan quality? Therefore, my main question: 1) Is there anything in the FS processing stream that would be a reasonable heuristic for determining if one scan processes better than another (all things being equal beside length of time being scanned)? For example, # of topological defects to correct, length of processing, amount of manual correction, or just visual quality? I understand some of this is related to what structure(s) I'm interested in measuring, but was wondering about it in a general fashion. This may come in handy down the road for us because many of our subjects are individuals with schizophrenia, and being in an enclosed space (let alone the time it takes to yield a high-quality scan) is traumatic enough for them. Thanks for any input. -Derin

4 4

spherical smoothing
by Cameron Ellis 14 Dec '06

14 Dec '06

Hi, I am trying to smooth fucntional data on the spherical surface and I see that there are a few options for how this can be done and was wondering if anyone has any suggestions for an optimum smoothing technique. I see one option is implementing smoothing as a tag in mri_vol2surf but I also see sphsmooth-sess as another option. Does mri_vol2surf call sphsmooth-sess and work off the same algorithm? If not, when would I use sphsmooth? I also see that func2sph offers the option of smoothing before resampling. When would this be a good idea? Thanks, Cameron

2 1

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Freesurfer December 2006