Freesurfer

freesurfer@nmr.mgh.harvard.edu

6 participants
30055 discussions

higher resolution reconstructions and mri_watershed
by Thomas, Adam (NIH/NIMH) 27 Sep '05

27 Sep '05

I've been playing around with reconstruction at higher resolutions and have been getting some unusual behavior so I decided to do some testing on bert. I convert bert to 0.7mm isotropic and 366 slices with: mri_convert -cs 0.7 ~/subjects/bert/mri/orig ~/subjects/bert0.7/mri/orig Then I fired up recon-all. Bert's surface came out beautifully without doing any manual editing. There are 276,245 nodes in the surface instead of 137,497 when reconstructed at 1.0mm. The inflated surface also looks a bit more convoluted at 0.7mm than at 1.0mm. However when I tried to recon bert at 0.5, mri_watershed gave some interesting errors: Skull Stripping Thu Sep 1 19:07:40 EDT 2005 /misc/radium0/adamt/subjects/bert0.5/mri mri_watershed T1 brain ************************************************************ The input file is T1 The output file is brain If this is incorrect, please exit quickly the program (Ctl-C) *************************WATERSHED************************** preflooding height equal to 25 percent Sorting... first estimation of the COG coord: x=252 y=303 z=228 r=259 first estimation of the main basin volume: 73641890 voxels Error w=White Matter =Intensity too high (>240)...valid input ? warning: non-standard value for imnr1 (512, usually 256) in file /misc/radium0/adamt/subjects/bert0.5/mri/T1/COR-.info warning: non-standard value for x (512, usually 256) in file /misc/radium0/adamt/subjects/bert0.5/mri/T1/COR-.info warning: non-standard value for y (512, usually 256) in file /misc/radium0/adamt/subjects/bert0.5/mri/T1/COR-.info warning: non-standard value for thick (0.0005, usually 0.001) in file /misc/radium0/adamt/subjects/bert0.5/mri/T1/COR-.info warning: non-standard value for psiz (0.0005, usually 0.001) in file /misc/radium0/adamt/subjects/bert0.5/mri/T1/COR-.inf o ERROR: mri_watershed exited with non-zero status Why would resampling to 0.5mm affect white matter intensity? Thanks, -Adam --- Adam Thomas adamt(a)nih.gov Functional MRI Facility, NIMH/NIH/DHHS 10 Center Dr, Room B1D708A Bethesda MD. 20892-1148 Phone:301-402-6351 Fax: 301-402-1370

3 2

anatomical stats
by Fornito, Alexander 26 Sep '05

26 Sep '05

HI, I'm finishing a study I completed using an older version of freesurfer (20040218), and have noticed that I get different values for surface area and grey matter volume when using mris_anatomical_stats for my labels depending on whether I use the mris_anatomical_stats that it is part of this older distribution, or the one packaged as part of the newer dev release (20050905). I noticed on a previous posting that Doug Greve identified an error with respect to how surface area was calculated (I've included it below) and was wandering if this is the reason for the discrepancy. Should I believe the figures calculated using the newer version, or will this also be associated with errors, since the surfaces were generated using the older version? Thanks, Alex I have just fixed a bug in FreeSurfer that involves computing the area > of some surface structures. For those of you who care, this was a bug > that emerged after converting our surface meshes from quadralaterals > to triangles. This change was made a few years ago. > > This change should have little effect on the reconstruction. It will > have an effect on programs that report the area of subsections of > cortex (but not total cortical area). At least three such programs > that are affected are: mris_compute_overlap, mris_label_area, and > mri_surfcluster. The bug causes the reported area to appear to be 1.5 > times what it really is. For example, in mri_surfcluster, a cluster > threshold of 300 mm^2 would actually be a threshold of 200 mm^2. > > For those in the NMR center, these three programs have been > updated. For those outside, you can get these programs from > > ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve > > New versions of freesurfer will, of course, have the update. > > Sorry for any inconvenience. > > doug >

2 1

Scuba/recon-all
by Satrajit Ghosh 26 Sep '05

26 Sep '05

I see the following statement printed several times through the recon-all processing and when I try to manipulate images in scuba. Is this something I can safely ignore? INFO: Volume /tmp/mritotal_5799/nu_8_dxyz.mnc cannot be found. Thanks, Satra

2 1

Freesurfer output question
by Michele Perry 26 Sep '05

26 Sep '05

Hi, I have some questions (sent to me by someone else, actually) about freesurfer output from make_cortex_table.csh. 1. Do integrated_rectified_mean_curvature & integrated_rectified_Gaussian_curvature refer to the shape/arc of the surface of the cortex either before or after inflation? Or are they descriptors of the distribution of the individual thickness measures... as might indicate if the measures were normally distributed (within an area) or skewed to the left or right? 2. Does the num_of_vertices represent the number of measures of thicknesses made for that region? 3. Would be possible to examine the distribution of thickness measures to identify focal regions of skewing... as might occur if there were micro-foci of atrophy within some areas? For example, values for the thicknesses corresponding to the 10th, 25th, 50th, 75th, and 90th percentile values within an area? Or even a median thickness, in addition to the mean?

2 1

mgz format
by Fornito, Alexander 26 Sep '05

26 Sep '05

Hi, Just wandering if someone could tell me what converting to mgh or mgz actually does to the image. COR changed it to 256x256x256, 1mm3, what happens with mgz/mgh? Thanks, Alex

3 2

Freesurfer staff,
by Anthony Cacace 26 Sep '05

26 Sep '05

Freesurfer staff, We're relatively new at running freesurfer, everything works fine until we run: recon-all -subjid 1801-4 -autorecon1 -mgz then we get the message: no run data found in: /home/amc/freesurfer_new/freesurfer/subjects/1801-4/mri Is there something simple here that we're missing? Tony

3 2

old version
by Fornito, Alexander 26 Sep '05

26 Sep '05

HI, I need to re-install an older version of freesurfer in order to finish a study I have been working on, but it seems that particular distribution is no longer on the website. I think it was the 20040218 distribution (or maybe that was that update for it, I'm not sure). The recon-all -version is $Id: recon-all-nmr,v 1.37 2004/01/06 22:21:48 tosa Exp $ Is there somewhere that I can still download this version from? Thanks, Alex

2 1

new release
by Kirk, Gregory 22 Sep '05

22 Sep '05

Hi freesurfer lords ! I see the grand release is released, do I need to get a new .licence file , i.e. fill out a new registration form online ? Thank you Greg

5 4

Two questions.
by Clifford A. Smith, PhD 22 Sep '05

22 Sep '05

First, I have successfully completed recon-all on my Mac. I created the subject directories using mksubjdirs prior to running recon-all. After running recon-all the files are all in the mri folder, and the respective folders (eg brain, T1, and filled) remain empty. Is it necessary to build the folders? Second, I would like to use the dti feature in tkmedit. I have the vector and fa volumes calculated with several computer packages in Analyze format. What does tkmedit need to read? A 4d volume with the three vector volumes plus the FA volume? Thanks. ------------------ Clifford Smith, PhD Assistant Professor of Behavioral Sciences Rush University Medical Center 312.942.2893

2 1

error: TR inconsistent
by agold＠nmr.mgh.harvard.edu 22 Sep '05

22 Sep '05

Hi, Our data is located in /space/kermit/8/users/milad and we had already analyzed this data set with 8 subjects and now are trying to add in 6 new subjects and analyze the final cohort of 14 subjects. During the mkanalysis-sess.new step, we received the following error message: ERROR: /autofs/space/kermit_008/users/milad/ext013d2 has inconsistent TR We checked the seq.info file: [kermit:milad] (nmr-std-env) more ext013d2/bold/seq.info sequencename recall nrows 64 ncols 64 nslcs 45 rowpixelsize 3.125 colpixelsize 3.125 slcpixelsize 3.0000001575331 ntrs 268 TR 3.0000 Turns out that the TR is consistent with the other subjects (i.e., TR = 3), but the slcpixelsize is smaller. It is 8 for all of the other subjects. E.g., see the subject listed below: [kermit:milad] (nmr-std-env) more ext012d1/bold/seq.info sequencename unknown nrows 64 ncols 64 nslcs 45 rowpixelsize 3.125 colpixelsize 3.125 slcpixelsize 8 ntrs 54 TR 3 It's unclear to us whether this is an issue that occurred during data acquisition or preprocessing. How should we go about changing this? Thanks! Andrea -- Andrea Gold Clinical Research Assistant Psychiatric Neuroimaging Research Program Massachusetts General Hospital Phone: (617) 726-5594 Fax: (617) 726-4078 agold2(a)partners.org -- The information transmitted in this email is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of or taking of any action in reliance upon this information by persons or entities other than the intended recipient is prohibited. If you received this email in error, please contact the sender and delete the material from any computer.

1 0

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