Hi all,
as previously mentioned Tim's remarks will be very important to any data converted by SPM's dicom toolbox and analyzed within FSL.
The scaling itself does not matter but it MUST be the same for the volumes of a given DTI data set or a fMRI time series.
So people have to rescale their data when using SPM's dicom toolbox for conversion. It may not be the problem in Stefan's data set but it is extremely important to recognize in general.
Best regards-
Andreas
-----Ursprüngliche Nachricht-----
Von: Tim Behrens [mailto:[log in to unmask]]
Gesendet: So 28.11.2004 18:55
An: [log in to unmask]
Cc:
Betreff: Re: [FSL] AW: [FSL] No results with FDT Diffusion
Hi Stefan -
Can you check in fslview whether the diffusion-weighted volumes have lower
signal intensity than the non-weighted ones.
A problem tha I have come across before with data which has been processed
with SPM is that SPM inserts a scale factor into the header ( I think it
use the header variable called funused1 ) which then gets completely
ignored by the FSL I/O routines..
Another way to check for this problem would be to run the following
command on each of the 3D volumes output from SPM ( before the merge!):
avwval volume funused1
This should be unset, or set to 1 in each case.
If this is not the problem, then I would be happy to take a look at the
data if you can tar it up and put it on an ftp site. Please do not email
it.
Thanks
Tim
-------------------------------------------------------------------------------
Tim Behrens
Centre for Functional MRI of the Brain
The John Radcliffe Hospital
Headley Way Oxford OX3 9DU
Oxford University
Work 01865 222782
Mobile 07980 884537
-------------------------------------------------------------------------------
On Sun, 28 Nov 2004, [iso-8859-1] Stefan Klöppel wrote:
> Dear Tim,
>
> thanks for the advice. I ran dtifit with spm-normalized data and received
> only blank images (of around 30 kB each). When doing the same with ecc
> volumes the error after slice 105 is 'child killed: SIGABRT'. The computer
> has 3 GB swap Ram so I don´t assume problems from that directions. I also
> checked pro recently described traps (e.g. UNIX carriage return).
>
> Any ideas?
>
> Stefan
>
>
> -----Ursprüngliche Nachricht-----
> Von: FSL - FMRIB's Software Library [mailto:[log in to unmask]]Im Auftrag
> von Tim Behrens
> Gesendet: Dienstag, 23. November 2004 14:29
> An: [log in to unmask]
> Betreff: Re: [FSL] No results with FDT Diffusion
>
> Dear Stefan - Have you run dtifit on this directory - this is always a
> good check to make sure that everything in the directory is ok, as it only
> takes a couple of minutes..
>
> Tim
>
>
> On Tue, 23 Nov 2004, Stefan Klöppel wrote:
>
> > Hi,
> > I have been trying to compute DWI-data (60 directions, 6 B0 images 4
> > repetitions, 264 files) motion corrected and normalized with spm. I left
> out
> > the eddy current correction. I used avwmerge to generate the 4d volume and
> > used the first B0-image as nodif-volume. There were no complaints by
> > bedpost_datacheck. After 30 hours I viewed the outputs of bedpost and
> found
> > very distorted images, hardly recognisable as a brain. Consequently, when
> > using probtrack, I received only blank images for every single seed voxel.
> >
> > When leaving out the whole pre-processing with spm but using fsl eddy
> > current correction (ecc) and volume 1 as reference volume I receive the
> > message: Both reference and input images have an sform matrix set (what
> does
> > it mean?) and ecc leaves me with a filesize below 1 MB (input data was
> > 200MB). Using avwsplit to determine the bvecs (I assume that they are
> > changed by ecc) produces nifti_read_buffer error. All extracted files
> except
> > the first are empty.
> > I am stuck. Is there a way out.
> > Thanks
> > Stefan
> >
>
> --
> ----------------------------------------------------------------------------
> ---
> Tim Behrens
> Centre for Functional MRI of the Brain
> The John Radcliffe Hospital
> Headley Way Oxford OX3 9DU
> Oxford University
> Work 01865 222782
> Mobile 07980 884537
> ----------------------------------------------------------------------------
> ---
>
|
