Hi Tim,
Does eddy_correct do anything special beyond the standard FLIRT affine
registration of the diffusion weighted volumes to a b=0 volume?
Geoff.
> -----Original Message-----
> From: FSL - FMRIB's Software Library
> [mailto:[log in to unmask]] On Behalf Of Tim Behrens
> Sent: 19 October 2004 14:18
> To: [log in to unmask]
> Subject: Re: [FSL] problems with FDT
>
>
> Hiya -
>
> Yep we're aware of the negative x dimension problem with analyze
> images. I think that all FSL programs write -ve x dimension
> in analyze.
> However, I would be surprised if this affects any future processing,
> seeing as FSL programs also ignore the sign of the dimension.
> (In fact, I
> believe dtifit is completely ignorant of the data dimensions).
>
>
> wrt eddy_correct, the effect that you are seeing is that
> small rotations
> needed for motion correction mean that there is no data left
> in the parts
> of the bottom or top slices.
>
> At the moment, these elements just have zeros placed in them
> - however, I
> think you're right that it would be better to pad with the
> nearest piece
> of "true" data. I have attached a replacement for the
> eddy_correct script
> which does this.
>
> if you download this file and do:
>
> mv eddy_correct ${FSLDIR}/bin
> chmod a+x ${FSLDIR}/bin/eddy_correct
>
> Then, next time you run eddy correction you shouldn't see this effect.
>
> Cheers
>
> Tim
>
>
>
>
>
>
>
> On Mon, 18 Oct 2004, Erin Gibson wrote:
>
> > I'm running into a few problems with FDT. I've changed the
> default save
> > format to Analyze, but the headers do not appear to be
> writing correctly.
> > The data.img/hdr file created with eddy_correct has a
> x-voxel dimension
> > that is negative and if DTIfit is run on that file, the
> results are just an
> > empty volume. If I correct the header, DTIfit produces
> reasonable output,
> > but those files also have negative valued x dimensions.
> >
> > The second issue is with eddy_correct. After running the
> correction on the
> > 4-D data, the first slice on all images has missing data
> and the exact
> > amount missing differs across all the orientation and
> baseline images. It
> > almost looks like a rotation correction, but it only
> affects the bottom and
> > top slices. Could there be something that I am not doing properly?
> >
> > The input image (s1_4D.img/hdr) and output image
> (data.img/hdr) can be found
> > here:
> > ftp://ftp.sunnybrook.utoronto.ca/pub/brainlab/fsl/
> >
>
> --
> --------------------------------------------------------------
> -----------------
> 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
> --------------------------------------------------------------
> -----------------
>
|