Sorry, jiscmail rejected the attachment for some reason...
On 17/10/2007, DRC SPM <[log in to unmask]> wrote:
> Hi Matt,
>
> I wonder why it didn't work? I've just run it through avw2fsl (and
> also changed /bin/sh to /bin/bash in the #! line for good measure, as
> I seem to recall Debian/Ubuntu was unhappy with sh), and tested it on
> a pair of images with different image dimensions and different q- and
> s-forms (from co-registration in SPM5) as follows, and the result
> looks fine.
>
> flirt_qsinit.sh src ref src2ref.mat
> flirt -in src -ref ref -out src2ref -init src2ref.mat -applyxfm
> fslview ref src2ref
>
> I've attached my copy in case it helps, if it doesn't, hopefully Mark
> or Steve could take a look at your data if you upload it? (Or I could,
> if you put it somewhere I can see it).
>
> Best,
> Ged
>
>
> On 17/10/2007, Matt Glasser <[log in to unmask]> wrote:
> > Thanks Ged.
> >
> > I couldn't get the script to work, even when I changed avw to fsl, but the
> > SPM method worked fine.
> >
> > Peace,
> >
> > Matt.
> >
> > -----Original Message-----
> > From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
> > Of Ged Ridgway
> > Sent: Tuesday, October 16, 2007 10:21 AM
> > To: [log in to unmask]
> > Subject: Re: [FSL] MNI Space in MRIcro and FSL
> >
> > Hi Matt,
> >
> > Mark J posted a script that can use the NIfTI header info to generate
> > a FLIRT transformation:
> > http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0701&L=FSL&P=R20004
> > I think if you apply this transformation with flirt, with the standard
> > as the -ref, then your FSLview results should match your MRIcron ones.
> >
> > I'm guessing MRIcron resamples a binary ROI in the space of the
> > standard image, with either nearest neighbour interpolation or linear
> > followed by re-binarisation at a threshold of 0.5 -- you might need to
> > find out its implementation details to exactly reproduce the results.
> > Possibly MRIcron has an option to write out the resampled ROI, that it
> > presumably stores internally for display purposes?
> >
> > Hope that works,
> > Ged.
> >
> > P.S. Another option, if you use SPM (sorry to mention the enemy on
> > this list again!), is to use "coreg"->"reslice" with the standard
> > image defining the space and the ROI in the "other images" (or words
> > to that effect). The resulting ROI images prefixed with "r" should
> > then overlay properly in FSLview. Check out spm_reslice if you want to
> > avoid using the GUI.
> >
> >
> > Matt Glasser wrote:
> > > Hi Steve,
> > >
> > > How do you suggest I go about correcting for this difference? I could
> > just
> > > do it with nudge, but I would prefer an automated way that does not have
> > the
> > > chance of human error.
> > >
> > > Thanks,
> > >
> > > Matt.
> > >
> > > -----Original Message-----
> > > From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
> > > Of Steve Smith
> > > Sent: Tuesday, October 16, 2007 8:27 AM
> > > To: [log in to unmask]
> > > Subject: Re: [FSL] MNI Space in MRIcro and FSL
> > >
> > > Hi,
> > >
> > > On 16 Oct 2007, at 12:58, Matt Glasser wrote:
> > >
> > >> Hi Steve,
> > >>
> > >> Well okay, but why would the ROIs line up with FSL's standard brain in
> > >> MRIcro, but not in FSL itself? I used the same standard space
> > >> image in both
> > >> programs
> > >
> > > Ah - I hadn't appreciated that - sorry...
> > >
> > >> and it was only misaligned in one of them, so I don't know how I
> > >> could use a flirt transform in this case.
> > >
> > > Right - in that case the difference presumably is because MRICRO uses
> > > the co-ordinate centre from the header in determining display
> > > centering, whereas FSLView only uses the raw voxels for working out
> > > display (the co-ordinate centre is only used in the reporting of the
> > > mm co-ordinates, not in the image display).
> > >
> > > Cheers, Steve.
> > >
> > >
> > >
> > >> Thanks,
> > >>
> > >> Matt.
> > >>
> > >> -----Original Message-----
> > >> From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On
> > >> Behalf
> > >> Of Steve Smith
> > >> Sent: Tuesday, October 16, 2007 2:00 AM
> > >> To: [log in to unmask]
> > >> Subject: Re: [FSL] MNI Space in MRIcro and FSL
> > >>
> > >> HI - I presume this is because of a change in field-of-view in the
> > >> different standard space images. One easy way to solve that is to use
> > >> flirt to estimate the transform between the standard space image in
> > >> SPM/mricro and the FSL standard space image, and use that instead of
> > >> eye.mat in your call to flirt resample the ROI.
> > >>
> > >> Cheers, Steve.
> > >>
> > >>
> > >> On 16 Oct 2007, at 00:17, Matt Glasser wrote:
> > >>
> > >>> Hi,
> > >>>
> > >>>
> > >>>
> > >>> I have a set of ROIs from SPM in MNI space that I want to load into
> > >>> FSL. When I take the standard FSL brain MNI152_T1_1mm_brain and
> > >>> load it in MRIcro and overlay the ROI, it looks fine, like the
> > >>> first screenshot: http://www.ma-tea.com/~Matt/MRIcroOnMNI.png
> > >>> However, if I resample the 4X4X4mm ROI so it can be overlayed on
> > >>> this brain in FSLView, using:
> > >>>
> > >>>
> > >>>
> > >>> flirt -in 4X4X4mmROI -ref MNI152_T1_1mm_brain -applyxfm -init
> > >>> eye.mat -out 1X1X1mmROI where eye.mat is the identity matrix:
> > >>>
> > >>>
> > >>>
> > >>> 1 0 0 0
> > >>>
> > >>> 0 1 0 0
> > >>>
> > >>> 0 0 1 0
> > >>>
> > >>> 0 0 0 1
> > >>>
> > >>>
> > >>>
> > >>> The result is incorrect, shown by the second screenshot in FSLView:
> > >>> http://www.ma-tea.com/~Matt/FSLOnMNI.png
> > >>>
> > >>>
> > >>>
> > >>> The ROI file header is the following:
> > >>>
> > >>>
> > >>>
> > >>> <nifti_image
> > >>>
> > >>> nifti_type = 'ANALYZE-7.5'
> > >>>
> > >>> image_offset = '0'
> > >>>
> > >>> ndim = '4'
> > >>>
> > >>> nx = '41'
> > >>>
> > >>> ny = '48'
> > >>>
> > >>> nz = '35'
> > >>>
> > >>> nt = '1'
> > >>>
> > >>> dx = '4'
> > >>>
> > >>> dy = '4'
> > >>>
> > >>> dz = '4'
> > >>>
> > >>> dt = '1'
> > >>>
> > >>> datatype = '16'
> > >>>
> > >>> nbyper = '4'
> > >>>
> > >>> byteorder = 'LSB_FIRST'
> > >>>
> > >>> cal_min = '0'
> > >>>
> > >>> cal_max = '1'
> > >>>
> > >>> descrip = 'FSL4.0'
> > >>>
> > >>> sform_code = '2'
> > >>>
> > >>> sto_xyz_matrix = '-4 0 0 80 0 4 0 -112 0 0 4 -52 0 0 0 1'
> > >>>
> > >>> num_ext = '0'
> > >>>
> > >>> />
> > >>>
> > >>>
> > >>>
> > >>> I would appreciate any assistance with this as I am unfamiliar with
> > >>> this type of issue if it has to do with the analyze file format or
> > >>> header.
> > >>>
> > >>>
> > >>>
> > >>> Thanks,
> > >>>
> > >>>
> > >>>
> > >>> Matt.
> > >>>
> > >>>
> > >>
> > >> ----------------------------------------------------------------------
> > >> --
> > >> ---
> > >> Stephen M. Smith, Professor of Biomedical Engineering
> > >> Associate Director, Oxford University FMRIB Centre
> > >>
> > >> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
> > >> +44 (0) 1865 222726 (fax 222717)
> > >> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
> > >> ----------------------------------------------------------------------
> > >> --
> > >> ---
> > >
> > >
> > > ------------------------------------------------------------------------
> > > ---
> > > Stephen M. Smith, Professor of Biomedical Engineering
> > > Associate Director, Oxford University FMRIB Centre
> > >
> > > FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
> > > +44 (0) 1865 222726 (fax 222717)
> > > [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
> > > ------------------------------------------------------------------------
> > > ---
> > >
> >
>
>
|