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
> ------------------------------------------------------------------------
> ---
>
|
