Deriving the mapping is a bit trickier if you use the normalise to MNI
space option, as the affine transform between MNI space and your
population average is not saved anywhere. Deriving this mapping is
possible, but would require additional coding.
The mapping from population average to individual is much easier. You'd
need to write out the deformation field for spatially normalising images
to their common average (y_*.nii). Then each point in the deformation
is a mapping to native-space coordinates in mm. Deriving the mm
coordinates in the original image can be achieved using a modified
version of:
P = 'y_XXX.nii';
native_mm_coords = [20,30,-10];
P = [repmat(P,3,1),[',1,1';',1,2';',1,3']];
V = spm_vol(P);
vox = V(1).mat\[native_mm_coords, 1]';
mm_coords = [spm_sample_vol(V(1),vox(1),vox(2),vox(3),1),...
spm_sample_vol(V(2),vox(1),vox(2),vox(3),1),...
spm_sample_vol(V(3),vox(1),vox(2),vox(3),1)]
Best regards,
-John
On Tue, 2009-11-03 at 12:34 +0000, Richard Binney wrote:
> Hi justin,
>
> thanks for your suggestion, however i am aware of this and hoping that
> the FIL might be able to tell me a way I can use the (better) DARTEL
> flow field transforms. PLus rather than warping a sphere/ROI back, I
> want to take a single coordinate set and draw the sphere in individual
> space so that it is not a subject-specfic version of the pshere but an
> actual sphere sampling the singal from the brain surrounding this
> coordinate (does that make sense)
>
> I have thought of a work around which I am gonna give a try....this
> involves creating a gaussain sphere around the MNI coordinaters of
> interest, inverse warp this using the flow field and then in the
> subject space, determine the coordinate at which the maximum vlaue of
> the ROI falls. this SHOULD be the warped MNI coordinate.WE'll see
>
> Any comments welcome SPM-ers
>
>
> On Mon, Nov 2, 2009 at 10:04 PM, Justin E Brown <[log in to unmask]>
> wrote:
> Hi Richard,
>
>
> I think that if you define the ROI from your group analysis
> using MNI space, you can inverse warp that ROI to individual
> subject spaces using the the normalise tool and *_seg_inv_sn
> files that were created when you normalized each fucntional
> scan.
>
>
> --
> Justin Brown
>
>
>
> On Mon, Nov 2, 2009 at 8:05 AM, Richard Binney
> <[log in to unmask]> wrote:
> Dear John et al.,
>
> I have used DARTEL to normalise functional data. Given
> that DARTEL produces huge files when normalising whole
> EPI runs, I ran the GLM on unsmoothed data in the
> subject space and wapred the contrast images for the
> random effects analysis. this makes several things a
> bit trickier for rookies like myself, as I will
> explain.....
>
> I want to be able to inverse warp a single set of
> coordinates (i.e. a single voxel coordinate) from MNI
> space, back to each of the individual subject spaces
> in order to perform a functional connectivity analysis
> using an ROI derived from a peak found in the group
> analysis.
>
> Is there a way I can use the DARTEL transform to
> perform forward and backward warps of a single
> coordinate set? Using segment is was previously
> possible to do this with get_orig_coord5.m
>
> This would be VERY useful for a number of things I do.
>
> Looking forward to your repsonses
>
> Richard
>
>
>
>
> --
> Justin E Brown
> Graduate Student, Neurosciences Program
> Stanford University, School of Medicine
> email. [log in to unmask]
>
>
--
John Ashburner <[log in to unmask]>
|