Hi,
Yes, it appears that this is a bug in img2talcoord (and also in
tal2imgcoord) - sorry.
An easy fix for any existing output is to take the y coordinate,
multiply by -1 and subtract 32 from the result.
That is: Y_correct = -1*Y - 32
This should give the same values as in the output from Feat (via cluster)
which is correct.
This was a result of not removing some of the medx y-coordinate
conventions correctly (the default behaviour without -flirt, which uses
medx conventions, is correct).
I've made new linux, macosx and alpha binaries for img2talcoord and
tal2imgcoord
that can be downloaded from
http://www.fmrib.ox.ac.uk/~mark/files/img2talcoord/
If anyone needs an irix or sun binary, let me know and I'll organise
something.
Sorry for the trouble,
Mark
Zrinka Bilusic wrote:
> Hello,
> as I was working on some things (possibly another question later on)
> with talairach coordinates and I have run into a problem. I tried to
> use img2talcoord to get tal coordinates for a list of voxels in the
> original space. here is the command I have used...
>
> img2talcoord -tal standard -img example_func -xfm
> example_func2standard.mat -vox -flirt original_coords.txt >
> mni_coords.txt
>
> I have run this within the reg/ folder, and original_coords.txt is a
> modified Imax_zstat1.txt file from the same analysis. I just wanted to
> confirm that I will get the same coordinates as are reported in the
> feat report (in cluster_zstat1_tal.html, and listed in
> Imax_zstat1_tal.txt). But I do not get the same numbers! Here are just
> a few coordinates from the list...
>
> original coordinates (voxels)
> 30 5 3
> 35 8 0
> 30 6 1
> 27 4 5
>
> corresponding tal coordinates from Imax_zstat1_tal.txt (result of feat
> analysis)
> 6.95 -92.4 19.1
> -9.38 -85.7 2.32
> 6.96 -91.3 8.14
> 16.8 -93.7 29.8
>
> and here are the coordinates output by img2talcoord
> 6.95002 60.3709 19.1
> -9.37855 53.7371 2.3233
> 6.96059 59.3127 8.13768
> 16.839 61.7015 29.7874
>
> As you can see, the y coordinates are way off. Why? I have
> doublechecked all my files: input functionals are in the same
> orientation (Posterior on top) as the standard file (mni standard);
> there is nothing out of the ordinary in the header files for the
> functional as well as the high res used in the registration. And the
> registration results are very good. Have I just used img2talcoord
> incorrectly, or is there something else that I am missing?
>
> Thank you!
>
> Zrinka Bilusic-Vezmar
> UCLA Brain Mapping
> 660 Charles Young Drive South
> Los Angeles, CA 90095
>
> 310-794-5060
> [log in to unmask]
> www.brainmapping.org
|