Dear Kevin,
In std2imgcoord the -img should represent the final coordinate space.
So if you are going from std 2 example_func then -img should be example_func.
I think you are also using the wrong coordinates. You say you are at MNI
coordinate -27 -3 -15, yet lamycor.txt contains 64 123 57. Where did this last
coordinate come from? Was it this process in MRIcro that you describe?
The input coordinates to std2imgcoord should contain standard space
coordinates (like -27 -3 -15). The *output* coordinates can be either mm or vox,
depending on the options, but the input should always be the MNI coordinates
for this. Sorry if that is not clear. When I tried to go to 64 123 57 in standard
space, it was indeed well outside the brain.
I've no idea what a 3D one-point standard space image is, so I'll ignore that.
Your fslmaths call is just fine.
All the best,
Mark
On 2 Mar 2011, at 15:14, Kevin Denny wrote:
> Here are the commands we used with std2imgcoord to transform an MNI coordinate point into subject's individual space. Our goal is to draw a 5mm sphere from that point in each individual space using fslmaths. Right now we are trying for a left amygdala seed at MNI coordinates -27 -3 -15
>
> std2imgcoord -img lAmy.nii.gz -std standard -xfm example_func2standard.mat -vox lamycor.txt
>
> First, for the input image: what should it be? In this case it is a 3D one-point standard space image (lAmy.nii.gz) made in MRIcron. I have seen instructions on here to use the highres image from reg folder and to use the example_func image. Which should we use as input?
>
> Second, to eventually draw the ROI in subject's space, fslmaths needs voxel coordinates. The lamycor.txt file has the coordinates 64 123 57. Is this correct? We ended up (from stdimgcoord) with the output 113 240 125 which is outside of the individual brain.
> To get out coordinates for the txt file, we entered the MNI coordinates in MRIcron in the "MNI Position window" and then used the voxel coordinates X Y and Z in the MRIcron main window at the top. Is this the wrong way to get voxel coordinates? Is there another way to convert MNI to voxel coordinates?
>
> After we have straightened this out, we want to make the ROI using fslmaths. Is this the correct command?
>
> fslmaths example_func.nii.gz -mul 0 -add 1 -roi 5 1 5 1 5 1 0 1 newroi
> (where the hypothetical xyz coordinates from std2imgcoord are 5 5 5 and the size of each is 1, tmin is 0 and tsize is 1 and output name is newroi)
>
> Thanks in advance,
> Kevin
>
|