Hi Mark, I think it may be because the voxels I'm using are all near the frontal poles. It seems to work fine with the other flirt option by transforming to subject space earlier so I'll just do it that way instead.

thanks for all your help!
Mengfei

On Mon, Aug 10, 2009 at 10:01 PM, Mark Jenkinson <[log in to unmask]> wrote:
Hi,

I'm surprised that a coordinate within the brain will be transformed outside
the FOV.  Have you checked coordinates that are not near the edge of the
brain?

Your flirt call is fine if you want to do that instead.

All the best,
       Mark



On 10 Aug 2009, at 20:27, Mengfei Huang wrote:

Hi Mark,

I just tried that and ended up with coordinates that fell off of my highres brain (y coordinate). Is this a reasonable problem to encounter due to inaccuracies in the transformation matrix, or would this indicate that something has gone wrong?

If it's the former, I'll go back and just transform the original mask into subject space first before picking the peak voxel. i believe that would be:
 flirt -in [mask in standard space] -ref reg/standard -applyxfm -init standard2highres.mat -out [mask in subject space] correct?

thanks!
Mengfei

On Mon, Aug 10, 2009 at 7:41 PM, Mark Jenkinson <[log in to unmask]> wrote:
Hi,

That's not quite right.
The syntax is:
 std2imgcoord -img reg/highres -std reg/standard -xfm reg/highres2standard.mat -vox coordfile
where coordfile contains your input coordinates (in standard space) as text.
The output goes to the screen.
To capture it in a file just end the line with:    > outputfile.txt
as the ">" operator in Unix directs the output (of any command) into the following file.

You are right that you give it the highres2standard.mat transformation, which it internally
inverts, to convert coordinates from standard to highres space.

All the best,
      Mark



On 10 Aug 2009, at 19:34, Mengfei Huang wrote:

Hi Mark,

so just to double check the syntax would be:

std2imgcoord -img <text file containing voxel coordinates> -std reg/standard -xfm <subject's hires2standard.mat> <output coordinate file in subject space> -vox (to output coordinates in voxels)

Is this correct? So even though you're using the hires2standard matrix, it converts from standard to hires?

Thank you!
Mengfei

On Mon, Aug 10, 2009 at 6:08 PM, Mark Jenkinson <[log in to unmask]> wrote:
Hi,

Just use the tool std2imgcoord.

All the best,
     Mark



On 10 Aug 2009, at 18:04, Mengfei H wrote:

Dear FSL,

Is it possible to use FLIRT to transform voxel coordinates (x,y,z) from standard
space back into subject space?  If so, what would be appropriate syntax?

I've seen it done for nii.gz images but not specific coordinates

Many thanks!
Mengfei




--
Mengfei Huang
Fulbright Scholar (UK) 2008-2009
MSc Neuroscience
Oxford University
BS Biological Sciences
Stanford University




--
Mengfei Huang
Fulbright Scholar (UK) 2008-2009
MSc Neuroscience
Oxford University
BS Biological Sciences
Stanford University




--
Mengfei Huang
Fulbright Scholar (UK) 2008-2009
MSc Neuroscience
Oxford University
BS Biological Sciences
Stanford University