Hi there AFNI folks:

I've read over a lot of the material on the web (and on this message board) about converting images in MNI space into tlrc space (for use with AFNI's volume rendering, 'where am I?' plugin, and tlrc colors), and am still somewhat unsatisfied with current solutions. However, I do have an idea for what might work well, and wanted to see if anyone else has tried doing this ...

I have taken the avg152T1 brain and created an AFNI orig dataset using to3d. Then, using the markers withing AFNI, I have tried to manually mark the locations of the ac, the pc, midsaggital points, and the edges of the brain. This will then pull the _entire_ MNI brain into tlrc space (though obviously with many caveats and errors). This dataset could then be used as a parent dataset for any image which is already in MNI space and is labeled as an AFNI orig dataset.

My question is .. has anyone done this before, and if so, what are the optimal locations (in MNI space) for the various markers? The first time I tried it, the results I got looked decent, but the readout from the AFNI 'where am I?' window in MNI coordinates did NOT agree with the original voxel locations in the MNI brain (which should be the best test of AFNI's internal tlrc -> MNI mapping, it seems to me). I guess a corollary question is ... is the internal AFNI tlrc -> MNI mapping the one developed by Matthew Brett, or something else?

My hope is that a 'standard' set of markers could be established so that when different people import images normalized in MNI space into AFNI, the markers used are always the same.

Here is what I am using thus far. I am definitely not an anatomist, however =), and would very much welcome any recommended adjustments.

Marker MNI location (+L/-R +P/-A +S/-I)
AC superior edge: 0 -2 -2
AC posterior margin: 0 0 -2
PC inferior edge: 0 -28 -4
1st midsaggital point: 0 -2 62 (probably unimportant as long as LR = 0)
2nd midsaggital point: 0 44 56 (ditto)

For these, the question is at what point in the probability density of the MNI brain do you draw a boundary:

Anterior edge: -74
Posterior edge: 106
Superior edge: 88
Inferior edge: -49
Left edge: 71
Right edge: -71

Like I said, this mapping leads to disagreement between AFNI's readout of MNI coordinates and the true MNI coordinate (which can be compared by clicking back on the 'orig' view once the tlrc transformation has been made). This is why I'd appreciate any recommended fixes =)

Here is the command I used to read the brain into AFNI:

to3d -anat -prefix avg152T1 -view tlrc -xFOV 91R-L -yFOV 127P-91A -zFOV 73I-109S -orient LPI 3Db:0:0:91:109:91:avg152T1.img

which should preserve the origin at 46, 64, 37.

Ed