It won't give you talairach coords in the label file, it will give you MNI305 coords. Use the View->Information in tkmedit to get MNI305 coords.
doug
Jed Singer wrote:
That almost works; tkmedit is giving talairach coordinates of, for a few examples:
14.7 23.9 -22.3 24.8 21.9 -18.8 35.6 20.8 -20.9 54.2 14.1 -17.1 61.1 10.9 -9.6 62.3 6.0 -1.8
Running mri_label2label with fsaverage as the target subject yields a label in which those same points have the coordinates:
14.877 25.802 -25.101 25.004 23.475 -20.993 35.985 22.497 -23.559 54.764 15.410 -19.482 61.693 11.670 -10.781 62.952 6.235 -1.732
For comparison purposes, the RAS coordinates in the original label are:
12.000 43.000 5.000 21.000 41.000 9.000 31.000 41.000 7.000 48.000 35.000 11.000 54.000 31.000 19.000 55.000 25.000 27.000
And talairach.xfm is: 1.1105 0.0582 0.0623 -0.1394 -0.0761 1.0376 0.1083 -21.5909 -0.0404 -0.0732 1.0813 -26.6708
Multiplying those RAS coordinates by the given talairach transformation gives coordinates of:
16.0028 22.6522 -24.8958 26.1303 20.3252 -20.7878 37.1110 19.3473 -23.3543 55.8898 12.2610 -19.2767 62.8184 8.5205 -10.5760 64.0778 3.0855 -1.5270
The three sets of values are close, but not the same. The discrepancy is really making me uncomfortable; any ideas what's going on?
Thanks so much, Jed
On Jul 22, 2011, at 5:39 PM, Douglas N Greve wrote:
You're so close! Use fsaverage as the target subject. The output coordinates will be in MNI305 space. doug
Jed Singer wrote:
I'm at my wits' end with what seems like it should be an easy question. I have a lot of volume labels that have been saved for a lot of brains. I want the Talairach coordinates corresponding to those points. It seems like mri_label2label should be able to do this, by putting in 'talairach' as the target subject, but this returns an error that the subject 'talairach' isn't found:
mri_label2label --srcsubject m00045 --srclabel label/all.label --trgsubject talairach --trglabel all_talairach.label --regmethod volume
It seems like I should be able to do it manually by loading the talairach.xfm generated by recon-all and multiplying that matrix by the RAS points in the labels; this yields results that are pretty close to those given in tkmedit, but not that close. Usually within a few, sometimes off by more.
Probably my inability to do this reflects some misunderstanding or lack of information, but I've been poring over documentation, forum, and wiki for hours now and feel like I've hit a wall. Any suggestions?
Thanks so much, Jed _______________________________________________