Hi Anastasia,

Choose any anatomical template, TT_N27, TT_avg152T1 or TT_icbm452.
The N27 looks good, since it is only one subject, but one might
prefer the others since they are averages.

If you particularly care about the accuracy of structures near the
AC-PC line, manual Talairach may be better. Otherwise, @auto_tlrc
should work just as well. Note that the Talairach transformation
is based on partitioning the brain into 12 pieces, and transforming
each of them.

The @auto_tlrc transformation is a single affine transformation.

----

You might use the "-SUFFIX NONE" option to @auto_tlrc. That way
there will be no '_at' suffix, and afni will understand the
relation between the +orig and +tlrc datasets. Without it, afni
has no way to know on what grid to put a dataset transformed to
+orig space.

----

You can apply the 1D matrix with any of 3dWarp, 3dAllineate or
3drotate.

For example, 3dWarp has options -matvec_in2out and -matvec_out2in.
Swapping between them means inverting the matrix in question. It
also has options -tta2mni and -mni2tta, for going between Talairach
and MNI-152 coordinate spaces.

- rick