Hi Jidan,

we construct fsaverage by mapping the icosahedron vertices to a set of subjects, then assigning that vertex to the position that is the average of the tal coords of those vertices.  Sorry, I don't understand the 2nd point. Maybe you can explain what you are trying to accomplish?

cheers,
Bruce

On Thu, 23 Jul 2009, Zhong Jidan wrote:

Hi Bruce,

Thanks for your reply. I'm not clear about two points you mentioned.

->To represent a folded surface we either (1) map to an individual subject,
or (2) use a volume transform such as tal and the vertex correspondence from
the sphere.reg to find the average coordinate of a vertex. This is how we
build the fsaverage and average7 surfaces.

For the (2) point, what do you exactly mean?  I?m confused. I thought your
fsaverage is derived by averaging 40 subjects' volume image and then
reconstruct the surface based on that. How can I use a volume transform such
as tal and the vertex correspondence from the sphere.reg to find the average
coordinate of a vertex? I'm totally lost with this sentence..


->You can use the mapping to paint the geometry of one subject onto another
folded surface if you want to visualize the geometric mapping (e.g. the curv
of one subject on the white surface of another).

For this one, I think you mean that, I can use one white surface as a base.
Then paint the curvature information on the white surface to see the
geometry. But this is not I want... To be specific, if I have a lh.sphere,
lh.sphere.reg, they share the curvature information, and the only difference
is the spherical coordinates. Then it is non sense to paint the curvature
information to another white surface to see the geometry because there is
only one .curv file. Another problem is, this lh.sphere may have different
number of points with the white surface. ... Not sure whether I understand
it correctly, hope for your suggestion.

Thanks,

Jidan




On Tue, Jul 21, 2009 at 8:26 PM, Bruce Fischl <fischl@nmr.mgh.harvard.edu>wrote:

Hi Jidan,

the atlas only exists in spherical coords. To represent a folded surface we
either (1) map to an individual subject, or (2) use a volume transform such
as tal and the vertex correspondence from the sphere.reg to find the average
coordinate of a vertex. This is how we build the fsaverage and average7
surfaces (we know this is a hack, but it's easy and a good visualization
tool).

You can use the mapping to paint the geometry of one subject onto another
folded surface if you want to visualize the geometric mapping (e.g. the curv
of one subject on the white surface of another).


cheers,
Bruce



On Tue, 21 Jul 2009, Zhong Jidan wrote:

 Hi Freesurfer experts,

I asked this question previously, but I found it problematic when
displayed
in your mailist. I'm sorry that the question still not solved and I feel
sorry to trouble you again.

In your sphere registration in freesurfer, the procedure is like:
 creating
the template.tif by mris_make_template. The template you use in
Freesurferis created by iterative registration of 40 subjects,

according to
"High-resolution inter-subject averaging and a coordinate system for the
cortical surface, Fischl, B., Sereno, M.I., Tootell, R.B.H., and Dale,
A.M.,
(1999). Human Brain Mapping, 8:272-284(1999)".
 So, after the template generation process, you will get a .tif file which
include the necessary infomation (like the means and variances of curv,
sul
from the aligned spheres). But,do you have the other information of this
final template, such as the sphere representation, folded surface
representation of this template?  I know that under
*/subjects/fsaverage/surf, there are some surface representations of the
average of the 40 subjects, but to my knowledge, they are just used for
visulazation and are not the surface representation of the template.tif
you
used, am I right?

2, subjects'  sphere registration to the template sphere
 In this process, we can get the  deformed subjects spheres( *.reg ),
which
have a one-to-one correspondance to the  original subject surfaces. Except
the .reg sphere with the cuvature information, do you have any other form
of
representation of the deformed sphere?  You know that there are other
kinds
of surface mapping methods, like Miller's Large Defformation deffeomrphic
surface mapping, they just do surface mapping using the folded surfaces.
After surface mapping, they will get the deformed folded surface which
would
be aligned with the template  folded surface. With the deformed subject
and
template folded surfaces, they can tell directly which sulcus or gyrus is
aligned well.  So, for your mapping, when I get the deformed  sphere, do
you
have any command or method to put the sphere back to the folded surface so
I
can see the suci and gyri directly?  If you also  have the surface
representation of the template, then i can superimpose them to see how
 good
the alignment is.

If you think I didn't state this problem clearly, please refer to an
example
in the following:

I found one reference using your sphere registration method. "Simpli˙˙ed
Intersubject Averaging on the Cortical Surface Using SUMA"Brenna D.
Argall,
Ziad S. Saad,and Michael S. Beauchamp"Human Brain Mapping 27:14
˙˙27(2006)"
You may see the attachment in :

https://mail.nmr.mgh.harvard.edu/pipermail//freesurfer/2009-May/010558.html


In "Spherical Morphing" section, They mentioned that "  Using the
mris_register [Fischl et al., 1999b] routine, each individual subject˙˙s
surface was registered to the FreeSurfer average7 template prior to node
number standardization. Standardization and averaging were then performed
on
the surfaces as described above" (using SUMA FYI).
---- From this part, I assume that all the deformed surfaces are in
spherical representation.

Then in the result part, in section "Intersubject Averaging of Functional
Data: Different Surface Methods", they mentioned they " in order to
compare
the AC˙˙PC method to these more complex algorithms, the FreeSurfer program

mris_register [used in Fischl et al., 1999b] was used to morph the
cortical
surface models to a prede˙˙ned template, and these morphed surface models
were then used to create a morphed surface average."

In Fig7C ˙˙Average surface created by averaging the same 28 subjects using

mris_register standardization. You can see that they show the average
surface in a folded surface representation, not a sphere.

Could you give me a hint that how they do this since you only have a
sphere
representation of the aligned surface?