The surf_*.mat files are produced by the <Xtract Brain> pulldown, and
requires Matlab 5.3 or above. The file contains two variables: faces
and vertices. The vertices matrix contains the co-ordinates (in millimeters)
of each vertex in the mesh. If the surface is derived from images
spatially normalised in Talairach space, then each row of the matrix
can be thought of as the Talairach co-ordinate of a vertex.
The each row of the faces matrix represents a triangular patch on a surface.
Each contains three elements, which is a list of vertices at the corners.
There are as many rows as there are triangular patches on the surface.
However, I'm not sure if the order of the vertices in a patch appears
clockwise or anticlockwise if viewed from outside the object. I'm not even
sure if the "isosurface" function in Matlab uses such a convention. If it
does, then I think that this ordering may change if the .mat files
associated with the images have negative determinants (one or three
negative voxel dimensions).
The help facility for <Xtract Brain> should give you some clues about how
the surfaces can be used.
I hope this helps,
-John
| I can fairly guess what is the format of the surf_*.mat files: The faces
| references three vertices.
|
| But which way round the vertices are listed? If I use the vectors from
| first to second vertex and from second to third vertex, is the cross
| product vector pointing outward? That would mean by looking from outside
| the vertices are counted counterclockwise.
|
| The coordinates of the vertices are in order X, Y, Z following the spm
| convention I guess?
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|