On 09/10/2012 08:30 AM, Tuomas Tolvanen wrote:

We need the description of sampling from 3D data on the surface (mri_vol2surf). We are specifically interested how the position is determined in 3D, when the fraction/distance from wm surface is announced. Is it e.g. the normal from one node (resulting in some overlap of 3D projections at the bottom of sulci), or is the neigbourhood somehow interpolated to get a better approximation of the normal direction.

It is just the normal direction.

Is the projection always a point projection (nearest neighbour or symmetrical interpolation)?

By default, it is nearest neighbor (can be changed with --interp)

Is mri_vol2surf using the voxels only once or is it possible that voxels are used multiple times?

It is possible (and likely) that a voxel is used more than once. This artifact is often reduced by spatial smoothing on the surface. The use of the same voxel multiple times then just contributes to the overall smoothness of the image. This is incorporated into the group statistics when the smoothness of the residuals is estimated.

If the method has been described in any publication, that would be helpful.

-Tuomas _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer