Hi all,
Surface plots created in SurfIce versus Freesurfer are quite different. When visualizing the same volumetric map in either software, the extent of activity or regional coverage appears to vary substantially. Freesurfer is apparently more conservative in its spatial spreading of projected volumetric data. But is it also more accurate in its representation? Are you aware of any comparative study or consensus on which is more trustworthy?
I’d highly appreciate your thoughts on this!
Many thanks,
Henning
As noted in the Surfice manual:
A major challenge with overlaying voxelwise data onto surface meshes is that there is not a one-to-one correspondence between the location of the voxels and the vertices of the mesh. Therefore, one must interpolate in order to determine the intensity at the location of a vertex. This is particularly challenging for data that have been thresholded as a low resolution voxelwise image. In this case, all voxels that do not survive the statistical threshold are artificially set to zero.
In general, I think it is worth using voxel-based images with MRIcroGL, and using Surfice for analyses where all analyses are done per-vertex. The dilation and flattening of a mesh will distort the shape, and the correspondence to your voxel-based image. If you must use a voxel-based image, I suggest you provide an unthresholded volume, so that the boundary of surviving voxels is not artificially surrounded by zeroed voxels (which will decrease the apparent size when interpolation is applied). I do wonder if this might be one reason for the discrepancy that you see: FreeSurfer may not be accounting for this. One other change you can make is with the Surfice preferences, where the Smooth voxel-based images impacts the interpolation.
1 Like