The average cortical thickness in humans is somewhere around 2.5mm, so you don't have multiple layers of voxels, you have barely over 1 layer on average, in theory. Diagonals make this more complicated, of course.
The vertices are also not laid out as a grid, they are instead a triangular mesh - similar to hexagonal close-packing, this allows more of them to fit in a defined surface area with a given distance between them than a grid layout (like voxels) would. As you hint at, the cortical surfaces have spherelike topology, and a number of those vertices are in the "medial wall", where the single-hemisphere surfaces cut through the corpus callosum. We use just under 30k vertices per hemisphere for the cortex itself, out of the 32k vertices in the full surface. Finally, the 2mm spacing is somewhat approximate (it was adjusted empirically, not directly controlled), and it is not entirely even across the anatomical surfaces - some difference in stretching is inevitable because a brain inflates to something more like a football than a sphere, and we use fully regular spherical topologies for our current atlases. Tim On Tue, Apr 17, 2018 at 9:11 AM, Amanda Mejia <[email protected]> wrote: > In the HCP FAQ and several HCP-related publications (e.g., A Multi-modal > Parcellation of Human Cerebral Cortex by Glasser et al.), there is a > statement about the surface vertices in 32K resolution and the 2mm > isotropic voxels being on a "similar scale". The aforementioned paper > further states that the average cortical vertex spacing is 2mm. If I'm > reading this correctly, this would imply that the average distance in two > dimensions between each vertex is similar to the distance between two 2mm > voxels in a *single* layer. This would seem to imply that since the > cortical ribbon in volumetric space is several voxels thick, each vertex > would be equivalent to multiple voxels. But this isn't consistent with the > fact that there are a similar number of cortical gray matter 2mm voxels and > cortical surface vertices (~60,000 across both hemispheres). It would seem > to me that the only way to keep the number of locations the same, while > transforming the data from multiple layers (of voxels) to a single layer > (of vertices) would be to make the vertices closer together than 2mm. Can > someone please clarify? > > Amanda > > -- > AMANDA F MEJIA, PhD > Assistant Professor > Department of Statistics > Indiana University Bloomington > https://mandymejia.wordpress.com/ > > _______________________________________________ > HCP-Users mailing list > [email protected] > http://lists.humanconnectome.org/mailman/listinfo/hcp-users > _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
