Astrocyte Endfeet Description

Astrocytes project branches that attach on the surface of the vasculature, wrapping around its surface, forming patches known as perivascular endfeet. Astrocytic endfeet are reconstructed during the ngv circuit building as two-dimensional triangulated surfaces. In order to approximate a three-dimensional object, a constant thickness is assigned to each two-dimensional endfoot surface.

The endfeet establish connections between astrocytes (glia) and the vasculature. Therefore, endfoot data is indexed by endfoot edge ids, where more than one value (e.g. points, triangles) corresponds to each id. For this reason the grouped properties file layout is used to store the endfeet data (See here for more details).

Group	Field	Type	Requirement	Description
/data	points	float32	Mandatory	Positions (x, y, z) of the mesh vertices in µm
/data	triangles	int64	Mandatory	Triangular connectivity of the point vertices
/data	surface_area	float32	Mandatory	Surface area in µm²
/data	surface_thickness	float32	Mandatory	Thickness in µm
/data	unreduced_surface_area	float32	Mandatory	Surface area in µm² before reduction by the target distribution
/offsets	points	int64	Mandatory	Ranges for the points dataset
/offsets	triangles	int64	Mandatory	Ranges for the triangles dataset

Geometry

Perivascular endfeet are stored as 2D surfaces in the 3D space. Their discretized geometry consists of points (vertices) and triangles, forming triangular meshes. Given that biological endfeet have three-dimensional geometries, a surface thickness has been assigned to each endfoot to roughly approximate its extent along the normals of the 2D surface.

Surface Areas

During the endfeet surface generation stage in the NGV building workflow, the surface meshes are first grown competitively on the surface of the vasculature until there is no available space. These initial surface areas are stored as the property unreduced_surface_area.

Then the mesh geometry is pruned, shrinking the endfeet, so that their surface area distribution matches the biological target distribution. The final surface areas are stored as surface_area and correspond to the surface area from the points and triangles stored in the file.