brainconn.similarity.edge_nei_overlap_bd

edge_nei_overlap_bd(CIJ)[source]

This function determines the neighbors of two nodes that are linked by an edge, and then computes their overlap. Connection matrix must be binary and directed. Entries of ‘EC’ that are ‘inf’ indicate that no edge is present. Entries of ‘EC’ that are 0 denote “local bridges”, i.e. edges that link completely non-overlapping neighborhoods. Low values of EC indicate edges that are “weak ties”.

If CIJ is weighted, the weights are ignored. Neighbors of a node can be linked by incoming, outgoing, or reciprocal connections.

Parameters:CIJ (NxN numpy.ndarray) – directed binary/weighted connection matrix
Returns:
  • EC (NxN numpy.ndarray) – edge neighborhood overlap matrix
  • ec (Kx1 numpy.ndarray) – edge neighborhood overlap per edge vector
  • degij (NxN numpy.ndarray) – degrees of node pairs connected by each edge