R/NetworkCovariance.R

In stnava/itkImageR: Examples using Rcpp to interface R and ITK

# NetworkCovariance.R From - van Wijk, B.C.M., Stam, C.J., Daffertshofer, A.: Comparing brain networks of
# different size and connectivity density using graph theory. PLoS One 5(10) (2010) e13701

networkCovariance <- function(n1, n2) {
  if (sum(abs(dim(n1) - dim(n2))) != 0) {
    print("Inputs must be of same dimension")
    return(NULL)
  }
  N <- dim(n1)[1]
  
  n1 <- n1 - mean(n1[upper.tri(n1)])
  n2 <- n2 - mean(n2[upper.tri(n2)])
  
  cov <- 2/(N * (N - 1)) * sum(n1[upper.tri(n1)] * n2[upper.tri(n2)])
  
  return(cov)
}