Network information and connected correlations. Author Elad Schneidman, Susanne Still, Michael Berry, William Bialek Publication Year 2003 Type Journal Article Abstract Entropy and information provide natural measures of correlation among elements in a network. We construct here the information theoretic analog of connected correlation functions: irreducible N-point correlation is measured by a decrease in entropy for the joint distribution of N variables relative to the maximum entropy allowed by all the observed N-1 variable distributions. We calculate the "connected information" terms for several examples and show that it also enables the decomposition of the information that is carried by a population of elements about an outside source. Keywords Models, Genetic, Models, Neurological, Models, Theoretical, Information Services Journal Phys Rev Lett Volume 91 Issue 23 Pages 238701 Date Published 2003 Dec 05 ISSN Number 0031-9007 DOI 10.1103/PhysRevLett.91.238701 Alternate Journal Phys Rev Lett PMID 14683220 PubMedGoogle ScholarBibTeXEndNote X3 XML