Inferring the coupling structure of complex systems from time series data is a challenging problem in applied science across many disciplines. Reliable and true coupling structure inference requires us to distinguish between both direct and indirect influences between the components within the system. In this work, we present our data-driven information theoretic approach called optimal causation entropy (oCSE) to identify the information flow associated with the actual coupling structure. We will include discussion of examples such as the functional brain network as inferred by fMRI. We will contrast this oCSE network from the popular correlation based brain inference, and dynamical roles these both play in healthy function.
Copyright. All Rights Reserved.