COUPLING CMLS AND THE SYNCHRONIZATION OF A MULTILAYER NEURAL COMPUTING SYSTEM

A coupled map lattice (CML) provides a mathematical model for both spatially extended physical and biological systems, and for a new type of parallel computing system. We analyse a general process to realize a multilayer structured CML by coupling m CMLs together with arbitrary coupling structure. As an application we use this coupling process in the study of a CML with multilayer planar architecture and non-symmetric hierarchical coupling between the layers. These CMLs model neural architectures and use chaotic maps as the basic elements in the lattice. We measure the complexity of the state of each layer and an increase in spatio-temporal coherence as one ascends a hierarchical feedforward layered network.