Corticonic Networks for Higher-Level Processing
Abstract
It is generally agreed that brain tissue is the most complex self-organizing matter known in the universe. In particular this applies to the cortex and all subcortical centers like the thalamus and the hippocampus with which the cortex interacts to carry out higher-level brain functions, such as perception, cognition, memory, language, control of complex motion, speech, and perhaps even awareness and consciousness. Understanding and creating models of how the cortex carries out such operations and implementing them in suitable fast and efficient algorithms or hardware will have far reaching implications for science, technology, and medicine, with the most obvious being the formulation and testing of models of higher-level brain functions and the design of future machines with brain-like intelligence. A set of equations seeking to model the way the cortex interacts with subcortical areas to produce certain higher-level brain functions is described. The equations are those of a network of parametrically coupled maps that incorporates salient properties of the cortex. Justifications for this approach and demonstration of its effectiveness for a parametrically coupled logistic map network (PCLMN) are presented. The PCLMN can self-organize under information-driven adaptation, is capable of handling dynamic (spatial-temporal) input patterns, furnishes an enormous number of attractors for inputs to choose from, plus it has other intriguing features that can be used in the design of intelligent systems. (3 figures, 4 refs.)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 25, 2004
- Accession Number
- ADA423292
Entities
People
- Nabil H. Farhat
Organizations
- University of Pennsylvania