Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

A lumped element modeling schema is presented that can be usefully applied to modeling and analysis of metabotropic processes affecting the signal processing properties of neurons. The schema is easily applied to qualitative models of such biochemical processes and facilitates the development of differential equation descriptions. The state variables within the model are applicable to augmenting and extending Hodgkin-Huxley-like models (H-H models) of neuron dynamics.

This tech brief presents an overview and general explanation of the Martian Research Program here at the University of Idaho's MRC Institute. It begins with a strategic overview of the long-term research objectives, followed by an overview of the current near-term research objectives. The Martian system in its current state of development is next explained and discussed. Finally, the specific research aims for the next twelve months are presented.

A companion tech brief (Wells, 2007a) introduced the Martian, a simple agent system serving as a platform for investigations of meanings-based, affect-driven unsupervised learning processes. The purpose of using a Martian as an investigation platform is to introduce a biologically homologous agent platform that is complex enough to introduce biologically reasonable constraints on the learning processes without introducing so much complexity of detail in the anatomy of the agent as to render analysis of learning phenomena untenable. This tech brief provides a basic definition of the lower sensorimotor anatomy of the Martian.

Entry needed.

This tech brief concerns the problem of the neural code and a new approach to the neural coding problem. Briefly stated, the problem of the neural code is this: What is the organization of the system or systems of information coding in the brain? The problem was first raised in 1956 by John von Neumann and to this day has remained one of the outstanding unsolved problems of theoretical neuroscience. Von Neumann saw the neural coding problem as analogous to a “brain language.”

This paper discusses the implementation of linking field synapses into the Wilson model of cortical neurons. To our knowledge nobody has investigated the effects of implementing such synapses into biologically accurate models of neural networks. The reason for this investigation is to propose a biologically accurate model for synchronization observed in the visual cortex of cats by Reinhard Eckhorn (Eckhorn 1990). Eckhorn has previously modeled this phenomenon using a linking field (LF) that links several neurons together to facilitate “group firing”, and as such, synchronization. The Eckhorn model is abstract and distance from biological parameters. The model is successful at modeling synchronization but not in a biologically-known manner. The main objective in our investigation is to take the LF from Eckhorn’s model and apply it to the Wilson model to achieve biologically interpretation of parameters while keeping the synchronization phenomenon intact.

Novel frequency-selective recurrent neural networks constructed from Eckhorn’s neuron model are described and a simple and precise theory of their operation is given. The networks are organized in basic cell groups, and larger networks are constructed by interconnecting these groups. Individual groups are modeled along the lines of the columnar organization of connectivity reported for neocortex. The cell groups are designed to be pulse rate selective “highpass filters”. “Bandpass filters”, i.e. assemblies that respond over a specific range of synchronous input pulse rates, are realized using pairs of highpass structures connected in a feedback arrangement. Signal propagation in chains of these networks is studied and certain interesting neurodynamics are described and explained. These include wave and packet generation and propagation characteristics, a rate-multiplier effect, evanescent wave propagation, dipulse propagation, and subharmonic, harmonic and anharmonic spectrum generation by these networks.

What is the nature of the cortical organization that gives rise to binding codes in the neural representation of objects and events? Unfortunately, the details of cortical circuit connection and many of the details of cortical cell physiology are far from completely understood. Furthermore, none of the current hypothetical structures proposed by various researchers are free from disagreement with experimental data at the detail level. In this paper those facts that are generally agreed to at the present time by a majority of neuroscientists are reviewed with the goal of putting together a comprehensive picture of the role of functional column structure in the neocortex. We first review the state of knowledge of neocortical organization. We then review the mathematical theories of oscillator and wave models in the context of functional column structure and discuss their implications for larger-scale modeling of object and event binding through retroactive multiregional feedback. In the course of this review, we will examine the implications for the chain-of-oscillators, synchrony, and wave propagation paradigms for modeling neural binding code mechanisms.

This paper is a tutorial review of the structure, composition, and statistical modeling of the organization of the neocortex. It begins with a general overview of the layered structure of the neocortex and its organization as a network of interconnected functional columns. Next it discusses the various classes of neurons that populate the neocortex using as a classification system the several generic types of signals produced by cortical neurons. This is followed by a discussion of characteristics in neuron-to-neuron signaling. Finally, it reviews some of the general trends found in the cortical organization.

This white paper describes the general objectives and organization for foundational research in computer architecture derived from models of the dynamical organization of cortical and non-cortical bio-circuits in the central nervous system (CNS). Our principal General Objectives are: 1) to understand how this dynamical organization in brain structure can be viewed in a computational context; 2) to discover what implications dynamical organization holds for information processing; and 3) to develop a new class of computing devices organized and designed to take advantage of these findings. Within the framework of these General Objectives there are Discipline Objectives in support of these General Objectives. This research will be carried out by an interdisciplinary team composed of researchers drawn from the disciplines of microelectronics, computer engineering, computer science, neurobiology, and computational neuroscience. The discoveries and findings from this research are expected to make fundamental contributions to the art and science of computing technology as well as to our basic knowledge of brain organization and the biological substrates of intelligence. It is expected that the experimental and theoretical findings of this research will profoundly affect the future course of computer design and computer science, and will establish a new break-through paradigm leading to more truly intelligent and robust computing devices. It is not improbable that the foundations laid by this research will lead to an entirely new industry of biomimetic neurocomputer systems based on dynamic link architectures.

A new graph-theoretic approach to neural structure identification with implications for information-theoretic optimization is proposed for dynamic link architecture neural networks. The method addresses the organization and interconnection of feature-representing multicellular units (MCUs) at the network architecture level. It incorporates indices for the identification of both dynamic and static data pathway and binding code links between neurons within an MCU and between different MCUs. It also proposes a new objective function for optimization of a measure of conditional entropies and mutual information properties of the system. The method is based on an extension of a method of graph-theoretic information flow analysis recently proposed by Akuzawa and Ohnishi, here called Akuzawa-Ohnishi Analysis (AOA).

This paper is a preliminary discussion of a new paradigm for the general architectural, information representation, operational and design strategy for a biologically-inspired, general purpose neurocomputer based on pulse-coded neural network methods. It presents both hardware and foundational “psychological” machine structure, and discusses key issues raised by this new paradigm. The model is based on experimental findings from both neuroscience and from developmental psychology, and proposes that the key differences between a general purpose neurocomputer and a general purpose digital computer lie in the operational characteristics of the machine that come from attempting to mimic the processes discovered by these sciences. Like most discussions of computer architecture, the model presented here is generally qualitative rather than quantitative. It does however provide a general framework for subsequent quantitative and mathematical research.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.

Entry needed.