Nexus: A Neural Simulator for Integrating Top-Down and Bottom-Up Modeling

Paul Sajda, Ko Sakai, Shih-Cheng Yen, and Leif H. Finkel

Department of Bioengineering and
Institute of Neurological Sciences
University of Pennsylvania
Philadelphia, PA 19104, U. S. A.

Abstract

We have developed the NEXUS simulation environment as a tool for modeling large-scale neural systems. The software is written in C and runs under UNIX. A unique aspect of NEXUS is that it is particularly suited for simulating hybrid neural models (i.e. systems integrating different modeling paradigms and/or architectures.) NEXUS is designed for large-scale simulations, and to facilitate model development, testing and analysis it incorporates several major features: network architectures based on topographic maps, programmable neural units, scalable and modular simulation, support for common learning paradigms including the generalized Hebb rule and backpropagation, and a user-friendly interface. These features make NEXUS a useful environment in which to study the ``perceptual'' properties of various network architectures.

[ Full Text in PDF ]