Webmaster:
bioweb (at) langate.gsu.edu

Brains & Computers, led by Melody Moore (Computer Information Systems), studies how the brain and nervous system can interface with computers. This includes real-time interactions between neurons and models, brain-computer interfaces via EEGs, and real-time interactions between animals and computers.
Upcoming group meetings

Seed Grants

Principle Investigators: Unil Perera / Yi Pan / Melody Moore
Title: Artificial Neural Networks with Silicon Circuits -- Simulation and Parallel Computing

Abstract:
Silicon p + -n-n + diodes have shown spontaneous pulsing while being driven by current and voltage at 4.2 K. There are similarities between these pulses and biological neuron spikes which are called action potentials. A simple circuit containing this diode with a load resistor and capacitor has been shown to be similar to Hodgking-Huxley equivalent circuits.  Period doubling, bifurcation, chaos, excitation, inhibition, summation over inputs, firing thresholds,  transient sensing, and overshoot, all of which are properties of biological neurons,  have been shown to occur in circuits with these diodes. Based on the theses similarities we have proposed a parallel processor that follows the functional mode of a real neuron in a biological retina. A prototype for this processor has been constructed and measurements have been made on the individual channels comprising it. An iterative map derived from the devise equations for the diode, which simulates a biological neuron was extended to a coupled neuron circuit consisting of two of these artificial neurons connected by a filter circuit, which was used as a single channel of a parallel asynchronous processor. The extended map output was studied under different conditions to determine the effect of the various parameters on the pulse pattern. As the control parameter is increased, fixed points (both stable and unstable) as well as a limit cycle appear. On the further increase, a Hope bifurcation is seen causing the disappearance of the limit cycle.   It will be beneficial to study the collective behavior of  a large number of channels  before implementing on an experimental  processor. In order to study the collective behavior in a larger scale, we first plan to simulate the behavior of the circuits with these Silicon diodes. Due to the time required to simulate the circuits, we also plan to implement the simulation program on a parallel platform.