ABSTRACT
The Hodgkin-Huxley equation provides the crucial insight for the electronic circuit designers that models of neural networks can be constructed from electronic devices or circuits whose conductance may be varied as a function of voltage or current values. The simplest such device is a single field effect transistor (FET) which, when operated below pinchoff, acts as a voltage-controlled conductance. This chapter shows feedback and feedforward networks which can be built from these elements. It describes alternative methods for hardware implementation of neural network models. The chapter describes hardware circuits that are inspired by the basic membrane equation which suggests synaptic conductance modulation as a mechanism for achieving shunting feedback and feedforward network interaction. The hardware models are abstraction tools which help study the behavior of the networks; mean intensity dependence and the effect of variation of inhibition to excitation strengths.
