Analog Very Large Scale Integration Implementation of Dendrite Segment with Voltage Dependent Spiking Behavior

I. INTRODUCIION The first silicon retina, designed by Mead and Mahowald in the late 1980's [I], was an analog Very Large Scale Integration (aVLSI) system. The success of early Integrated Circuit (IC) implementations of biologically accurate neural circuits led to the exploration of many other computational systems, as well as work on development of silicon models of single neurons [2]. These aVLSI systems are so successful because the mechanisms for signaling in the neural system, which are govemed by B o l t " statistics, can be captured by Metal Oxide Semiconductor Field Effect Transistor (MOSFEn circuits operating in their sub-threshold or low voltage regime. aVLSI is used in implementation of the silicon neurons and neural circuits because computations like addition, subtraction, expansion and compression are natural for analog circuits. It is these natural computational capabilities that make analog circuits so desirable for implementing biological functions. Since biological neurons and aVLSI circuits have similar data processing strategies, an analog circuit is a good representation of a biological system. The technology used for these ICs is Complimentary Metal Oxide Semiconductor (CMOS), a metbod of design that bas many advantages for simulating biological functions [3]. Some of these advantages include: wide use in data processing applications, cost-effectiveness, well-understood fabrication processes, and low power consumption.