A VLSI Array of Low-Power Spiking Neurons and Bistable Synapses With Spike-Timing Dependent Plasticity

Abstract

Abstract-We present a mixed-mode analog/digital VLSI device comprising an array of leaky integrate-and-fire (I&F) neurons, adaptive synapses with spike-timing dependent plasticity, and an asynchronous event based communication infrastructure that allows the user to (re)configure networks of spiking neurons with arbitrary topologies. The asynchronous communication protocol used by the silicon neurons to transmit spikes (events) off-chip and the silicon synapses to receive spikes from the outside is based on the "address-event representation" (AER). We describe the analog circuits designed to implement the silicon neurons and synapses and present experimental data showing the neuron's response properties and the synapses characteristics, in response to AER input spike trains. Our results indicate that these circuits can be used in massively parallel VLSI networks of I&F neurons to simulate real-time complex spike-based learning algorithms.Index Terms-Address-event representation (AER), analog VLSI, integrate-and-fire (I&F) neurons, neuromorphic circuits, spike-based learning, spike-timing dependent plasticity (STDP).

A Real-Time Event-Based Selective Attention System for Active Vision

“…The winning pixel will then produce a constant output current I wta , which is independent of the input, and source it to the pixel's leaky Integrate and Fire (I&F) neuron. This circuit, fully characterized in Indiveri et al 2006 [15], produces voltage pulses (spikes) at a rate which is proportional to it's input current. Each time a spike is emitted from a neuron, the address of the spiking pixel is encoded on a digital bus, instantaneously.…”

Managing Burstiness and Scalability in Event-Driven Models on the SpiNNaker Neuromimetic System

“…Here again there have been two threads of development. In the "neuromorphic" approach [20], chips use analogue circuitry to emulate as closely as possible the actual biophysics of real neurons [60]. The "neuroprocessor" approach [35], by contrast, attempts to use general-purpose digital components with an internal structure optimised for massively parallel neural processing.…”

Burst Synchronization in Two Pulse-Coupled Resonate-and-Fire Neuron Circuits

“…Functional networks of silicon spiking neurons are shown to be useful in a wide variety of applications [5]- [10]. Recent research efforts have concentrated on realtime event-based computation, so-called temporal cording, in which coincidence or synchrony detection plays essential roles in neural information processing, such as auditory perception [6], onset detection [8], and learning and memory [9]- [10].…”

“…Recent research efforts have concentrated on realtime event-based computation, so-called temporal cording, in which coincidence or synchrony detection plays essential roles in neural information processing, such as auditory perception [6], onset detection [8], and learning and memory [9]- [10]. Temporal filtering properties are also significant to extract temporal structure of spike sequences in which information may be encoded.…”

“…These circuits increase selectivity to inputs into a silicon spiking neuron due to synchrony detection and temporal filtering properties. Synaptic circuits with short-term plasticity [8], [10]- [12] and spike-timing-dependent plasticity [10], [13] can also increase computational performance of silicon spiking neural networks. However, they can work effectively only at network level.…”