Observation, Characterization and Modeling of Memristor Current Spikes

TL;DR

This paper demonstrates that memristors naturally produce fast, reproducible current spikes suitable for neuromorphic computing, supported by experimental evidence and theoretical modeling.

Contribution

It provides experimental evidence of memristor current spikes and develops models linking these spikes to memristive effects for neuromorphic circuit design.

Findings

Memristors exhibit fast, reproducible current spikes.

Current spikes can be modeled using memristor theories.

Spikes are suitable for neuromorphic architectures.

Abstract

Memristors have been compared to neurons (usually specifically the synapses) since 1976 but no experimental evidence has been offered for support for this position. Here we highlight that memristors naturally form fast-response, highly reproducible and repeatable current spikes which can be used in voltage-driven neuromorphic architecture. Ease of fitting current spikes with memristor theories both suggests that the spikes are part of the memristive effect and provides modeling capability for the design of neuromorphic circuits.