Engineering the Synaptic Kinetic Process into Memristive Device. (30th April 2018)

• Record Type: Journal Article
• Authors: Zhang, Ziyang; Wang, Yaoyuan; Li, Huanglong; Wu, Yujie; Wang, Guanrui; Shi, Luping
• Is Part Of: Advanced Electronic Materials. Volume 4:Number 6(2018)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2018-04-30
• Subjects: diffusive memristors -- dynamics -- memristors -- subcellular dynamics -- synaptic kinetics; Materials -- Electric properties -- Periodicals; Materials science -- Periodicals; Magnetic materials -- Periodicals; Electronic apparatus and appliances -- Periodicals; 537
• DOI: 10.1002/aelm.201800096
• Languages: English
• ISSNs: 2199-160X

Abstract

Abstract: The important synaptic functionalities, such as long‐term memory, short‐term memory, and experience‐dependent learning, are extensively emulated by the emerging memristive devices, which provide a main step toward brain inspired computing. The majority of the electronic synapses based on memristors usually focus on the emulation of the physiological behaviors at the cellular level. However, the research on the subcellular synaptic dynamic properties of electronic synapses is rarely reported. In this article, Ag/Ta2 O5 /Pt‐based memristors are utilized to emulate both the long‐term memory and short‐term facilitation of biological synapse. Under pulse‐train measurements, the device exhibits short‐term facilitation depending on varying the electrical pulse intensity, duration, and frequency. Based on the experimental results, one‐to‐one correspondence is established between the filament dynamics in our device and the subcellular synaptic dynamics during short‐term facilitation. By fitting the experimental results into the biological model, the dynamic displacements of the rear and head ends of the Ag filament are drawn analogy to the neurotransmitter and Ca 2+ dynamics. The dynamic process is explained by simulations of the nanoparticle dynamics in which the Ag filament is simplified to a 1D "stick." This research enables the emulation of synaptic function with improved biofidelity and broad applications to neuromorphic computing. Abstract : The synaptic dynamic process of memristive device is systematically investigated by fitting the experimental observations with Tsodyks and Markram model as well as nanoparticle dynamic simulation. One‐to‐one correspondence between the filament dynamics in memristive device and the subcellular synaptic dynamics during short‐term facilitation is established.