April 22 - 26, 2024
Seattle, Washington
May 7 - 9, 2024 (Virtual)
Symposium Supporters
2024 MRS Spring Meeting
SB10.09.08

Synaptic Plasticity on Demand for Neuromorphic Computing and Bio-Inspired Artificial Nerve

When and Where

Apr 25, 2024
4:30pm - 5:00pm
Room 429, Level 4, Summit

Presenter(s)

Tae-Woo Lee, Seoul National University

Co-Author(s)

Tae-Woo Lee1,Dae-Gyo Seo1,Gyeong-Tak Go1,Min-Jun Sung1

Seoul National University1

Abstract

Tae-Woo Lee1,Dae-Gyo Seo1,Gyeong-Tak Go1,Min-Jun Sung1

Seoul National University1
Biological nervous systems possess versatile attributes, serving diverse functions; for instance, the central nervous system (CNS) governs learning and memory, while the peripheral nervous system (PNS) is responsible for sensory perception. Consequently, there is a need to engineer artificial synapses tailored to adapt to performance requirements in various applications. Brain- inspired neuromorphic computing aims to emulate the learning and memory capabilities of the CNS, being inspired from the long-term potentiation (LTP) observed in biological synapses. The application of artificial synapses in the fields of nervetronics and neuroprosthetics requires the emulation of the short-term plasticity (STP), enabling the rapid signal transmission and fast responses akin to those in the biological PNS. To demonstrate the broad applicability, spanning areas including neuromorphic computing and bio-inspired nervetronics, our study has explored the modulation of STP and LTP using ion-gel gated polymer synaptic transistors (IGPSTs). We have modulated the polymer semiconductor (PSC) film’s crystallinity through post-deposition film annealing, self-assembly monolayer treatment, and the introduction of various sidechain length, leading to the conversion between STP and LTP properties in IGPSTs. Moreover, we have utilized a straightforward yet effective approach by blending two PSCs with the same backbone but different sidechains. This blend strategy has led to a substantial improvement in LTP characteristics, whereas IGPSTs employing each PSC individually show only STP properties. IGPSTs with enhanced LTP properties demonstrate their potential as neuromorphic computing devices, effectively simulating learning processes in artificial neural networks. Concurrently, IGPSTs featuring STP capabilities are used to demonstrate various artificial nervous systems, such as artificial reflex arcs, neuromuscular systems, and neuro-prosthetic nerves incorporating artificial proprioceptors. These pioneering studies on neuromorphic devices have expanded the scope of applications for artificial synapses and validate the feasibility of these innovative technologies.

Keywords

biomaterial

Symposium Organizers

Simone Fabiano, Linkoping University
Sahika Inal, King Abdullah University of Science and Technology
Naoji Matsuhisa, University of Tokyo
Sihong Wang, University of Chicago

Symposium Support

Bronze
IOP Publishing

Session Chairs

Simone Fabiano
Songsong Li

In this Session