Multifunctional Organic Materials, Devices, and Mechanisms for Neuroscience, Neuromorphic Computing, and Bioelectronics

Abstract

Neuromorphic computing has the potential to overcome limitations of traditional silicon technology in machine learning tasks. Recent advancements in large crossbar arrays and silicon-based asynchronous spiking neural networks have led to promising neuromorphic systems. However, developing compact parallel computing technology for integrating artificial neural networks into traditional hardware remains a challenge. Organic computational materials offer affordable, biocompatible neuromorphic devices with exceptional adjustability and energy-efficient switching. Here, the review investigates the advancements made in the development of organic neuromorphic devices. This review explores resistive switching mechanisms such as interface-regulated filament growth, molecular-electronic dynamics, nanowire-confined filament growth, and vacancy-assisted ion migration, while proposing methodologies to enhance state retention and conductance adjustment. The survey examines the challenges faced in implementing low-power neuromorphic computing, e.g., reducing device size and improving switching time. The review analyses the potential of these materials in adjustable, flexible, and low-power consumption applications, viz. biohybrid spiking circuits interacting with biological systems, systems that respond to specific events, robotics, intelligent agents, neuromorphic computing, neuromorphic bioelectronics, neuroscience, and other applications, and prospects of this technology.

1. Introduction

Deep learning and artificial intelligence algorithms are becoming increasingly essential in various applications, despite their similarities to the human brain. These algorithms are executed at a software level and rely on artificial neural networks, which are typically implemented on traditional von Neumann architecture computers [1–3]. The brain’s physical components, consisting of an interconnected and complex neuron system, function efficiently due to synapses’ chemical movements that control signal strength. The Hebbian learning principle posits that concurrently activated neurons create connections, serving as the foundation for memory and learning. This ability to adapt and change is primarily responsible for processing information within the brain, rendering it highly energy efficient compared to conventional computers, particularly in pattern classification and recognition.

Memristive devices, or memristors, are resistive switches with variable yet non-volatile electrical resistivity or optical reflectivity. These devices have been proven to exhibit synaptic functionality resembling Hebbian learning. They generate non-volatile memory arrays, perform basic pattern recognition task, and process information in hardware while consuming low energy. Researchers have demonstrated that neural network algorithms can be integrated into hardware, mimicking the brain’s efficiency and function on a small system [4–6].

To simulate the brain’s parallel operation, a memristive device-based network has to exhibit efficient parallel vector–matrix multiplication, symmetrical and linear programmable conductance states, and low energy consumption [7–9]. These features enable “blind” synaptic weight updates during learning. State-retention time requirements vary depending on the application, but longer times are generally preferred. For continuous learning scenarios, synaptic weights are regularly transferred to external memory, while for train-once inference-only applications, they are stored on-system for an extended period.