Advancing Biosensing With Memristive Technologies and Machine Learning: A Review

Abstract

The rapid advancement of the Internet of Things (IoT) explains the increasing demand for advanced computing technologies capable of processing massive datasets. Nowadays, memristive devices are emerging widely for efficient signal biosensing and edge-computing applications. Resistive switching enables memory and computation co-location, thereby overcoming data transfer challenges within traditional complementary metal–oxide–semiconductor (CMOS) technology and von Neumann architectures. Memristors, as nonvolatile memory structures, provide ultrafast and energy-efficient resistive switching required for enhancing electrochemical biosensing sensitivity by facilitating rapid and low-noise signal transduction. Memristors offer significant synaptic plasticity for neuromorphic complex computing, thus confirming the viability of memristive neuromorphic tools for artificial intelligence (AI)-based biosensing systems and human perception paradigms. This article critically examines recent advances in material engineering and device benchmarks, highlighting the importance of resistive switching mechanisms for enhancing bio/chemical sensing systems. In addition, this work reports the crucial role of memristive neuromorphic adaptability in AI-based biosensing systems and human perception toward efficient disease diagnosis and neuroprosthetics. By discussing these points, this work succeeded in linking the prospective advancement of memristive biosensing to memristive neuromorphic tools and machine learning (ML)-based healthcare systems. Empowering novel technologies with memristive switching suggests that vast innovative approaches toward healthcare systems are feasible, highlighting high-performing memristive bio/chemical and electrical biosensors with efficient electrical and thermal signal integrity.