Author(s):
Saanvi Suri and Nitish Saini

Abstract:
The ever-growing demands of AI and data-driven computing expose the inefficiencies of conventional CMOS, HPC, and AI workloads (GPUs & TPUs), which suffer from the von Neumann bottleneck. Bio-memristors offer a transformative alternative, merging memory and computation for real-time, energy-efficient processing. Inspired by synaptic plasticity, they utilize Electrochemical Metallization (ECM) and Valence Change Mechanism (VCM) for adaptive, multi-level conductance, key to neuromorphic computing. Recent advances in biomaterial-based memristors—incorporating plantderived cellulose nanofibers, saccharide-based electrolytes, DNA-based switching, and protein-assisted charge transport—enhance sustainability and biocompatibility while replicating parallel processing and in-memory computing. Additionally, quantum conductance effects enable ultra-precise, low-power synaptic modulation, further bridging artificial and biological intelligence. This review explores memristor evolution, key switching mechanisms, and bio-inspired designs, categorizing bio-memristors based on their resistive switching behavior and highlighting applications in neuromorphic AI, neuroprosthetics, and energy-efficient IoT. Finally, it addresses challenges in scalability, integration, and ethical considerations, paving the way for computing systems that learn and evolve like the human brain.

Pages: 873-898

Read Full Article