Ion-modulated oxide-based neuromorphic transistors for spatiotemporal information processing

Abstract

Unlike energy-intensive von Neumann systems, the human brain efficiently processes complex spatiotemporal information utilizing slow, dissipative ionic dynamics. To emulate this, ion-modulated oxide-based neuromorphic transistors have emerged as a compelling hardware platform, because they combine intrinsic ionic time constants with the scalability and functional versatility of oxide electronics. This review establishes a physical and architectural roadmap for spatiotemporal information processing in these devices by linking biological ionic mechanisms to modulation pathways, transistor structures, and representative computing functions. We show how ion-modulated oxide transistors evolve from basic temporal processing units to multi-terminal sensory fusion elements and ultimately to array-level adaptive computing hardware. Finally, we highlight the key bottlenecks and actionable future directions for achieving task-matched ionic dynamics, scalable integration, and real-time bio-inspired spatiotemporal intelligence.