Flexible Neuromorphic Electronics: From Synaptic Devices Toward Sensing-Memory-Computing Circuits
Neuromorphic electronic systems, inspired by the brain's parallel and distributed information processing, have emerged as a promising alternative to von Neumann architectures, which suffer from the memory wall and limited energy efficiency. However, seamless integration with biological tissue, particularly human skin, necessitates mechanical flexibility and conformability. Flexible neuromorphic electronics, combining neuromorphic computing with flexible substrates, enable brain-inspired processing with high efficiency and ultralow power consumption, targeting smart wearables, digital health, and brain-computer interfaces. Despite rapid advances in flexible synaptic devices, a cross-disciplinary synthesis connecting materials, device physics, circuit integration, and applications is lacking. This review systematically follows a device-to-system framework, first summarizing recent progress in flexible artificial synapses that emulate neural functions, then discussing neuromorphic circuits and systems, focusing on collaborative sensing-computing and heterogeneous integration strategies toward integrated sensing-memory-computing systems. Emerging applications in next-generation bio-intelligent systems, including wearables, health monitoring, and human-machine interaction, are highlighted. Key challenges and future directions are summarized to guide development of efficient, intelligent, and biocompatible bionic hardware. The review underscores the need for standardized metrics and scalable manufacturing to translate laboratory prototypes into practical flexible neuromorphic systems.