Bio-Inspired Spike-Timing-Dependent Plasticity Learning with Metal Halide Perovskites: Toward Artificial Synaptic Functionality
Recent advances in neuromorphic engineering have sparked a convergence between nanotechnology and neuroscience, where emerging devices such as memristors are being explored to replicate fundamental learning mechanisms observed in the brain. One such mechanism, spike-timing-dependent plasticity (STDP...
| Autores: | , , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2026 |
| País: | España |
| Institución: | Universidad de Sevilla (US) |
| Repositorio: | idUS. Depósito de Investigación de la Universidad de Sevilla |
| OAI Identifier: | oai:dnet:idus________::7e198c451cefa81cb23203d33fdb99aa |
| Acceso en línea: | https://hdl.handle.net/11441/186080 https://doi.org/10.1021/acsami.5c21545 |
| Access Level: | acceso abierto |
| Palabra clave: | halide perovskite memristo spike-timing-dependent plasticity (STDP) neuromorphic computing synaptic plasticity noise robustness triplet-STDP |
| Sumario: | Recent advances in neuromorphic engineering have sparked a convergence between nanotechnology and neuroscience, where emerging devices such as memristors are being explored to replicate fundamental learning mechanisms observed in the brain. One such mechanism, spike-timing-dependent plasticity (STDP), encodes synaptic changes based on the precise timing between pre- and postsynaptic spikes, and has been widely adopted in machine intelligence and computational neuroscience. In this work, we demonstrate that a halide perovskite memristor (Cs3Bi2I6Br3) can effectively simulate biologically plausible STDP dynamics. We fabricate and characterize the MHP-based device, and develop a dynamic physical model capturing its voltage- and history-dependent switching behavior. Using biologically inspired biphasic voltage pulses, the model replicates classic STDP characteristics including long-term potentiation (LTP), long-term depression (LTD), and the canonical asymmetric learning window. Further analysis shows that the memristor supports advanced features such as triplet-STDP and synaptic memory consolidation. Importantly, the STDP behavior remains stable across 100 independent trials with biologically realistic voltage noise, exhibiting less than 0.03% variation in synaptic weight. These results suggest that the inherent physical dynamics of halide perovskites enable bioinspired learning without external programming or algorithmic supervision. By bridging molecular-scale materials physics with spike-based computation, our findings lay the groundwork for implementing scalable, low-power, and noisetolerant synaptic learning in next-generation neuromorphic computing systems |
|---|