Controlling Synaptic Plasticity via Metal-Organic Frameworks to Develop an Artificial Retina for Information Filtering

Artificial synapse-based neuroprosthetic systems have emerged as a promising strategy for restoring impaired sensory functions. However, an essential challenge remains the ability to selectively recognize meaningful stimuli while suppressing irrelevant information, similar to the attention mechanisms of the human nervous system. In this work, we introduce an information-selective artificial retina that combines artificial synaptic devices with a signal integration module to achieve both sensory perception and attention-inspired signal processing. The synaptic characteristics are regulated by metal–organic framework (MOF) layers, where the temporal response is tailored through the structural properties of the MOFs. Four isoreticular Zr-based MOFs containing identical Zr₆O₄(OH)₄ secondary building units were systematically investigated to reveal the relationship between framework architecture and synaptic behavior. MOFs with smaller pore sizes preferentially promote short-term synaptic responses, whereas larger-pore structures with stronger ionic interactions facilitate long-term signal retention. By integrating short-term and long-term artificial synapses into a 6 × 6 pixel artificial retina together with a signal integration device, the proposed system effectively performs signal accumulation and selectively enhances significant visual information. This attention-inspired artificial retina closely mimics biological sensory information processing and offers a promising platform for next-generation neuroprosthetic devices and intelligent neuromorphic systems.