Wu_2026_Nat.Neurosci_29_234

Synthesizing perceivable artificial neural inputs independent of typical sensory channels remains a fundamental challenge in developing next-generation brain-machine interfaces. Establishing a minimally invasive, wirelessly effective and miniaturized platform with long-term stability is crucial for creating research methods and clinically meaningful biointerfaces capable of mediating artificial perceptual feedback. Here we demonstrate a miniaturized, fully implantable transcranial optogenetic neural stimulator designed to generate artificial perceptions by patterning large cortical ensembles wirelessly in real time. Experimentally validated numerical simulations characterized light and heat propagation, whereas neuronal responses were assessed by in vivo electrophysiology and molecular methods. Cue discrimination during operant learning demonstrated the wireless genesis of artificial percepts sensed by mice, where spatial distance across large cortical networks and sequential order-based analyses of discrimination predicted performance. These conceptual and technical advances expand understanding of artificially patterned neural activity and its perception by the brain to guide the evolution of next-generation all-optical brain-machine communication.