Catenaa, Friday, November 07, 2025-Researchers at the University of Southern California have developed artificial neurons that replicate the behavior of real brain cells, marking a major advance toward energy-efficient, brain-like computing.
The study, published in Nature Electronics, describes how the team built “diffusive memristor” neurons that mimic the brain’s electrochemical signaling using atoms instead of electrons. Each neuron occupies the space of a single transistor, potentially reducing chip size and energy use by several orders of magnitude.
Led by USC engineering professor Joshua Yang, the researchers demonstrated that silver ions moving within the device emulate biological ion flows that drive neural activity.
This physical replication allows the neurons to perform computation as efficiently as natural systems, offering a path toward neuromorphic hardware that learns and processes data like the human brain.
The breakthrough could accelerate the development of artificial general intelligence by enabling chips that compute and learn through hardware rather than software.
The system’s compact design also supports scaling for use in future AI devices that consume far less energy than current digital processors.
Yang said the next step will be integrating large networks of these artificial neurons to test how closely they match the brain’s adaptability and power efficiency. The work was supported by the U.S. National Science Foundation and the Army Research Office.
