Researchers in the Department of Neurotechnology at Lobachevsky University have created a neuromorphic network that mirrors how brain cells interact. For the first time, digital stand-ins for astrocytes, the supportive cells in the brain, have been integrated into artificial intelligence systems. This innovation yielded a notable 20% boost in the network’s performance, a development shared with socialbites.ca by officials at the National Center for Physics and Mathematics (NCFM).
The team explains that incorporating astrocytes into neuromorphic AI marks a significant leap in neural network design. In the brain, these cells help manage information flow, adapt as tasks evolve, and support memory. With digital astrocytes, researchers can steer neuronal dynamics more precisely and implement working memory within AI models. The study highlighted that this could be the missing piece in training modern neural networks, addressing long-standing challenges in learning efficiency and adaptability.
Experts note that such networks may shed light on little understood processes behind how information is encoded, transformed, and stored in the brain. They may also illuminate factors that contribute to age-related neurodegenerative conditions and other brain phenomena, offering clues for both science and medicine.
Looking ahead, the use of astrocyte-inspired components could pave the way for a new generation of intelligent information and computing platforms. These platforms would process vast streams of data by leveraging principles observed in brain function, enabling more capable neuromorphic software for medical analytics, IT data classification, and the construction of virtual models of neural systems. The potential applications extend across clinical research, healthcare technology, and data science, aligning with trends toward more brain-like computing architectures.
The funding behind this work came from a grant from the Russian Science Foundation, underscoring the global interest in brain-inspired computing approaches. In related advances, researchers have begun exploring AI methods crafted to forecast survival outcomes for brain cancer patients, illustrating how neuromorphic ideas connect to real-world medical decision making.