AI framework speeds up brain neuron modeling

Cedars-Sinai investigators worked with a multi-institutional team to develop a new artificial intelligence framework that can accurately, quickly and efficiently create virtual models of brain neurons. The achievement could accelerate discoveries in brain function research and ultimately lead to better treatments for brain disorders.

The study’s findings were presented at the 39th Conference on Neural Information Processing Systems (NeurIPS 2025) in San Diego. The paper is available on the arXiv preprint server.

“Computational modeling of brain neurons has become an important tool for studying their activity and interactions,” said co-author Costas Anastassiou, Ph.D., associate professor of Neurology, Neurosurgery and Biomedical Sciences at Cedars-Sinai. “But traditional models are hindered by limitations including the cost of computer resources, data availability and cumbersome handling.

“Our new framework tackles this problem by operating at speeds thousands of times faster than existing methods while remaining so biologically accurate that it can capture the variability of actual brain neurons, unlike current approaches. The framework can also generate an unlimited number of virtual neurons, better reflecting the diversity and variability of actual biological neurons.”

Anastassiou added that the framework opens a pathway toward modeling larger-scale brain circuits that could improve understanding of the relationships between gene expression, electrical activity and networking in neurons. The investigators named their invention NOBLE, for Neural Operator with Biologically-informed Latent Embeddings.

“I am very happy to see this interdisciplinary and inter-institutional collaboration,” said Anima Anandkumar, Ph.D., Bren Professor of Computing and Mathematical Sciences at Caltech and a co-author of the study. “Neural operators are designed to capture the complex dynamics seen in biological neurons, and this is the first large-scale AI framework validated with experimental human cortex data.”