A newly developed computer chip from researchers at Peking University and the Chinese Academy of Sciences achieves speeds up to 478 times faster than Nvidia A100 GPU systems in reconstructing complex brain surfaces. The 40-nanometer chip integrates an artificial neural network within its memory, overcoming computational limits and enabling real-time modelling of brain structures, a feat detailed in a study published in Science on Thursday. This advancement promises to accelerate diagnostics for conditions like Alzheimer’s disease and enhance brain-machine interfaces. Lead author Yang Yuchao, a professor at Peking University’s school of integrated circuits, says that this breakthrough opens up new possibilities for brain-computer interfaces and the diagnosis and treatment of brain diseases, and that personalised and dynamic digital brain twins will become possible.

This performance leap stems from the chip’s ability to perform both data storage and computation within a single memory array, overcoming computational limitations. The team anticipates the technology will facilitate intraoperative neuronavigation and early Alzheimer’s disease screening.

It also provides a hardware foundation that can operate in real time for intraoperative neuronavigation [a navigation system for surgery], early screening for Alzheimer’s disease and personalised interventions.

Published in Science on Thursday, the study demonstrates the chip’s ability to reconstruct complex brain surfaces in under half a second, a time considerably shorter than that required by Nvidia A100 GPU systems. The team reports performance gains ranging from 50 to 478 times faster than those systems, overcoming computational limitations previously hindering real-time modelling of brain structures. The chip’s capacity for rapid, accurate modelling suggests a substantial advancement in neurotechnology and its potential clinical applications.