 |
| Revolutionizing Neuroscience: This image depicts a conceptual breakthrough in brain-mapping technology, featuring a new AI chip. With processing speeds 478 times faster than current standards, it enables real-time cortex reconstruction and diagnoses in under half a second. This innovation holds immense promise for advancing surgical navigation, early Alzheimer's screening, and brain-computer interfaces. This image was generated by AI. |
Breakthrough in Brain-Mapping Technology
Chinese researchers have developed a new artificial intelligence chip capable of reconstructing complex brain structures in real time. This breakthrough has the potential to dramatically accelerate the diagnosis of brain diseases, enhance neurosurgery, and propel the development of advanced brain-computer interfaces.
According to the research team from Peking University and the Chinese Academy of Sciences, the chip can map the brain’s intricately folded cortex in under half a second. For this highly specialized task, the hardware delivers processing performance up to 478 times faster than NVIDIA’s widely used A100 GPU. The findings were officially published in the journal Science.
Combining Memory and Computing
Traditional computer architectures often struggle with these demanding calculations because their memory and processors operate separately. This constant transferring of data creates a major performance bottleneck.
The newly developed 40-nanometre computing chip solves this issue by integrating an artificial neural network directly into its hardware using a computing-in-memory architecture.
The new chip performs data storage and computation within the exact same memory array.
This drastically reduces the delays caused by data transfers between separate units.
By merging these functions, the system can reconstruct the highly folded cortical surface in less than 0.5 seconds.
Depending on the specific brain-mapping workload, it outperforms systems powered by the NVIDIA A100 GPU by a factor of 50 to 478 times.
Harnessing Phase-Change Memristors
One of the key innovations behind the chip’s millisecond-scale processing speed is its use of phase-change memristors.
Historically, engineers have treated "conductance drift" in emerging memory tech as a severe hardware limitation. Instead, the researchers ingeniously repurposed this drift to perform neural dynamic computations. This fresh approach enables high-fidelity calculations, improves overall computing speed, and significantly reduces energy power consumption.
The Future of Brain Disease Diagnosis and Medical Care
Mapping the human brain is computationally demanding because the cortex's complex folded structure dramatically increases its surface area while fitting snugly inside the skull. With this new technology, researchers believe they can unlock several vital medical applications where rapid mapping is essential.
Potential uses include:
Surgical Assistance: Real-time neuronavigation during complex brain surgery.
Disease Detection: Early screening and diagnosis for conditions like Alzheimer’s disease.
Advanced Tech: Faster brain-computer interface development.
Custom Treatments: Personalised brain modelling and clinical decision support.
Lead researcher Yang Yuchao, a professor at Peking University’s School of Integrated Circuits, noted that this breakthrough could eventually enable the creation of personalised "digital brain twins." These highly accurate digital models would provide invaluable support for rapid diagnosis and precision treatment planning.