Researchers at University College London have developed NextBrain, an AI-assisted brain atlas that enables unprecedented detailed visualization of the human brain, representing a major advancement in neuroscience and neuroimaging.
Published in Nature, this new tool allows analysis of live patient MRI scans with microscopic-level precision previously achievable only through post-mortem tissue examination.
The atlas captures hundreds of interconnected brain regions and their subregions, providing crucial insights into how these structures change during development, aging, and disease progression.
Six-Year Development Process
The creation of NextBrain involved a meticulous six-year process resembling an intricate puzzle assembly. Using post-mortem tissue from five human brains, researchers carefully dissected each brain into 10,000 sections, stained them for structural identification, captured microscopic images, and reconstructed them into 3D digital models.
Prior to dissection, each brain underwent MRI scanning to provide a reconstruction blueprint. Artificial intelligence played a crucial role in aligning microscopic images with MRI data, accounting for technical variations and ensuring seamless integration without gaps or overlaps.
Technical Innovation and AI Integration
The atlas development leveraged advanced AI algorithms to accelerate what would have taken decades manually. Researchers labeled 333 distinct brain regions across each of the five digital 3D models, creating a composite “average” brain that generalizes to all adults.
This approach enables automatic inference of detailed structural information from MRI scans of both living and deceased subjects, bridging the critical gap between microscopic histology and live imaging capabilities.
Validation and Performance Testing
NextBrain underwent rigorous validation across thousands of MRI datasets, demonstrating reliable brain region identification across varying imaging conditions and scanner types. In one experiment
the atlas automatically labeled regions in ultra-high-resolution MRI scans with remarkable accuracy matching manual labeling, even for small subregions like hippocampal subdivisions. Another test involving over 3,000 live MRI scans enabled more detailed analysis of age-related volume changes than existing tools permit.
Clinical Applications and Research Potential
“NextBrain provides an unparalleled map of brain cytoarchitecture,” stated Dr. Zane Jaunmuktane from UCL’s Queen Square Institute of Neurology. “This opens doors to detecting the earliest signs of neurological disorders like Alzheimer’s long before symptoms appear, enhancing our ability to understand, monitor, and ultimately prevent these devastating diseases.”
The atlas allows researchers to rapidly identify hundreds of brain regions in living patients while maintaining histological-level precision, significantly accelerating neuroscience discovery.
Accessibility and Implementation
All underlying data, tools, and annotations used in NextBrain have been publicly released through the FreeSurfer neuroimaging platform, along with visualization tools and educational resources. This open-access approach ensures global researchers can immediately benefit from the technology.
The atlas integration with established platforms like FreeSurfer facilitates seamless adoption into existing research workflows while providing new capabilities for detailed brain structure analysis in both clinical and research settings.
Related topics
