A new, freely available brain atlas that uses artificial intelligence to map the human brain in unprecedented detail has been developed by a team of international researchers. This new tool, called NextBrain, is poised to revolutionize neuroscience and neuroimaging by allowing scientists and clinicians to visualize the brain’s intricate structures in living individuals with a level of precision that was previously only possible through the examination of post-mortem tissue. The atlas provides a high-resolution, three-dimensional map of the brain’s cellular architecture, which will accelerate research into brain development, aging, and a wide range of neurological and psychiatric disorders, including Alzheimer’s disease.
The NextBrain atlas overcomes a significant limitation of existing brain maps, which can identify major brain structures in magnetic resonance imaging (MRI) scans but lack the ability to delineate their finer sub-regions. These smaller, distinct areas within larger brain structures are often where the earliest signs of disease appear. For example, specific sub-regions of the hippocampus are known to be affected differently during the progression of Alzheimer’s disease. By providing a tool that can quickly and accurately identify these sub-regions in the MRI scans of living patients, NextBrain opens the door to earlier diagnosis, improved monitoring of disease progression, and the development of more targeted treatments for a host of debilitating conditions.
A New Standard in Brain Mapping
The NextBrain atlas represents a major leap forward in our ability to understand the complexities of the human brain. It provides a comprehensive and detailed map of 333 distinct brain regions, many of which were previously indistinguishable in conventional MRI scans. This level of detail is akin to having a high-resolution street map of a city, where previously only a map of the major highways was available. The atlas is designed to be generalizable to all adult humans, meaning it can be used to automatically analyze MRI scans from a diverse range of individuals, regardless of the specific scanner or imaging conditions used. This universal applicability is a key feature that will enable researchers around the world to benefit from this powerful new tool.
Bridging the Gap Between Microscopy and MRI
A fundamental challenge in neuroscience has been the disconnect between the cellular-level detail visible in post-mortem brain tissue under a microscope and the lower-resolution images of the living brain obtained through MRI. NextBrain effectively bridges this gap by integrating the precision of microscopy with the non-invasive nature of MRI. This allows for the in-depth study of brain anatomy in living individuals, which is crucial for understanding how the brain changes over time, both in healthy aging and in the context of disease. The atlas provides an unparalleled map of the brain’s cellular architecture, enabling rapid, accurate, and accessible analysis of brain images in living individuals.
The Six-Year Journey of Creation
The development of the NextBrain atlas was a meticulous and labor-intensive process that took a team of researchers six years to complete. The project involved the use of post-mortem tissue from five human brains, which were donated to science. The researchers have compared the process to assembling a complex, three-dimensional jigsaw puzzle. This painstaking work has resulted in a tool that will save other researchers countless hours of manual labor and provide a standardized framework for brain imaging analysis.
From Tissue Slices to a 3D Model
Before beginning the dissection process, the researchers conducted MRI scans of the intact brains to create a reference image, much like the picture on the front of a jigsaw puzzle box. Each brain was then carefully sectioned into 10,000 thin slices. These slices were stained with special dyes to identify different brain structures and then photographed under a microscope. The high-resolution digital images of these stained tissue slices were then painstakingly reassembled into a three-dimensional digital model of each brain. The final atlas is an average of the five individual brain models, ensuring that it is a robust and representative map of the adult human brain.
Artificial Intelligence at the Core
Artificial intelligence was a critical component in the creation and functionality of the NextBrain atlas. The vast amount of data generated from the 10,000 tissue slices per brain would have been nearly impossible to process and assemble manually. AI algorithms were used to align the microscopic images with the MRI scans, a process that accounted for the differences in the two imaging techniques and ensured that the final 3D model was accurate and free of gaps or overlaps. The use of AI not only accelerated the creation of the atlas but also brought a new level of precision and consistency to the process.
AI-Powered Brain Region Labeling
Once the 3D models were assembled, the researchers began the task of labeling the 333 distinct brain regions. This is a highly specialized and time-consuming process that would have taken decades to complete manually. AI was employed to significantly speed up this process, allowing the team to complete the labeling in a fraction of the time. The AI was trained to recognize the unique features of each brain region, enabling it to automatically identify and label them in the digital models. This AI-powered labeling is also a key feature of the NextBrain atlas in its application, as it can automatically identify and label brain regions in the MRI scans of living individuals in a matter of minutes.
Validating the Atlas with Thousands of Scans
To ensure the accuracy and reliability of the NextBrain atlas, the researchers tested it on thousands of MRI datasets. These tests demonstrated the ability of the atlas to consistently and accurately identify brain regions across a wide variety of imaging conditions and scanner types. In one experiment, the team used the atlas to automatically label brain regions in a publicly available, ultra-high-resolution MRI scan. The results of the automated labeling closely matched the manual labels that had been created by experts, even for very small and difficult-to-distinguish areas such as the sub-regions of the hippocampus.
In another experiment, the researchers applied NextBrain to over 3,000 MRI scans of living individuals to study age-related changes in brain volume. The atlas enabled a more detailed and nuanced analysis of aging patterns than was possible with existing tools. These validation studies have provided strong evidence of the power and potential of the NextBrain atlas to advance our understanding of the human brain.
Open-Source Access and Future Implications
The creators of the NextBrain atlas have made it, along with all the underlying data, tools, and annotations, freely available to the global research community through the FreeSurfer neuroimaging platform. This open-source approach is intended to accelerate discovery in brain science and facilitate its translation into better diagnostic tools and treatments for a wide range of neurological and psychiatric conditions. The availability of this powerful new tool is expected to have a significant impact on research into conditions such as Alzheimer’s disease, epilepsy, schizophrenia, and many others.
The ability of NextBrain to detect subtle changes in brain structure holds the promise of identifying the earliest signs of neurological diseases, long before the onset of clinical symptoms. This could lead to earlier interventions and a better prognosis for many patients. The atlas will also be an invaluable tool for researchers working to understand the fundamental principles of brain organization and function. As the use of the NextBrain atlas becomes more widespread, it is expected to lead to a wealth of new insights into the complexities of the human brain and to pave the way for a new era of precision medicine in neurology and psychiatry.
