Watch the video in this story. The images are groundbreaking. The video demonstrates that the communication pathways (fibre pathways) in the brain's white matter are organized in layers that run parallel to each other and interlock like a zipper, even when they cross. This new knowledge is based on many years of pure research at DTU Compute and Hvidovre Hospital. The findings, now presented through a video, are attracting considerable attention in the fields of computer science and neuroscience.
"For years, researchers worldwide have debated how the brain's networks are organized. It was known that many fibre pathways go from point A to point B. However, the interesting aspect was how they crossed each other. It was believed that there was some chaos, with fibre pathways lying haphazardly in the soft, white brain matter surrounded by cerebrospinal fluid. But this is not the case. The organisation is much less complicated," says Professor and Research Leader Tim Dyrby.
"Although this is basic research with tissue samples from animals, it could potentially impact the treatment of neurological diseases such as Alzheimer's or ALS, where fibre pathways are degraded, and early diagnosis and intervention are crucial."
Mathematical Models Zoom in on the Structure
The brain functions through neurons (nerve cells) that communicate with other neurons, muscles, and organs via electrical impulses. Signals are transmitted through axons, the long thread-like fibre pathways that extend from neurons. This transmission is crucial for coordinating movements, sensing, and regulating vital body functions.
Today's MRI scans of the brain have such low resolution that only the overall structures can be seen. However, using mathematical models and artificial intelligence (AI), researchers have discovered how to 'zoom in' on the scan images with unprecedented detail, allowing them to see the brain's microstructure and the organisation of axons in 3D.
"An American research group started a huge debate about 13 years ago when they proposed a theory that fibre pathways run parallel to each other and cross each other like a cross when they don't. We see the same effect and can document the theory visually at the finest level simply because we combine different imaging techniques," says Tim Dyrby.
Non-Invasive AI Tool to Detect Disease Early
The image analysis models are developed based on data from brain tissue from deceased animals. The tissue was examined using X-ray synchrotrons in Germany and France. The equipment functions like special microscopes to capture 3D images and can zoom in on individual axons. The data was then compared with images of tissue examined through MRI scans.
Tim Dyrby's DTU team, specialising in image analysis, translated the data into mathematics and statistics and developed AI models to make the invisible visible, allowing them to 'zoom in' on 3D scan images.
"Today, tissue samples are only taken from the brains of deceased individuals. Our hope is that the AI models can function as a non-invasive tool to examine living brain tissue, allowing early detection of disease signs and hopefully treating neurological diseases using an MRI scanner", says Tim Dyrby.
Video Provides Deeper Understanding
The research results are published in the journal eLife and accompanied by the aforementioned video that has attracted attention.
Associate Professor Hans Martin Kjer from DTU Compute used the Blender program to visualize data using colour codes, showing how the axons are organized.
Red indicates one direction, green another, and blue a third. Normally, only tissue slices are shown, but here, the 3D appearance is displayed.
The video, which visualizes the surprising findings, has already gained significant attention in both computer science and neuroscience, with thousands of online views. Recently, Tim Dyrby gave a presentation at an international forum where people were very impressed by the level of detail:
"In neuroscience, 3D visualization of large datasets is a new niche that is growing and has suddenly gained enormous attention because the potential is now visible. Previously, one could show some beautiful 3D data, but no one really knew how to handle the massive data collected from X-ray synchrotrons. We have become much better at this. Now you can visually show some results in 3D, calculate something, and extract information from them."
Tim Dyrby and his team have continuously published parts of the research, but it is only now, after 5-6 years, that they have presented the complete picture, partly through the video.
