Researchers have combined MRI scans with physics-informed artificial intelligence to estimate fluid flow in the brain that helps clear metabolic waste. The glymphatic system, first described in 2012, is linked to diseases such as Alzheimer’s. Professor Douglas Kelley of the University of Rochester explains that MRI can image whole brains but cannot capture the very slow flow velocities in living tissue.
The team trained neural networks on videos showing dye spreading across brain tissue. From those videos the AI deduced both how fast the fluid moves and how permeable the tissue is. The study appears in Science Advances.
The results show two main removal pathways: a faster route around open regions such as the brain surface and a much slower flow through deep tissue. So far the researchers have baseline measurements in animals like mice and plan to compare healthy and sick brains and to study humans in future work.
Difficult words
- physics-informed — computer methods that use physical laws
- artificial intelligence — computer systems that learn from data
- glymphatic system — brain network that removes waste
- metabolic waste — waste products made by body cells
- permeable — allowing liquid or gas to pass through
- pathway — a route for movement or flowpathways
- baseline — basic measurement used for later comparison
- neural network — computer model made of connected unitsneural networks
- flow — movement of liquid or gas in space
Tip: hover, focus or tap highlighted words in the article to see quick definitions while you read or listen.
Discussion questions
- Why is it useful to have baseline measurements in animals before studying humans?
- How could knowing the brain's removal pathways help with diseases like Alzheimer’s?
Related articles
Keeping healthy habits during the holidays
The holiday season often disrupts regular health routines. Samantha Harden of Virginia Tech recommends realistic, small changes—like habit stacking, gamifying activities, involving others and travel tips—and a broader view of well‑being so breaks are not failure.
New training method helps models do long multiplication
Researchers studied why modern language models fail at long multiplication and compared standard fine-tuning with an Implicit Chain of Thought (ICoT) method. ICoT models learned to store intermediate results and reached perfect accuracy.
