A new machine learning algorithm has identified 478 compounds with the potential to reverse brain aging and protect against neurodegenerative diseases. The breakthrough, led by researchers at the Luxembourg Centre for Systems Biomedicine and CIC bioGUNE, could accelerate the development of drugs to combat cognitive decline as the global population rapidly ages.
The Aging Brain and the Transcriptome
The core problem is clear: aging is the biggest risk factor for neurodegenerative diseases like Alzheimer’s. With over two billion people expected to be over 60 by 2050, finding ways to protect brain health in later life is no longer just a scientific challenge – it’s a public health imperative. This research doesn’t focus on genes themselves, but on the transcriptome —the RNA molecules that show which genes are active. This approach is more dynamic than just looking at DNA, because gene activity changes with age and disease.
How the Algorithm Works
Researchers analyzed brain samples from 778 healthy individuals aged 20 to 97. The machine learning model learned to predict biological age with remarkable accuracy (within five years) based on the activity of just 365 key genes. Surprisingly, most of these genes aren’t directly involved in brain function; they regulate DNA repair and overall aging processes. This suggests that slowing down aging systemically can protect the brain.
When applied to samples from patients with Alzheimer’s or traumatic brain injury, the algorithm consistently showed their brains were biologically older than expected—sometimes by as much as 15 years in people aged 60-70. This confirms that neurodegeneration is linked to accelerated aging at the molecular level.
Identifying Rejuvenating Compounds
The algorithm then scanned data from thousands of neurons, searching for gene expression patterns that reduced predicted age. The result: a list of 478 drugs with potential rejuvenating effects. While many of these compounds haven’t been tested for lifespan extension, the algorithm’s predictions are a starting point for targeted research.
Early Validation in Mice
To test the model’s accuracy, researchers gave three of the predicted compounds to old mice over four weeks. The mice showed significant improvements in memory and reduced anxiety, along with genetic changes that made their brain cells appear younger. This suggests that the algorithm isn’t just identifying correlations, but causal links between compounds and brain rejuvenation.
The Future of Anti-Aging Drug Discovery
Currently, the anti-aging field lacks systematic methods for drug development. This machine learning platform offers a structured way to identify promising candidates. However, the identified compounds require further validation across multiple biological systems before they can be considered effective treatments.
The hundreds of compounds predicted by this platform represent an extensive opportunity for future research and therapeutic development.
The goal is clear: to develop drugs that not only slow down aging but actively reverse it, protecting brain function for a growing population. The systematic approach provided by this new machine learning method represents a major step toward that future.
