Researchers at the UK Dementia Research Institute in Edinburgh are using artificial intelligence to hunt for treatments for brain diseases that may already exist but have not yet been identified.
The team analyzes patient data, including voice recordings and eye scans, as well as lab-grown brain cells, to see whether drugs already approved for other conditions could be used to treat diseases such as motor neurone disease (MND). By applying algorithms that detect disease patterns and predict which medicines might work, the scientists aim to cut the timeline for finding effective treatments from decades to just a few years.
A patient's perspective
That goal is shared by Steven Barrett, a trial participant who was diagnosed with MND 10 years ago. Barrett had been looking forward to an active retirement after a long career in the civil service when he noticed numbness in his leg. A few years later came the diagnosis of MND, a degenerative neurological condition with no known cure.
"MND is a horrible disease, it strips you of who you are," Barrett told the BBC at his home in Alloa, Scotland. "It rips any sense of future that you may feel that you had planned for yourself. All that goes."
Barrett said his family also did not see the disease coming. He showed the BBC photos of himself at work, at parties, and at his son's wedding. Despite the grim prognosis, he described the clinical trials as a "bright light" of hope for himself and for others living with MND or similar conditions.
One such trial, MND-SMART, tests several drugs simultaneously rather than giving one group a treatment and another a placebo. "For me the research is much more than taking a tablet. It's taking a tablet with the intention of delivering outcomes, that may or may not help me but help others," Barrett said.
How the research works
The institute is also building a database of people with Parkinson's disease, dementia, and MND. Clinicians collect iris scans and voice recordings and use AI to process and organize large amounts of data, looking for subtle changes that could serve as early warning signs of future health problems.
In addition, researchers take blood samples from volunteers and turn them into stem cells, which are then cultivated into groups of brain cells known as neurones. Existing drugs are tested on multiple batches of these neurones using a combination of robots, traditional lab equipment, and computers running specialized algorithms.
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
These machine-learning algorithms have been trained to recognize drugs that can shift the neurological disease signature back toward a healthy one. Drugs that the AI suggests may be effective can then move into clinical trials involving people like Barrett.
The potential of repurposing drugs
There are about 1,500 drugs that have been developed and approved to treat other conditions. Institute chief executive Professor Siddharthan Chandran said it is possible that even one of those drugs could also work in the brain without anyone knowing it yet.
"The brain is the most complicated organ in the body, so we've got to contend with the paradox of that complexity," Chandran told the BBC. He added that until recently, scientists had to rely on less sophisticated methods of study. "A combination of AI and new technologies mean we can now do things which would have been unbelievable when I was at medical school."
Because the drugs already exist and are approved, repurposing them can be more straightforward than starting from scratch. Developing a completely new drug and bringing it to market can take more than 10 years. Chandran and his team believe their work could mean affordable, effective treatments for neurological conditions arrive much sooner.
Broader AI research and a recent setback
The Edinburgh research is not the first to explore how AI can unearth potential solutions hidden in large medical datasets. Scientists at the Massachusetts Institute of Technology in the United States have used generative AI to identify new antibiotic compounds that might treat superbugs such as gonorrhoea and conditions like Parkinson's disease. In 2024, researchers at Harvard University developed a neural network model called TxGNN to surface existing drugs that could be used for rare conditions.
But the field has also seen setbacks. A recent review of the Alzheimer's drugs lecanemab and donanemab, once hailed as breakthroughs, found that although they slowed disease progression, the effect was not significant enough to make a meaningful difference for patients. The review examined 17 studies involving 20,342 volunteers and looked at drugs that remove amyloid, a misfolded protein present in the disease, from the brain. The conclusion sparked backlash from other scientists.
Despite that, Professor Chandran remains confident that the field is at a "tipping point of change" in neurological research and understanding.