Artificial Intelligence diagnoses Alzheimer’s and beats the scientists
In the UK, 225,000 people will develop dementia from Alzheimer’s this year alone, according to estimates by the Alzheimer’s Society - that’s one person every three minutes. Hundreds of thousands of families are affected annually by the heartbreaking deterioration of loved ones to this cruel and under-researched disease.
Alzheimer’s is a degenerative brain disease that’s the leading cause of dementia in elderly
patients. Dementia is a catch-all term to describe various brain conditions that can disrupt cognitive functions, causing confusion and distress for the patient. Proteins build up in the brain to cause plaques and tangles, disrupting the connections between nerve cells leading to memory loss and problems with communication and orientation, and eventually death. Currently, Alzheimer’s often goes undetected until it is too late, as testing relies on methods that are too specialised, expensive and invasive to roll out widely.
There is currently no cure for Alzheimer’s but researchers are racing to enable earlier diagnosis of the disease, allowing treatment steps to be made. Early diagnosis of Alzheimer’s can provide patients with the crucial extra time needed to make lifestyle changes that can slow down the disease’s progression. Drugs that are currently in development are more effective the earlier they are given, but research into the area is desperately underfunded and breakthroughs are rare. In the UK five times fewer researchers choose to work on dementia research than they do cancer research - but a team of Italian researchers are bringing their skills into the fight with the help of computers.
Scientists Nicola Amoroso and Marianna La Rocca at Italy’s University of Bari have been looking into using non-invasive MRI imaging to detect alterations in the brain’s neural connections. The team have developed a machine-learning algorithm using 67 images of healthy brains and Alzheimer’s affected brains, gathered from a database held by the University of Southern California in Los Angeles. The idea of the research was to train the algorithm to ‘spot the difference’ between samples of healthy and diseased brains, by teaching it to learn to analyse small sections of different brain scans. The algorithm compared the neuroconnectivity of each region of the scans and was able to correctly identify alterations around 86% of the time.
In an additional exciting development, the algorithm was able to detect mild cognitive impairment (MCI) in around one third of the samples it analysed. Mild cognitive impairment is a decline in cognitive skills like memory and reasoning that can be seen as a precursor to the patient developing Alzheimer’s some years later. In correctly identifying MCI in patients, the algorithm could identify changes in the brain almost a decade before clinical symptoms of Alzheimer’s appear. It’s not currently known how many older people with MCI go on to develop Alzheimer’s, but further research in this area could help to clarify this.
So what’s next? As no cure or therapeutic treatment currently exists for Alzheimer’s, the benefits of these findings are somewhat limited, but with earlier discovery of warning signs like MCI patients will be empowered to make changes that may slow the onset of the disease. It is also heartening to see a diagnostic method in development that can be made widely available relatively cheaply and in a non-invasive way. The research needn’t stop here, either, La Rocca says: ‘it’s a method that’s very versatile’. Her team are now looking into ways to extend the technique into identifying biomarkers in patients, and helping with earlier diagnosis of other brain diseases like Parkinson’s.
Medical breakthroughs are often a long journey, and require hundreds if not thousands of hours of dedicated work from talented professionals and, in this case, computerised algorithms. If technology is able to take human ingenuity and improve upon it, then that’s certainly a development to be welcomed! The journey is not yet over for Alzheimer’s researchers, but hopefully thanks to Amorosa and La Rocca’s team, we’re a few steps closer to understanding and tackling this life-changing disease.