A new artificial intelligence machine can identify changes in the brains of people most likely to get Alzheimer’s disease ten-years before doctors can diagnose from symptoms alone.

67 non-invasive MRI scans were analysed – of which 38 had Alzheimer’s and 29 did not. The researchers at the University of Bari in Italy used their machine-learning algorithm that could identify the structural changes in the brain caused by Alzheimer’s.  

The original concept was to teach the algorithm to classify and identify the difference between those with Alzheimer’s and those without. The researchers found the best results came after the brain regions were analysed were the size of 2250 to 3200 cubic millimetres. Interestingly, these measurements were comparable to the anatomical structures associated with Alzheimer’s.

In more tests, researchers used a different control group of 148. 52 of the subjects were healthy, whereas 48 had Alzheimer’s disease and mild cognitive impairment but were known to have developed Alzheimer’s two and a half to nine years later.

Not only was the algorithm able to differentiate between the healthy brain and the one with Alzheimer’s, but it could also distinguish between healthy brains and those with mild cognitive impairment with an accuracy of 84%.

This shows that the algorithm can identify changes in the brain that lead to Alzheimer’s, ten-years before any symptoms appear.

Speaking on the discovery, Marianna La Rocca, researcher at the University of Bari, said:

“Nowadays, cerebrospinal fluid analyses and brain imaging using radioactive tracers can tell us to what extent the brain is covered with plaques and tangles, and are able to predict relatively accurately who is at high risk of developing Alzheimer’s 10 years later”

To find the latest developments surrounding machine learning in medical imaging, register for your free ticket to the Medical Imaging Convention here.

Additionally, if you are interested in exhibiting at next year’s convention and meeting thousands of decision makers actively looking to purchase products and services, please contact Nick Woore, Event Director, now on: 

Tel: 0117 929 6097