A team of researchers have created an artificial intelligence that can identify the early signs of someone who is likely to get Alzheimer’s disease a full 10-years before their first symptoms will show.
While diagnosing Alzheimer’s has become easier, the most effective way to diagnose it early is through an invasive, expensive procedure that simply isn’t available to everyone.
What makes this study so important then is that it has shown how computers are able to effectively offer a comparable degree of accuracy but without the costs both in terms of time and money.
The team from the University of Bari in Italy were able to do this by harnessing an artificial intelligence that then uses machine learning to analyse images and then accurately recognise the differences between them.
In this case those images were MRI scans. Initially the team fed the algorithm 67 scans, 38 of which were from people who had the disease and the remaining 29 were healthy.
This allowed the algorithm to establish a baseline from which to work. They then fed it a further 148 scans, but this time they broke up the group into three parts.
52 were healthy, 48 had Alzheimer’s disease and a further 48 had Mild Cognitive Impairment (MCI) who had then gone on to develop Alzheimer’s in a few years time.
What the team found was that the algorithm could identify the differences between those with and without the disease with an accuracy of around 86%.
More importantly however it was able to tell the difference between a brain that was healthy and a brain that showed MCI with an accuracy of 84%.
Machine learning has huge potential in the diagnoses of diseases like Alzheimer’s. Earlier this year another team from Case Western Reserve University also used machine learning to identify the early signs of Alzheimer’s within MRI scans.
As with the study in Italy, this earlier project also trained their algorithm to tell the difference between healthy brains and brains that showed the signs of mild cognitive impairment.
The next step for both projects is to increase their algorithm’s baseline knowledge. Put simply, the more scans it can see and analyse the more accurate it can become.
Alzheimer’s disease is the most common type of dementia, a progressive and irreversible neurological disease which affects multiple brain functions and affects an estimated 850,000 people in the UK.
The exact cause is unknown although a number of things are thought to increase likelihood including old age, a family history of the condition, previous severe head injuries and lifestyle factors associated with cardiovascular disease.