A new method for analysing brain scans using artificial intelligence (AI) could make it possible to detect signs of Alzheimer’s years before the true onset of the disease. The technique’s accuracy has been rated by its creators as 80%, but tests still need to be carried out on more data.
A method for analysing MRI (magnetic resonance imaging) scans using artificial intelligence (AI) technology could be used to detect the precursory signs of Alzheimer’s disease. Thanks to this algorithm-based technique, it might be possible to identify these signs ten years before the first symptoms of the disease appear.
This is what has been suggested by researchers from the University of Bari (Italy) and from a brain rehabilitation centre in London (United Kingdom), in a study published on 7th September. Scientists studied MRI scans from 67 people, 38 of whom had Alzheimer’s disease, using an algorithm that analyses the brain’s neuronal connectivity.
84 to 86% accurate
Once the results had been studied, the specialists then tested their new algorithm on 148 new MRI scans. These digital images were taken from different types of patients, some of whom had Alzheimer’s disease and others who displayed mild cognitive impairment that might lead to the development of the disease in the future.
The method developed by the authors of the study was found to be capable of detecting the disease with 86% accuracy and detecting mild cognitive impairment with an accuracy of 84%.
Focusing on variations within the human brain.
The AI technology is able to detect structural changes within the test subjects’ brains well before the disease itself becomes apparent. The algorithm focuses on the way in which the human brain restructures its neuronal chains and the connections between the neurones.
As reported in Numerama, in July 2016 a team of Dutch researchers had already tested a method for the early prevention of Alzheimer’s disease by analysing medical imagery using an artificial intelligence system. This technique proved to be 82 to 90% accurate. The system focused on machine learning to recognise early arterial spin patterns in MRI scans.
Awaiting further tests
Though the results are promising and give great hope to patients experiencing problems with their memory, behaviour or thoughts, the research carried out by the British and Italian specialists needs further investigation. The researchers used just one database, the ADNI (Alzheimer’s Disease Neuroimaging Initiative), so further tests are set to be carried out on other scans.
Contact Allianz Partners
Chine, Aug 7, 2017
The Chinese government last week revealed its strategy to make up the ground that is has lost in the artificial intelligence race, and eventually step up to become world leader in the f [...]
Oct 2, 2017
Far more than any other sector of activity, insurance companies are ploughing significant investments into integrating artificial intelligence into their systems. The various companies [...]