Computer algorithm can accurately predict, will develop if a person has Alzheimer’s disease two years before the actual symptoms. This fact is devoted to a new study at McGill University.
Machine makes a correct prediction in 84% of cases. The researchers are pleased that the present level of development of artificial intelligence technology allows you to choose the right patients for clinical trials and medication in anticipation of the beginning of the disease. And this, in turn, can help delay the progression of the disease and delay its debilitating consequences.
The results of a study published in the journal «Neurobiology of aging» suggests that technology can in time to warn doctors about the inevitable diagnosis of Alzheimer’s disease, and it is much better to help in the fight against disease, than how the patient describes his health.
Alzheimer’s is very difficult to predict. Usually scientists rely on the amyloid (the protein that accumulates and leads to cognitive damage). It can be seen when scanning the brain, but the presence of amyloid does not always lead to dementia (Alzheimer’s disease is the most common form).
Algorithm AI walked three steps on the path from development to testing. Was first written the software, and then the researchers conducted a learning machine and testing the entire system. The program showed the brain scans of 200 volunteers, so she could understand how the brain of amyloid formations.
Then the machine showed the results of the scan, another 270 people who have been observed in mild cognitive impairment, but the required the disease is still not apparent. In 43 of them, eventually developed Alzheimer’s disease. Using the previously obtained knowledge, the algorithm could accurately predict 84% of cases, when a person is manifested dementia.
