A study by researchers at the Research Center for Neuroscience at St. George’s University in London has identified the possibilities of the methods used to record the early symptoms of Alzheimer’s disease through analysis of speech and artificial intelligence, Medical Express reported.
The results of the study, published in the journal Frontiers in Computer Science, show that although artificial intelligence is used to analyze speech, it is very important what task the subject will be given.
So far, research has shown that Alzheimer’s disease affects speech very early. Therefore, early diagnosis of this problem is important.
The researchers’ analysis showed that setting patients the task of telling a story like Cinderella instead of a movie scene, for example, required them to integrate multiple characters and events into a series they remembered, and this helped early diagnosis best.
In a study of 50 patients, the results of the AI analysis showed that the machine detected symptoms with 78 percent accuracy when retelling the Cinderella story, compared with 76 percent accuracy when retelling an episode of a TV series.
“Our results show that by controlling the tasks used to diagnose Alzheimer’s, we are able to detect the disease more accurately through speech analysis,” said study leader Natasha Clark, a PhD student at St. George’s.
–
