Three algorithms that can be used to detect Alzheimer’s disease in patients while having telephone conversations have been developed by researchers working at the Department of Public Health, McCann Healthcare Worldwide Japan Inc.
Previous research has shown that some of the first signs of Alzheimer’s include speaking slower than normal and pausing more frequently during conversations.
More effective than human diagnostics
In this new study, researchers have replaced humans listening in and analyzing phone conversations with a computer running a machine learning algorithm.
Three different machine learning algorithms were designed to study speech patterns . They were all taught to identify the signs of Alzheimer’s disease through voice recordings from an ongoing dementia program in Japan.
Then other voice recordings were used to test the algorithms, and the researchers found that, on average, they were as good or slightly better than the Telephone Interview Test for Cognitive Status (TICS-J) and did not return any false positives.
The researchers suggest that their algorithms could be used to provide a cheaper and more accessible way of early testing for Alzheimer’s disease.