By listening to how a person speaks, artificial intelligence can diagnose Alzheimer’s.
Alzheimer’s affects millions of people around the world every year. The disease damages brain function and is difficult to detect at an early stage.
By comparing the language of a healthy person with the person suffering from Alzheimer’s, it turns out that there are clear differences. This is where machine learning comes in.
For example, someone with Alzheimer’s disease may need to change certain words in their sentences. The person likes to exchange nouns for pronouns – instead of saying “The boy sat down on the chair” it is easier to say “He sat down on it”.
Small changes like these easily go unnoticed by those around you. Language analysis with AI , however, shows what may be going on.
Research at the Stevens Institute of Technology has shown a link between the onset of Alzheimer’s and the individual’s speech. Through so-called machine learning, the researchers’ software has learned to feel the difference in the number of healthy people and people with Alzheimer’s, respectively.
Training gives skill: the more comparisons the software makes, the better the AI has become at detecting early Alzheimer’s. The researcher who led the project, KP Subbalakshmi, says that their method diagnoses Alzheimer’s with ninety-five percent certainty – “it is absolutely state of the art”.
If all goes well, this solution offers a cost-effective and for the patient convenient way to test for onset Alzheimer’s.
Will Alexa say in the future that something seems wrong with the language we use at home in the kitchen? Or will it be our smart watch that gives us a warning?
Let the future come faster!
