By simply listening to your speech, artificial intelligence can determine if you will develop Alzheimer's disease
According to popular belief, artificial intelligence has improved our quality of life in a variety of fields, including education, the workplace, and medicine.
There is no known cure for Alzheimer's disease, so researchers have been looking for the best early detection methods for decades in an effort to prevent or postpone the onset of the disease.
It is crucial to identify individuals with mild cognitive impairment (MCI) early on before it develops into Alzheimer's disease. Both prompt treatment and successful enrollment in all kinds of clinical trials depend on this.
Therefore, Boston University researchers used advances in artificial intelligence to create a technique that, by analyzing patients' speech, can predict when mild cognitive impairment will turn into Alzheimer's disease.
The researchers used audio recordings they made while interviewing 166 people to arrive at this conclusion.
Ninety of them had mild cognitive impairment, which is a sign of Alzheimer's disease, and 76 of them continued to have mild cognitive impairment for six years after the follow-up began.
They processed the spoken language data and different demographic variables, like age, using machine learning algorithms based on these recordings. They then produced linguistic and acoustic features that were fed into predictive models.
Therefore, Boston University researchers used advances in artificial intelligence to create a technique that, by analyzing patients' speech, can predict when mild cognitive impairment will turn into Alzheimer's disease.
The researchers used audio recordings they made while interviewing 166 people to arrive at this conclusion.
Ninety of them had mild cognitive impairment, which is a sign of Alzheimer's disease, and 76 of them continued to have mild cognitive impairment for six years after the follow-up began.
They processed the spoken language data and different demographic variables, like age, using machine learning algorithms based on these recordings. They then produced linguistic and acoustic features that were fed into predictive models.
When predicting which people with mild cognitive impairment would develop Alzheimer's disease within six years, the best model had 78.5% accuracy and 81.1% sensitivity.
This indicates that the AI model could predict cognitive decline by using voice recognition to identify digital biomarkers.
The benefit of this approach is that it uses a more accessible and reasonably priced tool—like a mobile app—instead of a costly one.
In addition, this may facilitate the ongoing assessment of Alzheimer's disease risk.
This study was conducted last summer, so hopefully some progress will be made in recent months while we wait for new results.