Summary: A new machine-learning algorithm can accurately detect cognitive impairment by analyzing audio recordings.

Source: Boston University

It takes a lot of time and money to diagnose Alzheimer’s disease. After conducting long-term neuropsychological tests in the body, clinicians should copy, review, and analyze each response in detail.

But researchers at Boston University have developed a new tool that automates the process and ultimately allows it to run online. Machine learning: A computational model can detect cognitive impairment from neuropsychological tests – no physical appointment.

Their findings were published in Journal of Alzheimer’s and Alzheimer’s disease Alzheimer’s Association.

“This approach brings us one step closer to the initial intervention,” said Ioanis Pascalidis, author of the paper and Honorary Professor of Engineering at BU College.

He said early detection of Alzheimer’s disease could lead to large clinical trials targeting individuals in the early stages of the disease and may lead to clinical interventions that reduce cognitive decline. Number of people to be examined quickly.

The research team trained more than 1,000 individuals in Framingham Heart Study, cardiovascular disease and other physiological conditions using a long-term BU-led project using audio-recorded neuropsychological interviews.

Using automated online speech recognition tools – “Hey, Google!” – and computers, using a machine learning technique called a natural language processing technique that allows their programs to copy the interviews and then duplicate the numbers.

The final model uses demographic data, text encoding, and real neurologists and neuropsychologists to assess an individual’s chances and severity of cognitive impairment.

Pascalidis says the model not only accurately identifies healthy individuals and people with dementia, but also distinguishes between those with mild cognitive impairment and those with dementia. Thus, the quality of the recording and the way people spoke — their speech was windy or incessant — was less important than what they said.

“We are surprised that the flow of speech or other audio features is not so important; we believe it is important to translate interviews and evaluate cognitive impairment based on AI analysis,” said Pascalidis, also the new director of BU’s Rafik B. Hariri Institute of Computing and Computational Science and Engineering.

Although the team still wants to verify the results from other sources, the findings suggest that their devices can help clinicians diagnose cognitive impairment using audio recordings, including those coming from virtual or tele health appointments.

Diagnosis before the onset of symptoms

The model suggests that they may be more useful than other parts of the neuropsychological test for determining an individual’s cognitive impairment. The researchers’ model divides test copies into different sections based on clinical trials.

For example, the Boston Name Test – when clinics ask individuals to name a picture – they found it to be very informative for the diagnosis of dementia.

“This will allow clinics to allocate resources in a way that allows them to conduct further tests even before the onset of symptoms,” says Pascalisis.

Early diagnosis of Alzheimer’s disease is important not only for patients and their caregivers to develop effective treatment and support plans, but also for medical researchers working to reduce and prevent the development of Alzheimer’s disease.

“Our model clinics can help patients assess their chances of cognitive decline,” says Pascalidis.

watch out

Do you want to join the research effort?

The research team is looking for volunteers Take an online survey and submit an anonymous cognitive test-The results will be used to provide personal awareness assessments and also help the team refine their AI model.

So the news of AI and Alzheimer’s disease research

Author Molly Gluk
Source: Boston University
Contact Molly Gluck – Boston University
Image The image is in the public domain.

Preliminary study Closed access.
”Simple Cognitive Disorder and Autism Scale Automatic Separation of Audio Copies The process of setting up a natural languageBy John Pascalidis et al. Alzheimer’s and dementia

Draft

Simple Cognitive Disorder and Autism Scale Automatic Separation of Audio Copies The process of setting up a natural language

Introduction

Automated computational evaluation of neuropsychological tests allows for a comprehensive, cost-effective diagnosis of dementia.

Methods

A new natural language processing approach has been developed and validated to identify different levels of forgetfulness based on the automatic recording of digital audio recordings of subjects in the Framingham Heart Study (Framingham Heart Study).n = 1084). Copies from the experiment were recorded into quantitative data, and several models were tested and tested using this data and the demographic characteristics of the participants.

Results

The average area below the AUC on the test data reached 92.6%, 88.0% and 74.4% Normal cognition from Dementia, Normal or Mind Cognitive Impairment (MCI) from Dementia, and Normal from MCI, respectively.

Conversation

The proposed approach provides a complete automatic diagnosis of MCI and Alzheimer’s disease based on the registered neuropsychological test, which provides the opportunity to develop a remote control device that is easily adapted to any language.