A simple language test developed with artificial intelligence (AI) can predict, with a high degree of accuracy, which people with normal perception are likely to have the disease, a new study finds. Alzheimer’s (AD).

The results show that this test has a 70% accuracy rate in predicting AD onset years before cognitive decline begins and is more accurate than traditional prediction methods, such as psychometric testing. nerve.

Study author Guillermo Cecchi, PhD, Center for Computational Biology, IBM Research, Yorktown Heights, New York told Medscape Medical News, “The bottom line is that we’re finding small highlights that when you put them together will provide a significant amount of information.”

The researchers hope the new results will eventually lead to the use of simple, low-cost voice probes to detect dementia early and monitor its progression.

Cecci said: “The value of [loại thử nghiệm] This is it can be done quickly; it’s non-intrusive and can be done at any time.

The study was published online October 22 in eClinicalMedicine.

Cookie thief

An important priority in AD research is the development of early interventions to reduce risk, delay onset, and/or slow disease progression. This requires identifying patients who are likely to benefit from such interventions, and this is where the language comes in.

Research shows that different aspects of language are an important component of age-related cognitive decline. Even ordinary language abilities, such as naming objects, join vast brain networks.

To conduct the study, the investigators used data from the Framingham Heart Study (FHS), a longitudinal investigation of more than 5000 individuals over several decades. As part of the FHS, participants completed an energy neuropsychological test that included the cookie-stealing image task (CTT) from the Boston Aphasia Diagnostic Test.

In this experiment, study participants were asked to describe in writing a picture of stealing cookies. The painting depicts three figures in a kitchen – a woman at a sink overflowing with water; a boy reaches for a jar of cookies in the cupboard; and a little girl expecting a cookie from the boy. The investigators extracted linguistic variables from the responses to CTT. In total, 87 linguistic variables were calculated.

Using these variables, the researchers developed computer models to predict whether a participant had mild cognitive impairment (MCI) leading to AD.

A review panel with at least one neurologist and one neuropsychologist considered possible cases of cognitive decline and dementia. They diagnosed AD based on criteria from the National Institute of Neurological and Communication Disorders and the Association of Stroke-Alzheimer-Related Disorders.

Patient cases included those who developed MCI due to AD at age 85 or earlier. Age, sex, and education matched controls included those who remained dementia-free until at least 85 years old.

From more than 3000 responses from CTT, the researchers created models to examine 80 participants – 40 cases of disease and 40 controls.

Analyzes suggest that the onset of future AD is associated with repetition, typos, and telegraph speech, defined as speech lacking a fluent and continuous grammatical structure.

Another important variable, Cecchi said, is the “granularity” of the “reference” or lack of such a reference, for example, referring to the older woman in the photo as “mother” or “wife” ” instead of using the more generic term. , “women”.

Using linguistic variables yielded significant predictive power, with an area under the curve of 0.74 and an accuracy of 0.70. The median time to diagnosis of AD was 7.59 years.

The researchers note that this model is based on data collected when the study participants were cognitively normal. The study found that predictions were harder for participants with a college degree than for less educated participants.

Language competence is a behavioral marker of education and occupation, both of which have been suggested to increase “cognitive reserves,” the investigators note.

Research shows that it is much easier to predict conversion to AD in women than in men. The authors noted that the frequency of AD was significantly higher in women than in men, and that the rate of progression after onset of cognitive decline was faster than in women.

Better performance with predictive models that use linguistic variables than prediction models that incorporate more of the traditional variables associated with AD risk, such as neuropsychological test scores, information demographic information and APOE status. For disease predictions based on a combination of these traditional variables, the accuracy was 59 percent, says Cecchi.

The authors note that neuropsychiatric tests and other biomarkers, including assessment of cerebrospinal fluid and brain imaging, have been used to predict progression of MCI to AD. In addition, there have been very promising results using nerve light chains to treat disease progression in asymptomatic early-stage patients of similar AD.

“However, these are still technology or supply requirements and require expert involvement,” they note.

Cecchi and his colleagues hope this new study will lead to the development of a simple, accessible tool for accurately assessing risk for AD.

Cecchi envisions a simple test that could be used routinely by clinicians and family members to monitor a patient’s disease progression, eliminating the need for lengthy hospital appointments. doctor’s office.

Screened patients identified as at-risk may be able to make lifestyle changes to help delay cognitive decline. These can include following a healthy diet, increasing physical activity and social and cognitive inclusion, Cecchi says.

Additionally, doctors can use the tool to identify patients who might benefit from enrolling in clinical trials of potential preventive therapies, he said.

Interesting but new research

Commenting on the study for Medscape Medical News, PhD Heather Snyder, vice president of medical and scientific relations, Alzheimer’s Association, described the study as “interesting” and “intriguing” but cautioned that it remains is a “new project”.

Research to identify speech or language biomarkers of cognitive change is “certainly” one of the “emerging new areas,” says Snyder. The Alzheimer’s Association has funded research in this area.

“We’re seeing more and more studies looking at and trying to understand vocal and vocal patterns and that can be an indicator of what might be going on in the brain,” she said.

However, Snyder points out that the current study used a fairly small and “quite defined” group of individuals.

She also notes that language is tied to hearing and other senses. “So this is not so simple.”

She stressed that there is certainly still some work to be done before this type of test becomes useful as a diagnostic tool.

Source: Medscape

Link: https://www.medscape.com/viewarticle/939618#vp_1

Self-translated article please do not reup.

Author: Roxie Duong