Can We Use Language to Predict Alzheimer’s Dementia Onset?
Artificial intelligence using linguistic dysfunction can help identify future dementia.
Artificial intelligence (AI) using a short language test is moderately accurate in determining whether seemingly healthy individuals will develop Alzheimer’s disease in the future. Today we turn to new research from IBM Research. First, I want to look at some of the warning signs and symptoms of dementia.
The Alzheimer’s Association offers that 50 million individuals live with Alzheimer’s disease or other forms of dementia worldwide. AD is a degenerative brain disease and is the most common type of dementia. Dementia is a broad term that describes a constellation of symptoms.
Memory loss that disrupts one’s activities of daily life may be a symptom of dementia. Here are some selected warning sign and symptoms offered by the Alzheimer’s Association:
- Memory loss that impairs daily life. Here, we are not talking about occasionally forgetting names or appointments and remembering them later. Alzheimer’s disease’s common symptom is forgetfulness of recently learned information or asking the same questions repeatedly. Some begin to rely on family members for things they formerly handled on their own.
- Challenges in planning or solving problems. Can’t keep track of your monthly bills anymore? Such issues can be a symptom of dementia. Making the occasional mistake doing your household bills is more likely age-related change.
- Difficulties completing familiar tasks. Forgotten how to drive to a usual location? You may need a clinical evaluation.
- Confusion about time or place. Here, we are not talking about the confusion about the day of the week and remembering it later.
- Trouble understanding visual images and spatial relationships. Visual problems can be a symptom associated with dementia. Such problems can lead to difficulties in reading or driving.
- New-onset problems with speaking or writing. For example, someone with early dementia may stop mid-conversation and not be able to continue.
- Misplacing things. Of course, we all do this at times but typically can retrace our steps to find a lost object.
- Personality or mood change.
- Withdrawal from work or social obligations.
- Decreased or poor judgment.
If you have any concerns about symptoms such as these, please see a valued healthcare professional. You may be able to find ways to delay the disease’s progression or discover a clinical trial that offers the promise of helping you or others.
We turn to this observation: Language competence is intimately related to mental functioning. Now we have news that a simple language test (developed with artificial intelligence tools) can accurately predict which cognitively normal individuals will develop Alzheimer’s dementia in the future.
The test has a 70 percent accuracy rate in predicting Alzheimer’s disease before the cognitive decline is apparent. The novel approach appears more accurate than historical practices, such as neuropsychological testing.
Someday, the researchers observe that we may have simple non-invasive probes that detect early dementia and help monitor its progression.
Study details: The cookie model
We start with the premise that language components are essential indicators of age-associated cognitive decline. Even simple tasks such as naming objects can activate brain networks broadly.
The researchers used data from the Framingham [USA] Heart Study. This investigation tracked over 5,000 individuals over several decades. Researchers had subjects complete a neuropsychological battery from the Boston Aphasia Diagnostic Examination.
As a part of the test, participants provide a written description of a so-called cookie-theft picture. The illustration shows three characters in a kitchen, including a woman at an overflowing sink, a boy reaching into a cookie jar, and a girl expecting to get a cookie from the boy.
Researchers extracted linguistic variables from the responses to the test, identifying 87 variables. The study investigators then developed computer models to predict whether a study subject would develop mild cognitive impairment and subsequent Alzheimer’s dementia.
And voila, a first-generation simple tool to identify those at high risk for developing Alzheimer’s dementia. Thank you for joining me today.