Right on cue, the press is responding to President Biden’s manifest mental lapses by rolling out the so-called experts. “Neurologists say blanking on the names of acquaintances or having difficulty remembering dates from the past, especially when under stress, can simply be part of normal aging,” says a Feb. 10 NBC News report.
Such pieces typically note that only a medical doctor can diagnose cognitive impairment and dementia. That’s true, but it’s also true that dementia and mild cognitive impairment are hugely underdiagnosed. One study found a nondiagnosis rate of more than 90% for mild cognitive impairment, suggesting 7.4 million Americans unknowingly live with that condition.
That’s partially because the cognitive tests doctors use to diagnose these conditions miss many cases. Many older people also don’t get tested because they consider their failing memories to be a normal part of aging or stress. But diagnostic tests now in development could determine whether someone is mentally impaired simply based on how he speaks or walks, or even with a self-administered blood test.
Studies have found that gaits can start to slow more than a decade before patients are diagnosed with dementia. Those with different forms of dementia also exhibit different gait abnormalities. Gait rhythm has been correlated with mental processing speeds. Even subtle change in the brain can affect the way people walk.
One study found that those with a frontal gait disorder—short, shuffling steps—are more than four times as likely to develop vascular dementia. By using machine learning, scientists have developed a wearable technology that can identify Alzheimer’s and Lewy Body disease merely based on the way someone walks.
Artificial intelligence is also enabling scientists to analyze voice recordings to detect patterns in patients with early cognitive impairment. These models have found that people with dementia, even in its early stages, speak with different pitches, volume, cadences and intonations than elderly people who are cognitively normal.
The human ear might not pick up these subtle differences, but artificial intelligence can. A team of researchers from IBM and Japan’s University of Tsukuba has even developed a self-administered prototype app that can accurately diagnose someone with early-stage Alzheimer’s and mild cognitive impairment merely based on voice recordings. Cognitively impaired people not only speak more slowly and with more pauses, but their voices have more irregular vibrations.
Researchers at England’s University of Sheffield developed a cognitive voice diagnostic tool that they found to be as accurate as the “pen and paper” cognitive tests administered at doctors’ offices. Such voice evaluations—which are also quicker than the alternatives—may be used to screen patients for early signs of dementia.
Numerous labs have developed blood tests that use artificial intelligence to identify blood biomarkers linked to dementia. Alzheimer’s and other forms of dementia result from changes in the brain that can generate specific proteins in the blood. These proteins, such as glial fibrillary acidic protein, often show up in the blood even before brain scans show changes.
A study last year presented at the Alzheimer’s Association International Conference found that a blood test measuring specific biomarkers correctly identified Alzheimer’s in approximately 85% of cases later diagnosed by specialists. By comparison, primary-care doctors using cognitive tests and brain scans accurately diagnosed only 55% of cases. Scientists also found that such biomarkers could be isolated with merely a finger-prick of blood, creating the possibility that such tests could be administered at home.
Artificial intelligence is even enabling blood tests that can predict whether a patient will develop dementia a decade later. A study that used blood samples from more than 50,000 people in the U.K. found four proteins were associated with the subsequent development of different types of dementia over 14 years.
The study, published last week in the journal Nature Aging, determined that levels of two blood proteins began to change at least 10 years before a dementia diagnosis. Researchers created an algorithmic model that combined the blood-test results with demographic variables such as age, sex, education and family history. Their model could predict who would develop dementia with 89% accuracy.
How many seniors would want to know if they’re destined to develop dementia? Wouldn’t most prefer blissful ignorance? Maybe. But as early-cancer diagnosis significantly improves prognoses, the same could be true for Alzheimer’s, thanks to new treatments.
Eli Lilly’s Alzheimer’s drug, Donanemab, is expected to receive Food and Drug Administration approval soon. The drug works by removing amyloid plaque in the brain that can trigger a cascade of neurological degeneration. Lilly’s trials showed that patients treated in the disease’s early stages fared significantly better. This holds the promise that the treatment could be used to prevent Alzheimer’s in patients who are reported to have high odds of developing the disease.
Alas, regulators could slow these life-changing advances. The FDA is trying to make it harder to launch tests that use artificial intelligence to diagnose diseases by deeming them medical devices. The Centers for Medicare and Medicaid Services is likewise restricting access to novel Alzheimer’s drugs.
Mr. Biden may not want to come to terms with his cognitive decline. But why does his administration want to deny Americans access to tools that could help their own?