Can you predict who will get Alzheimer’s just by looking at written patterns years before symptoms appear?
According to a new study by IBM researchers, the answer is yes.
And, they and others say that Alzheimer’s is only just beginning. People with many types of neuropathy have special language patterns that, investigators suspect, could be an early warning sign of their disease.
For the Alzheimer’s disease study, researchers looked at a group of 80 men and women aged 80 – half with Alzheimer’s and others not. But seven and a half years before, it was all cognitively normal.
The men and women participated in the Tim Framingham Study, a lengthy federal research effort that requires regular physical and cognitive tests. As part of it, they took a written test before any of them had Alzheimer’s disease, asking subjects to describe a drawing of a boy standing on an unstable chair and reaching for a bottle. biscuits on high shelf while a woman turns her back on him. , forgotten with an overflow sink.
The researchers tested the subjects’ word usage with an artificial intelligence program looking for slight differences in language. It identifies a group of subjects that are more repetitive in their word usage at a time ago when they were all cognitively normal. These subjects also make mistakes, such as incorrect spelling or capitalization and they use telegram language, meaning that the language has a simple grammatical structure and lacks subject and words like ” the “,” is “and” are “.
Members of that group turned out to be the ones who developed Alzheimer’s disease.
The AI program predicted, with a 75% accuracy, who would get Alzheimer’s, according to results recently published in The Lancet’s EClinicalMedicine.
“We have no prior assumption that word use is,” said Ajay Royyuru, vice president of health and life science research at the IBM Thomas J. Watson Research Center in Yorktown Heights, NY. will show anything.
Alzheimer’s researchers have been intrigued, saying that when there are ways to slow or end the disease – a goal so far elusive – it’s important to have simple tests possible. Early warning is that without intervention, a person will develop encephalopathy.
“What’s going on here is very clever,” said Dr. Jason Karlawish, an Alzheimer’s researcher at the University of Pennsylvania. “With a large number of speeches or articles, can you give the signal?”
For years, researchers have analyzed voice and voice changes in people with symptoms of neuropathy – Alzheimer’s, ALS, Parkinson’s, frontal dementia, bipolar and psychiatric disease. schizophrenia, among others.
However, Dr. Michael Weiner, who studies Alzheimer’s disease at the University of California, San Francisco, said the IBM report created a new platform.
“This is the first report I have seen that has taken perfectly normal people and correctly predicted who will have problems for years to come,” he said.
The hope is to expand the work of Alzheimer’s disease to find subtle changes in language use by people who don’t have obvious symptoms but will continue to develop other neurological diseases.
According to Dr. Murray Grossman, professor of neurology at the University of Pennsylvania and director of the university’s front dementia center, each neuropathy produces unique changes in speech, possibly happened long before the time of diagnosis.
He studied speech in patients with frontal behavioral dementia, a disorder caused by the gradual loss of nerves in the frontal lobe of the brain. These patients exhibit lethargy and impaired judgment, self-control, and empathy that have been shown to be difficult to objectively quantify.
Speech is different, Dr. Grossman says, because changes can be measured.
In the early stages of the disease, there are changes in the patient’s speech rate, with seemingly random distributional pauses. Word usage has also changed – the patient uses less abstract words.
These changes are directly related to changes in the anterior parts of the brain, says Grossman. And they seem to be ubiquitous, not unique to the English language.
Dr. Adam Boxer, director of the clinical research unit of neuroscience at the University of California, San Francisco, is also studying frontal dementia. His tool is a smartphone app. His subjects are healthy people who have inherited a genetic predisposition to develop the disease. His method was to show subjects a picture and ask them to write down a description of what they saw.
“We want to measure changes very early, five to 10 years before they show symptoms,” he said.
“The great thing about smartphones,” adds Dr. Boxer, “is that you can do anything.” Researchers can ask people to talk for a minute about something that happened that day, he says, or repeat sounds like tatatatata.
Dr. Boxer said he and others are focusing on speech because they want the tests to be non-invasive and inexpensive.
Dr. Cheryl Corcoran, a psychiatrist at the Icahn School of Medicine in Mount Sinai, New York, hopes to use the changes in voice to predict which adolescents and young adults are at high risk of schizophrenia. Paralysis can continue to develop the disease.
Drugs to treat schizophrenia can help people who are about to get sick, but the challenge is to determine who the patient will be. One in four people experience symptoms that go away from time to time, and about one third never develop schizophrenia even though their occasional symptoms persist.
Guillermo Cecchi, an IBM researcher who also recently participated in the Alzheimer’s study, studied the speech in 34 patients of Dr. Corcoran, looking for “the flight of ideas”, meaning where the patient deviates from speaking and turns ideas in different ways of direction. He also looked for “language poverty”, meaning the use of simple syntactic structures and short sentences.
In addition, Dr. Cecchi and his colleagues studied another small group of 96 patients in Los Angeles – 59 of whom sometimes have paranoia. The rest are healthy people and people with schizophrenia. He asks these subjects to retell a story they have just heard, and he looks for narrative patterns like that.
In both groups, the artificial intelligence program could predict, with an accuracy of 85%, who would develop schizophrenia three years later.
“There have been a lot of small studies finding similar signals,” says Dr. Corcoran. At this point, she said, “we’re not at the level yet able to tell people if they’re at risk.”
Dr. Cecchi was encouraged, although he realized that research was still primitive.
“For us, the first priority is to do science with precision and scale,” he said. “We should have more models. There are over 60 million psychiatric interviews in the US each year, but none of them use the tools we have.