Early Detection of Alzheimer’s and Parkinson’s Through Speech Patterns
November 20th, 2015 by NHI
For some time, researchers have sought ways to detect neurological disease at a very early stage.
Early diagnosis, before patients show up in a doctor’s office with symptoms, could delay the onset of the disease or even offer a shot at preventing it altogether.
Recent developments in computer learning and technology give us hope that this can be achieved through analysis of speech patterns. Here’s what they’ve found. . .
Continued Below. . .
Colonoscopy cover-up—are you being swindled by this medical sham?
2015 will be the year one of American medicine’s most profitable cover-ups is blown to pieces.
Over 14 million colonoscopies are performed each year.
But despite its prevelance, there are devastating facts most American doctors are “keeping mum” all at your expense.
The most important being…you have 3 unbelievably easy and effective alternatives right in front of you.
Imagine—NO more painful probes and with these results, you can feel safe knowing you’re helping to protect yourself from colon cancer.
Today, one medical insider is releasing his scathing report—absolutely FREE. It’s yours—but I can’t guarantee how long they will last. Claim yours while you can.
Not to beat around the bush, it appears that the “dumber” our use of language gets, the more likely we are to get a neurological disease. Declining vocabulary. . .fewer and simpler ideas in a sentence. . .long pauses. . .and repeating the same phrases are just a few of the many speech signals that point toward a medical problem.
Researchers from Arizona State University were fascinated by The Nun Study published in 1986 in which the letters and writings of 700 nuns were analyzed.
It compared the number of new ideas in a sentence to its length. It was found that the greater this “idea density” the lower the risk of dementia in later life.
Bush vs Reagan
The researchers wondered if this also applied to speech, so they compared the non-scripted press conferences of President Reagan with those of the first President Bush.
They found that the former, who was diagnosed with Alzheimer’s five years after he left office, had significant changes in his speech.
Over time he spoke a reduced number of unique words and used more non-specific nouns like “thing” when he could not think of the right word. He used more conversational fillers and repeated himself more. None of these speech pattern irregularities were apparent with President Bush.
A Five Minute Test in the Doctor’s Office
Prof. David Shaffer of Binghamton University, New York, is another researcher who looks for clues in speech. He and his team have compiled a hundred indicators. He believes there must be an “identifiable fingerprint” for Alzheimer’s patients.
In his tests, subjects are shown a picture and asked to describe it.
Some indicators are the length of pauses and their frequency, short and clipped sentences, complexity of sentences and repetition of phrases.
They are developing a software program that can be used as a diagnostic tool in any doctor’s office.
His colleague Dr Shawn Berkowitz holds out the potential of creating “a five minute test in their office and says, you know, you have a 70% chance that this could be Alzheimer’s. That’s our dream really.”
Working on the same problem is Professor Sona Patel from Seton Hall University in New Jersey.
Volunteers’ brain activity and vowel sounds are analyzed while hooked up to an electroencephalogram.
She too is looking to provide software a doctor can use with a score that indicates the likelihood of a neurological condition. She is even considering a smartphone app to allow patients to test themselves.
Speech Patterns Help Diagnose Parkinson’s
Changes in speech patterns are already assessed in Parkinson’s patients. A tremor in the voice, and soft, breathy speech are early indicators of the disease.
Max Little, Ph.D., is a University of Oxford applied mathematician and currently project director of the Parkinson’s Voice Initiative at MIT. He’s been developing methods of detection from voice recordings for some years.
He says, “Vocal effects can actually be quite subtle, in some cases, but with any digital microphone, and using precision voice analysis software in combination with the latest in machine learning, which is very advanced by now, we can now quantify exactly where somebody lies on a continuum between health and disease using voice signals alone.”
In 2012 he and his colleagues were able to demonstrate with 98.6% accuracy whether a subject was healthy or had Parkinson’s.
He hopes an ultra-low cost, 30-second self-administered test will be available to the public in the near future.