Treating disease from a distance. Once science-fiction, it’s becoming a reality thanks to technology. Telehealth allows patients and providers to connect with each other by phone or medical PC to discuss a patient’s problems. Remote patient monitoring (RPM) allows providers to monitor a patient’s health from a distance through devices such as wearables. 

There’s a third option, one that has the medical community excited. Called vocal biomarkers, this technology allows providers to evaluate and even diagnose patients’ issues like COVID-19 just by the sound of their voice. 

Sounds, again, like science-fiction? Like telehealth and RPM, it may be more common in the years to come. We’ll be covering how vocal biomarkers work and their effects on three major diseases: Parkinson’s, coronary artery disease (CAD), and COVID-19.  

Vocal Biomarkers: The Science Behind the Sound

To understand vocal biomarkers, we need to first discuss biomarkers.

A “biomarker” or “biological marker” is defined by the National Institute of Health (NIH) as a medical condition observed on a patient from the outside. Examples range from blood pressure and heart rate to complex genetic tests.

Biomarkers are usually determined by a provider or medical staff and are always measurable in some way. 

Vocal biomarkers are medical conditions that can be deduced from one’s voice. A person’s voice is created by three major components: the lungs, the larynx, and the articulators ﹘ the tongue, the palate, and the mouth muscles. These components plus the brain work together for maximum precision. 

Pitch, tone, rhythm, speaking rate, as well as breathing and coughing change when people become ill. The characteristic cough with the common cold is a simple example. Based on this premise, the technology works in the following manner:

  • A recording of a person’s voice is made. They may be asked to read text to a medical tablet, describe a personal experience, voice a vowel for as long as they can (example: “aaaaaah”), or force a cough. Note there’s no special recorder used in vocal biomarking. Recordings made on a standard smartphone work just fine. 
  • That recording is translated into an image called a spectrogram. 
  • Artificial intelligence (AI) and machine learning (ML) algorithms are used to look for correlations between that spectrogram and a variety of illnesses and symptoms. These small changes are detected using computer vision techniques. The result is stored as a vocal biomarker.
  • The vocal biomarker is then tested with a different group of patients to see how well it performs. The goal is more matches between the biomarker and the illness. Like most AI and ML systems, accuracy increases as more data is entered to strengthen the algorithm.

Vocal Biomarker Case Studies 

Healthcare has been quick to jump onto the technology and with good reason. 

Providers can gather more information on a patient’s condition during telehealth visits. Imagine being able to determine their cough is actually a sign of an upcoming heart attack. Or they’re in the early stages of a COVID-19 infection. Such insight would not only avoid a costly office visit, but protection from highly infectious diseases.

Technologies developed from vocal biomarkers can also allow a limited form of RPM. The provider could call a patient suffering from Parkinson’s and determine if their meds are working based on their speech. The advantage over current systems is there is no need for cumbersome monitoring in-home systems or even wearables. All that’s needed is a phone connection. 

Dr. Jaskanwal Deep Singh Sara notes about vocal biomarkers: “We’re not suggesting that voice analysis technology would replace doctors or replace existing methods of health care delivery, but we think there’s a huge opportunity for voice technology to act as an adjunct to existing strategies. Providing a voice sample is very intuitive and even enjoyable for patients, and it could become a scalable means for us to enhance patient management.”

Research is still on-going to bring vocal biomarkers to the healthcare industry. The following major case studies look especially promising. 

Parkinson’s Diseases – Signs from Speech

Parkinson’s disease is a brain disorder. Nerve cells in the basal ganglia, an area of the brain that controls movement, become impaired and even die. This causes unintended or uncontrollable movements in patients, such as shaking, stiffness, and causes difficulty with balance and coordination. 

It also affects speech. Patients affected by Parkinson’s disease experience loss of voice volume and begin to speak with haste. Up to 78 percent of such patients showed speech changes in the early stages of the disease according to the NIH. The hope is that earlier detection can get patients quicker into proper treatment programs. This is especially useful with Parkinson’s which currently has no known cure.

Coronary Artery Disease – Hearing Heart Attacks 

Coronary artery disease (CAD) is when not enough blood reaches the heart through the arteries. This is usually caused by the build-up of plaque in the arteries themselves. The results can range from mild chest pain to a lethal heart attack. 

The American College of Cardiology earlier this year released the results of a study on the use of vocal biomarkers in predicting CAD. It showed participants who scored high in vocal biomarkers were 2.6 times more likely to develop CAD than those with lower scores. They were also three times more likely to have plaque build-up as well. The hope is that early detection can help patients receive proper treatment like medication and weight control advice. 

COVID-19 – Surprising Link to Alzheimer’s

It’s unsurprising the healthcare industry looked for any correlation between voice biomarkers and the coronavirus

The ability to determine if a patient is positive with the disease simply by their voice would have many advantages. Patients could be safely tested via telehealth. There would be no need for swabs and labs, which would also drastically cut costs for providers. 

A correlation was detected. Interestingly, it used the same vocal biomarkers for detecting Alzheimer disease. Even more exciting is that COVID-19 positive patients who were asymptomatic, or showed no outward signs of illness, could be detected via the technique. Healthcare companies like Vocalis Health are looking to bring such screening tools to market. 

As Dr. Shady Hassan, co-founder of Vocalis Health, points out about studies and tools: “PCR testing is being used to screen for COVID-19, which is an extremely expensive, resource-intensive and time-consuming approach. Instead of misusing PCR testing for screening, utilizing a highly scalable screening tool like VocalisCheck can fill a significant gap in the current approach to COVID screening, with the ability to effectively funnel those with high risk of infection to the appropriate diagnostic test.”

Closing Comment 

The ability to detect major illnesses like COVID-19 from the sound of one’s voice seems like something straight from science-fiction. Yet the science behind it is sound, and results from research like detecting Parkinson’s and COVID-19 through it are looking promising. 

If your medical company is looking into the use of vocal biomarkers in the field, contact a representative from Cybernet. Also follow Cybernet on Facebook, Twitter, and Linkedin to stay up to date on this and other relevant topics.