Vocal biomarkers have amazing potential in reforming diagnostics through their accuracy, speed and cost-effectiveness in the areas of mental illnesses as well as physical diseases such as Parkinson’s or coronary artery disease.
Your voice betrays you, princess!
The line has a long tradition in literature and cinematography. Its earliest presence might be in the Tales of One Thousand and One Nights by storyteller Scheherezade. In one of her stories her voice reveals a princess dressed in male cloths to a dervish; and he uses this exact line to tell her, he knows he is in fact female.
It has been commonplace since ancient times to spot liars based on their voices. Secret services such as the FBI currently also uses speech patterns in determining the truth value of statements. In an interview with CNBC, former FBI negotiator, Chris Voss said, “only 7 per cent of a message is based on the words, while 38 per cent comes from the tone of voice and 55 per cent from the speaker’s body language and face”. In the television series Lie to Me! the main character played by Tim Roth solved countless crimes based on body language as well as through recognizing lies from the intonation and cadence of the bad guys.
How to define vocal biomarkers?
According to the latest scientific studies, it is definitely not negative, if your voice betrays you. On the contrary! The characteristics of your voice – or as medicine labels them, vocal biomarkers – reveals a lot about your health; and help in detecting serious diseases and health risks.
The term “biomarker”, the shortened version of “biological marker” refers to medical signs, which indicate the medical state observed from outside the patient. So while patients sense symptoms, medical professionals measure biomarkers. Currently, they take into account all kinds of objective, quantifiable biomarkers ranging from biochemical, radiology markers to various health parameters. And as you could have guessed already, vocal biomarkers are medical signs deducted from the features of your voice.
Amazing potential in reforming diagnostics
Vocal biomarkers have an amazing potential in reforming diagnostics through their accuracy, speed and cost-effectiveness.
They are able to detect some diseases earlier than an average check-up process; and an earlier diagnosis could essentially be the difference between life and death in relation to certain illnesses. It is an amazing area of medicine, and the field is buzzing. More and more start-ups are eager to join in: Beyond Verbal, Sonde Health or the Berlin-based Audio Profiling. The tech giant, IBM is teaming its Watson AI supercomputer with academic researchers to try to predict from speech patterns whether patients are likely to develop a psychotic disorder. Even the US Army got interested! In May, 2016 it launched a partnership with MIT researchers with the goal of developing an FDA-approved device to detect brain injury. But that’s not all!
Let me show you the potential areas where diagnosticians could use vocal biomarkers successfully!
What if coronary artery disease could be identified over a phone call?
An Israeli company, Beyond Verbal deals with emotion analytics and provides voice analysis software. It has announced that its algorithms were successful in helping to detect the presence of coronary artery disease (CAD) in a group of patients.
The research was presented at the American Heart Association Scientific Sessions in New Orleans, Louisiana in December 2016; carried out in cooperation with Mayo Clinic. It involved a double-blind study with 120 patients undergoing both an angiography and a voice analysis; and a group of controls. Beyond Verbal used a smartphone app to measure their voice signal prior to a coronary angiograph. One voice characteristic in particular indicated an almost 20-fold increase in the likelihood of CAD. Yuval Mor, CEO of Beyond Verbal said that these vocal features are not audible by the human ear alone.
Now, imagine all the implications of the research! Imagine how easy it would be if medical professionals could identify patients with CAD over a phone call! There would be no need to go to the doctors’ office, wait for hours for costly examinations and days for the results.
Say some words, I tell you whether you are depressed!
Sonde Health Inc., a Boston-based company develops a voice-based technology platform for monitoring and diagnosing mental and physical medical conditions, with the help of a technology licensed from the MIT Lincoln Laboratory. Researchers designed the technology to enable analysis of brief voice samples to screen and monitor for a range of mental and physical medical concerns based on subtle changes in acoustic characteristics of the speaker’s voice.
“The ability to help recognize early signs of psychiatric illness and monitor treatment responses on devices that people already own is an important step in moving from reactive to preventive care,” said Aimee Danielson, Ph.D., Director, Women’s Mental Health Program at MedStar Georgetown University Hospital. “This would be particularly useful in conditions that are chronically underdiagnosed, like perinatal mood and anxiety disorders, including postpartum depression.”
Another Boston-based company, Cogito, is developing a voice analysis app. The US Department of Veterans Affairs started to use it already to monitor the mood of service members. Its testing also started on patients with bipolar disorder and depression. It is amazing how they are able to detect depression in individuals based on a brief sample of spoken words! I hope the technology will be widely available soon.
What if voice could reveal Parkinson’s disease?
It is time-consuming, difficult and expensive to carry out standard medical tests for Parkinson’s disease. Scientists found though that the chronic illness affects limb movements as much as voice – and some of them started to develop a technology for easier diagnosing the illness. Through voice.
For proving the efficiency of the new diagnostic tool, a couple of researchers started the Parkinson’s Voice Initiative. It aims to record 10,000 voices across the world – to collect enough recordings to introduce the jaw-dropping technology on a wider scale.
This could enable some radical breakthroughs, because voice-based tests are as accurate as clinical tests, but additionally, medical professionals could administer them remotely, and patients can do the tests themselves. Also, they are high speed (take less than 30 seconds), and are ultra low cost. So, they are massively scalable. Amazing, isn’t it? Would you like to facilitate the great cause? You’ve got the chance to participate in their smartphone data-gathering or PVA Voice test, so do it!
Start-ups and researchers interested in vocal biomarkers, let’s unite!
In 2016, Beyond Verbal announced the launch of Beyond mHealth Research Platform. It aims to collaborate with research institutes, hospitals, businesses, and universities to collectively search for unique markers in voice samples.
This is a fantastic idea since more data could lead to more accurate results and better solutions for patients. One of them might be the use of AI algorithms for analyzing billions of voice samples and suggesting a potential diagnosis. Beyond Verbal, which has gathered more than 2.5 million voice samples in 40 languages, wants to build such an algorithm with a virtual assistance. Thus, it encourages other institutions, researchers and others interested in collecting voice data through smart cars, the Internet of Things devices in smart homes or personal assistants like Siri and Alexa. It is a wonderful idea and I hope they will succeed!
It’s obvious that we need more research in the field of vocal biomarkers. We also need not to forget the ethical issues concerning the voice recordings. Although algorithms analyzing samples may not be interested in the content of the speech; nevertheless listening to someone talking to someone else over the phone and analyzing the vocal biomarkers in it, could constitute a serious breach of privacy. What should we do about that?
Well, regulators need to keep up with the development of the field of vocal biomarkers, and consider privacy issues before the sea of vocal analytics apps reach the market.