IS VOICE A NEW BIOMARKER? While most of us have a voice change during puberty and another as aging weakens our voice, a third voice shift may signal the onset of disease or a mental health decline. Today, we look at the promise of vocal biomarkers to aid in diagnosing and managing conditions such as dementia, mild cognitive impairment, depression, and more. Voice is a health indicator.
If I ask to check your vital signs, you probably think I am going for your pulse, blood pressure, or breathing rate. But these investigations are often not particularly specific to a disease or condition.
Now comes news that voice patterns may provide important psychological and physical health clues. I can imagine a day when we talk into our cell smartphone or computer and get a medical diagnosis.
Voice is a health indicator
While you or others may not notice a change in your voice, voice analysis may help identify concerning medical conditions earlier than is possible with current approaches. Recognizing this promise, the US National Institutes of Health has created the Voice as a Biomarker of Health project.
Artificial intelligence (AI) has opened the door to new opportunities in digital health. Voice technology appears promising, with the global voice market expected to be up to $5.4 billion.
Virtual assistants on smartphones or smart home devices, such as connected speakers, are becoming common. Here are some statistics from 2019:
In 2019, 31 percent of smartphone users worldwide used voice tech at least once a week, and 20 percent of queries on Google’s mobile app and Android devices were voice searches.
Biomarkers
A biomarker is “a factor objectively measured and evaluated which represents a biological or pathogenic process, or a pharmacological response to a therapeutic intervention, which can be used as a surrogate marker of a clinical endpoint.”
A vocal biomarker is a signature or combination of voice audio signal features associated with a clinical outcome. The vocal marker can help to:
- Monitor patients
- Diagnose a condition
- Grade the severity or stages of a disease or drug development.
Vocal biomarkers and Alzheimer’s disease
Before we see other Alzheimer’s symptoms, slight changes in voice or language may be apparent. We may also identify such alterations in the early stages of mild cognitive impairment.
Both of these conditions affect verbal fluency; individuals may be hesitant to speak, and their speech rate may slow. Some experience word-finding challenges or add frequent filler sounds (such as “uh”).
Moreover, individuals with dementia or other cognitive impairment may exhibit repetitions, speak novel words (neologisms), or have grammar or word simplification. In addition, there can be variations in pitch and modulation, as well as speech rhythm. Some individuals introduce implausible and irrelevant details.
You can see how voice features can potentially become a non-invasive biomarker for an earlier mild cognitive impairment or dementia.
Vocal IS A mental health indicator
Stress can change our voice. Smartphone-based self-assessed stress is correlated with voice features.
High levels of stress affect our verbal interactivity. When cortisol levels are high (as is more common among those with depression), voice symptoms are more common. We can use voice characteristics to suggest a diagnosis of depression.
Researchers used an automated telephone system to correlate vocal and acoustic measures of depression severity and treatment response.
Voice analysis may target other mental health conditions, including schizophrenia. This condition may be marked by disjointed idea flow, digressions from the topic, or nonsensical word associations. Those with bipolar disease or a histrionic personality disorder may have circumstantial speech patterns.
Voice is a COVID-19 indicator
Might voice recognition software help us grade a respiratory infection’s severity (such as COVID-19)? The answer is not surprisingly yes; COVID-19 is primarily a respiratory condition affecting breathing and voice. It can cause a dry cough, sore throat, voice change, and breathing rhythms.
Cambridge University (England) researchers used vocal biomarkers to help diagnose COVID-19 infection. Scientists at the Massachusetts Institute of Technology (USA) have demonstrated the promise of cough recordings.
To learn more about the recording of voice, please go here:
https://www.mayoclinicproceedings.org/article/S0025-6196%2818%2930030-2/fulltext
Voice biomarkers — Into the future
Twelve research institutions are exploring using artificial intelligence and machine learning to help doctors diagnose and manage diseases, such as depression, based on voice.
The United States National Institutes of Health (NIH) is providing $14 million in funding over four years. The study includes institutions in the USA and Canada.
This project is one of several funded by a new NIH Common Fund program called Bridge to Artificial Intelligence (Bridge2AI). French-American AI biotechnology startup Owkin is supplying the database technology.
The Voice as a Biomarker of Health project aims to “ethically collect hundreds of thousands of human voices while ensuring diversity and patients’ privacy.” Researchers will train machine learning models to identify diseases by detecting human voice changes at a low cost.
Imagine a world in which wearable devices with voice (or movement) analysis would alert us if we needed to seek a medical evaluation.
Thank you for joining me for this look at the promise of vocal biomarkers to aid in diagnosing and managing conditions such as dementia, mild cognitive impairment, depression, and more. Please remember to sign up to follow me on this blog site. Thank you!
The information I provided in this blog is for educational purposes only and does not substitute for professional medical advice. Please consult a medical professional or healthcare provider if you seek medical advice, diagnoses, or treatment. I am not liable for risks or issues associated with using or acting upon the information in this blog.
