Detection of COVID-19 Via Voice and Acoustic Biomarkers

Issue 17 - Special Series: Voice + Acoustics Front and Center in Healthcare

The content and distribution of Azafran’s INSIGHTS newsletter is focused to our LP, incubator, research, investment and partner ecosystem. As we look to build a two-way dialogue benefitting our collective efforts, each month we highlight important news and our approach to the emerging intersection of deep technology, end to end solutions and platforms driven by voice, acoustics/sensory and imagery.

This article is one section of an entire issue of INSIGHTS. Please sign up to receive access to past and new issues as they are published.

Subscribe to our newsletter

Detection of COVID-19 Via Voice and Acoustic Biomarkers

A Framework for Biomarkers of COVID-19 Based on Coordination of Speech-Production Subsystems

Following is an excerpt from a recently published paper in the IEEE Open Journal of Engineering in Medicine and Biology (May 2020). While a bit deeper into the technical aspects than we typically travel in INSIGHTS editions, the paper shows the leading edge of thinking around the power and full spectrum of the acoustic instrument that humans now have at our disposal.

The fact that we can go this deep, this soon into the COVID-19 crisis bodes well as a proof point towards the power of voice and acoustics in the future of healthcare, opening the door for non-invasive, real-time analysis of our condition and changes in those conditions by the minute, hour, day, in a doctor’s office, hospital or at home.

Welcome to the future, today - please visit for a full copy of the paper. The excerpt below has not been modified and is available via Creative Commons license - link here.

Contributors: Thomas F. Quatieri , Fellow, IEEE, Tanya Talkar , Member, IEEE, and Jeffrey S. Palmer , Senior Member, IEEE

COVID-19 is often characterized by specific dysfunction in respiratory physiology including the diaphragm and other parts of the lower respiratory tract, thereby affecting patterns of breathing during inhalation and exhalation of air from the lungs. In speech production, during the exhalation stage, air from the lungs moves through the other essential vocal subsystems, i.e., through the trachea and larynx and into the vocal tract pharyngeal, oral and nasal cavities (Fig. 1). The manner in which we breathe in speaking, including the rate and length of an exhalation (coupled to the number of words in a phrase or sentence), and its intensity and variability, highly influences the quality of our voice.

For example, the loudness, aspiration (‘breathiness’), steadiness of fundamental frequency or ‘pitch’ during phonation, and the mechanism by which we alter speaking rate all effect vocal quality. Furthermore, the respiratory system is highly coordinated with these primarily laryngeal-based subsystems. Likewise, in turn, laryngeal activity is finely coupled to articulation in the oral and nasal cavities. Although impact on speech subsystems and their coordination are often perceptually obvious with a condition involving inflammation, these changes can be subtle in the asymptomatic stage of an illness, i.e., incubation and early prodromal periods, or in recovery.

We have established a framework for discovery of vocal biomarkers of COVID-19 based on the coordination of subsystems of speech production involving respiration, phonation, and articulation. Our preliminary results, using a very limited data set, hint at support of the hypothesis that biomarkers derived from measures of vocal subsystem coordination provide an indicator, possibly in its asymptomatic stage, of COVID-19 impact on respiratory function.

Given a sample size of five subjects and a growing list of COVID-19 symptoms, however, validation of our hypothesis will clearly require additional data and analysis to address potential confounders such as different recording environments and channels, unbalanced data quantities, and changes in underlying vocal status from pre-to-post time recordings. It will also be important to expand the suite of vocal features, introducing neurophysiological modeling of subsystem interactions, to address the increasing evidence of neurological insult arising from COVID-19 and feature specificity relative to typical flu and flu-like symptoms.

© Copyright 2017 - 2020, Azafran Capital Partners, Inc.