By Kylie Foy, MIT Lincoln Laboratory — 9 July 2020
MIT Lincoln Laboratory researchers are detecting vocal changes in COVID-19 patients even when those changes are too subtle for people to hear or notice in themselves. By processing speech recordings of people infected with COVID-19 but not yet showing symptoms, researchers found evidence of vocal biomarkers — measurable indicators of the disease. A technology letter describing this research was published in IEEE Open Journal of Engineering in Medicine and Biology.
The biomarkers stem from disruptions the infection causes in the movement of muscles across the respiratory, laryngeal, and articulatory systems. The researchers found decreased complexity of movement in COVID-19 interviews compared to pre-COVID-19 interviews, particularly in the coupling between respiratory and larynx motion.
The team is now looking at using mobile apps to implement this research, with a partnership underway to integrate vocal screening for COVID-19 into existing apps. “A sensing system integrated into a mobile app could pick up on infections early, before people feel sick or, especially, for subsets of people who don’t ever feel sick or show symptoms,” said research group leader Jeffrey Palmer.