American students from the MIT (Massachusetts Institue of Technology) developed a smartphone app that can help users detect if they are infected with the novel coronavirus based on the way they cough.
The scientists found out that all infected patients, even asymptomatic ones, have a very particular trait in the way they cough.
The human ear can not detect the sounds that are specific to the condition provoked by the novel coronavirus, but computers can, according to pulmonologist Lieven Dupont.
“It uses software, with artificial intelligence (AI),” he added.
It was initially aimed to recognize patients with Alzheimer’s disease, but with some tweaking and adaptations, it can also detect infection with the novel Coronavirus.
The app derives four parameters from the way someone speaks and coughs – their vocal cords strength, mind state, muscular degradation, and lung capacity.
To adjust and calibrate the software, the researchers gathered over seventy thousand recordings of forced-coughs sent in by volunteers.
About 2,500 of the reports came from patients who tested positive for the disease, including asymptomatic patients.
When the program processed all the coughs, it detected 98.5% of infected people.
The algorithm isn’t supposed to detect people who show symptoms, co-developer Brian Subirana stressed that the tool’s strong point is its ability to discern asymptomatic people from healthy individuals with regular coughs.
Researchers have found that asymptomatic Covid-19 patients may have distinctive coughs. Indecipherable to the human ear, the differences can be picked up by artificial intelligence. Cellphone-recorded coughs could provide a convenient screening tool. https://t.co/OyHqC2KqNk pic.twitter.com/0kYdEIXQkd
— Massachusetts Institute of Technology (MIT) (@MIT) October 29, 2020