AI recognizes corona infected people by the coughing sound |

Researchers at the Massachusetts Institute of Technology (MIT) have developed an AI-powered mobile app that uses the sound made when coughing to diagnose COVID-19. […]

Juan Brian Subirana’s team trained the smart analysis software with tens of thousands of coughing noises from healthy and infected people. In parallel, the subjects were tested for COVID-19.

Pre-screening app as a target

“Our tool could reduce the spread of the pandemic if everyone uses it before going to a classroom, factory or restaurant,” says Subirana. The team is now working with Japan’s Takeda Pharmaceutical Company and hospitals around the world to develop a free pre-screening app based on the AI ​​model.

Even before the devastating pandemic began at the beginning of this year, MIT scientists had investigated coughing noises as a means of targeted diagnosis. They aimed at detecting pneumonia and asthma. At MIT, the focus was then more and more on the possibility of recognizing Alzheimer’s disease on the basis of coughing noises. This disease is characterized not only by the destruction of brain cells, but also by changes in the vocal cords due to neuromuscular degradation.

Similarities to Alzheimer’s

When the pandemic broke out, there were indications that people infected with corona also experience neurological symptoms such as temporary neuromuscular problems. “So we thought, why don’t we try these Alzheimer’s biomarkers for COVID-19 too,” says Subirana. There were similarities between the coughing sounds made by people with Alzheimer’s disease and those infected with the virus. Tests showed that the artificial intelligence (AI) -based process recognized 98.5 percent of those infected.