American researchers are about to launch an app that could play an important role in containing the pandemic. Because it filters a large proportion of those who are infected who do not have symptoms of their own, but still can infect others.
Experts currently stress it almost daily: The number of new corona infections is also so high because more are being tested. In other words, there are probably a lot of infected people in the spring who were not attracting attention. This is probably also related to the fact that Covid-19 is completely symptom-free in some affected people. Therefore, they see no reason to go to the doctor. So you carry the virus unknowingly and you could infect others. But does it really exist, completely free of symptoms? According to scientists at the Massachusetts Institute of Technology (MIT), this is an illusion. They are convinced that the virus can certainly bring about change. This is exactly what they showed in a study – and programming artificial intelligence (AI) to recognize these differences. This new approach could help limit the spread of the pandemic.
Artificial cough as a data base
MIT researchers have a lot of experience with such types of diagnosis. Because they programmed pre-Covid-19 algorithms and trained them to analyze coughing sounds recorded via smartphones. They were originally looking for ways to diagnose diseases such as pneumonia or asthma. In the next step, they taught the AI to contribute to the diagnosis of Alzheimer’s. Because this is not only related to memory loss, but also to physical changes. Among other things, neuromuscular degradation can lead to weaker vocal cords. They used their findings in the current study to diagnose Covid 19.
++ infected with corona at work: is that really an accident at work? ++
The scientists initially collected tens of thousands of recordings, including artificial coughs and speech samples. Participants presented their sounds through web browsers and devices such as mobile phones and laptops, not through a standard professional recording device. They used this data to train artificial intelligence. They then tested whether the algorithms would recognize cough samples from 2,500 Covid 19 patients – the hit rate was 98.5%. They then checked how many of those infected had not noticed any complaints. The AI seems to have identified all people without symptoms with Covid-19. However, MIT has not revealed how many were.
Alzheimer’s biomarkers transferred to Covid-19
In fact, the scientists programmed not only one algorithm for their AI, but several. They again focused themselves on their research on the diagnosis of Alzheimer’s. To do this, they had developed four biomarkers, namely vocal cord strength, mood, pulmonary and respiratory performance, and muscle breakdown. Your thesis: The way a person produces sounds changes when he or she has Corona.
“The sounds of talking and coughing are influenced by the vocal cords and surrounding organs. This means when you talk, part of the conversation is basically like coughing, and vice versa. It also means that things we can easily glean from fluency in speaking, including factors such as gender, native language, or even an individual’s emotional state, can be coughed. In fact, the mood is rooted in the way you cough, ”said Subirana. “So we thought, why don’t we try these Alzheimer’s biomarkers for Covid?”
App for early detection
The researchers are currently in the process of converting their program into a user-friendly app. They hope it could become a useful pre-screening tool. The user could cough into their smartphone every morning to analyze any changes. What is clear to the scientists, however, is that a clear result would have to be confirmed by the official Covid-19 test. “Effective implementation of this group diagnostic tool could reduce the spread of the pandemic if everyone uses it before going to a classroom, factory or restaurant,” said Brian Subirana, co-author of MIT.
Read more about Covid-19:
Looking for a job for engineers