Brilliant apps may be able to give you a coronavirus test instantly – BGR

• Many people infected with the new coronavirus are asymptomatic or presymptomatic, but they can still be transmitted.
• MIT researchers have developed what could be a great screening tool that provides a COVID-19 test instantly each time you leave home.
• Using machine learning algorithms trained to detect Alzheimer’s disease through forced cough, researchers believe that the same technique can be used to detect asymptomatic COVID-19 cases.

---

The combination of factors makes it especially difficult to tame a new coronavirus. First of all, it spreads incredibly easily, especially indoors. Second, it should take 2 to 14 days for symptoms to appear. It is the same even if there are no symptoms at all. Many people get asymptomatic illnesses, but they can still spread it to others. This causes an infected person, which is the third function of the virus. Infect before symptoms appear A person who looks perfectly healthy, even if not asymptomatic, can pass on the disease to another person.

Capturing asymptomatic and stopping the spread of asymptomatic (including presymptomatic infections) may help governments reduce and control the spread. And in the future, a simple app that runs on the phone or smart speaker could provide asymptomatic people with the COVID-19 test immediately after a single cough.

People who cough are no longer asymptomatic because coughing is a symptom. But what MIT researchers can do depends on forced coughing. A healthy person can cough on command, and it may be enough for artificial intelligence (AI) to tell you if you have been infected with the new coronavirus. Such apps can provide immediate results without the need for professionally managed testing. The results may not always be perfect, but a positive diagnosis means that the person needs to seek immediate confirmation from the PCR test.

This app relies on research completed by MIT researchers for Alzheimer’s disease. They used three machine learning algorithms to train AI and recognized signs of Alzheimer’s disease, which causes not only memory loss, but also neuromuscular deterioration that damages the vocal cords. Vocal cords are associated with coughing, and people may not be able to distinguish Alzheimer’s cough by ear, but AI can.

One of the neural networks used by the solution was taught to distinguish sounds and associate them with vocal cord strength. AI learned how to interpret the sound “mmmm” from a 1,000-hour audiobook speech to picking up the words “them,” “the,” and “then.” The second neural network was trained to pick up emotional states because people with Alzheimer’s disease show emotions such as frustration and apathy more often than others. A third machine learning algorithm worked in a cough database to measure changes in lung and respiratory performance.

Combined with an algorithm that detects muscle deterioration, the resulting AI provided a voice recording of the cough. AI has been found to be able to detect Alzheimer’s disease very accurately through forced coughing. When the new coronavirus pandemic began, researchers tried the same principle.

“The sound of speaking and coughing is affected by the vocal cords and surrounding organs. That is, when speaking, part of the speech is like a cough and vice versa. Also, fluent speech. What can be easily derived from, AI also means that you can easily pick up a person’s gender, native language, and even emotional state from a cough. In fact, emotions are embedded in the way you cough. ” Brian Subirana Were told MIT News.. “So why not try these Alzheimer’s disease biomarkers? [to see if they’re relevant] For COVID. “

The researchers went to work, created a website where participants could register a series of forced coughs, investigated symptoms and created official COVID-19 test results. Scientists have collected more than 70,000 recordings with about 200,000 coughs. Among them, 2,500 came from people who tested positive, including asymptomatic.

The team trained AI on over 4,200 samples using 2,500 coughs from healthy people and over 2,500 coughs. The AI ​​then set about processing the remaining 1,000 samples. Researchers have stated in a paper on the issue that this experiment resulted in “significant similarities between Alzheimer’s disease and discrimination against Covid.”

The COVID-19 model picked up 98.5% of coughs from people with COVID-19, and it sounds very promising. “I think this shows that even asymptomatic, using Covid changes the way sound is produced,” says Subirana.

The team is working on incorporating the model into an app that, if approved by the FDA, can be used as an affordable or free screening tool. Even if the tool is not 100% accurate, it may be enough to move people to take the test. Combined with rapid testing and contact tracing apps, such screening tools can have dramatic impacts.

Even if an asymptomatic person doesn’t think anything is wrong, it’s no wonder they are affected by the virus. Blood test and CT scan Previous studies have shown that the body is affected. They may show no symptoms, but their immune system is still in an invisible battle. Therefore, it is no wonder that muscle deterioration can temporarily affect the vocal cords and cough.

MIT papers are fully available With this link..