MIT team cuff detector identifies 97% of COVID-19 cases even among asymptomatic people

Part of the challenge in controlling the coronavirus epidemic is in quickly identifying and isolating infected people – not particularly easy when COVID-19 symptoms are not always noticeable, especially early. Now scientists have developed a new artificial intelligence model that can detect the virus with a simple forced cough.

Evidence suggests that AI can distinguish cough that cannot be heard with the human ear, and if the detection system can be incorporated into a device like a smartphone, the research team thinks it is a useful Can be an early screening tool.

The work builds on research that was already taking place in detecting Alzheimer’s through coughing and interactions. Once the epidemic began to spread, the team instead drew attention to COVID-19, which had already learned how the disease could cause very small speech changes and other noises that we We do.

“Talking and coughing sounds are influenced by vocal cords and surrounding organs,” says research scientist Brian Subirana of the Massachusetts Institute of Technology (MIT).

“This means that when you talk, part of your talk is like a cough, and vice versa.”

“It also means that the things we easily get from fluent speech, AI can only pick up from coughing, including things like a person’s gender, mother tongue, or even emotional state. Really emotions. Is how you cough. “

Alzheimer’s research was redesigned for COVID-19, which included a neural network called ResNet50. It was trained on one thousand hours of human speech, then on a dataset of spoken words in different emotional states, and then on a cough database for changes in lung and respiratory performance.

When the three models were combined, a layer of noise was used to filter out the strong cough from the weak ones. Approximately 2,500 people captured the recording of a cough, confirmed to be COVID-19, with AI correctly identifying 97.1 percent of them – and 100 percent in asymptomatic cases.

This is an impressive result, but there is still more work to be done. Researchers emphasize that its main value lies in determining the difference between healthy cough and unhealthy cough in asymptomatic people – not actually diagnosing COVID-19, which would require a proper test. In other words, it is an early warning system.

“Effective implementation of this group’s diagnostic tool can reduce the spread of the epidemic if everyone uses it before going to a classroom, factory, or restaurant,” Subirana says.

The fact that the test is non-invasive actually free to run and to implement adds to its potential utility – while not being designed Diagnosis of People with COVID-19 who are already showing symptoms can tell you if you should isolate and get proper testing done when there are no major symptoms of the virus.

Researchers now want to test the engine on a more diverse set of data, and see if there are other factors involved in reaching the detection rate in such an impressive way. If it falls into the stage of the phone app, then obviously there will be privacy implications, as some of us want to listen to our devices for signs of constantly getting sick.

Once we start putting the coronovirus epidemic behind us, new research can help with cough and Alzheimer’s studies. The data suggest that neural networks require only minor tweaking to adapt to each situation.

“Our research uncovers a striking similarity between Alzheimer’s and COVID discrimination,” the researchers write in their published paper.

“The exact same biomarker can be used as a differentiation tool for both, suggesting that perhaps, in addition to temperature, pressure, or pulse, there are some high-level biomarkers that are particularly specialized in conditions. Can diagnose conditions that have been cut most times. “

Has been published in IEEE Open Journal of Engineering in Medicine and Biology.


Leave a Reply

Your email address will not be published.