Researchers from MIT recently developed an A.I. model which can distinguish between asymptomatic patients of COVID-19 and healthy individuals, MIT News reports. The algorithm for this A.I. relies on subtle differences between the way that healthy people and asymptomatic people cough.
Asymptomatic patients are often at a higher risk for spreading the virus since they do not exhibit any obvious signs of being sick. The algorithm developed has been shown to be 100 percent effective in detecting asymptomatic carriers of COVID-19. This new development could be a breakthrough for COVID-19 testing since it would be a convenient screening procedure for asymptomatic patients.
The team of researchers over at MIT developed the algorithm by listening to forced cough recordings. This A.I. model is already being tested in clinical trials and it is hoped that soon it will be approved by the FDA, which will help with the remote and non-invasive testing.
Developing the algorithm
The algorithm for this model is based on previous A.I. models developed by MIT for diagnosing Alzheimer’s. Back in April when the pandemic was gaining steam, the team decided to collect data by collecting as many cough recordings as they could. This was done by setting up a website where people could submit their recordings through phones or laptops. This was followed by the participants filling out a survey where they described their symptoms, whether they were COVID-19 positive and if they had been tested in a lab or had self-diagnosed. This resulted in over 70,000 recordings being collected out of which 2500 belonged to people confirmed to have the coronavirus (this included people who were asymptomatic).
The paper detailing this A.I. model was published in IEEE Journal of Engineering in Medicine and Biology on 30th of September, and was authored by Brian Subirana, Jordi Laguarta and Ferran Hueto, of MIT’s Auto-ID Laboratory. After the model was developed, it was fed new cough recordings and it was able to correctly identify 98.5 percent of the people who had tested positive for coronavirus which included identifying 100% of people who had tested positive for the virus but were asymptomatic. “The effective implementation of this group diagnostic tool could diminish the spread of the pandemic if everyone uses it before going to a classroom, a factory, or a restaurant,” says co-author Brian Subirana.
The research team is currently working on developing a free pre screening app based on their A.I. algorithm. They are also working on collecting more cough recordings in order to have more diverse data which will help the A.I. model diagnose people from different age groups and ethnic backgrounds.
Do you want to publish on Apple News, Google News, and more? Join our writing community, improve your writing skills, and be read by hundreds of thousands around the world!