The road to finding a treatment for Covid-19 still looks long, and even though we are closer than ever before to finding a vaccine, the virus is proving to be vicious as it is rapidly hitting countries across the globe with a second wave.
The number of cases is on the rise and testing everyone is proving to be difficult, especially with the reported rates of the false-negative results being as high as 50%. To make things even tougher, there is a high number of asymptomatic patients who contracted Covid-19 but do not show any symptoms, which increases the transmission of the virus. And while finding a treatment or a vaccine is essential, the most effective way to stop the spread of the virus is finding a more reliable and readily available testing.
To solve this problem and allow for an easy and more accurate testing, researchers have been looking to use the sound of coughs as a way to detect the virus. As per an article on BBC, Cambridge University was able to identify positive Covid-19 cases with an 80% success rate by using the sound of coughs. However, they are not the only ones. MIT (Massachusetts Institute of Technology) began using an AI (artificial intelligence) model to help detect Covid-19. The AI model was initially developed to detect Alzheimer’s disease using cough recordings. According to MIT, due to similar neurological symptoms between Alzheimer’s and Covid-19 patients, researches decided to check whether the AI model can help detect coronavirus. This was done by asking both healthy and infected people to send a recording of forced coughs through a website. More than 70K forced cough recordings were gathered, in which about 2,500 were from people infected with Covid-19.
The researchers were then able to test the AI model by feeding it the cough recordings in order to detect the differences between healthy and infected people. It was then deduced that four features, namely muscular degradation, vocal cord strength, sentiment and respiratory and lung performance, helped in determining the differences in coughs. The research team at MIT said that the algorithm had a 98.5% accuracy rate in detecting the coughs of confirmed cases and a 100% accuracy rate in detecting asymptomatic cases.
One of the key aspects to the virus’ rapid spread is the fact that many people who have contacted the virus are asymptomatic, which means they do not show any symptoms. This is where the strength of this tool lies. The use of this AI tool will help people know if they have the virus even without showing any symptoms, which can aid in stopping the spread of the virus.
“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...” - Brian Subirana, MIT research scientist.
While the test is still not approved by the FDA, it still looks promising and might greatly aid in the fight against the virus. The researchers are now working on gathering more samples from hospitals to strengthen their model, as integrating the AI model in a user friendly app that can be available to everyone.
The views/opinions expressed in this article are those of the author and do not necessarily reflect the views and opinions of Al Bawaba Business or its affiliates.
© 2000 - 2021 Al Bawaba (www.albawaba.com)