Asymptomatic patients are part of the population that often goes unnoticed for having coronavirus in the first place. In doing so, they can pass on the virus unknowingly to many other people, most of whom may develop serious complications because of the virus. Now a recent development by researchers reveals a special model that can detect and differentiate between symptomatic and asymptomatic patients through a mere recording of how they cough.
The novel application operates using a model of artificial intelligence and can be used on a smartphone. Developed by researchers from the Massachusetts Institute of Technology (MIT) in the United States, the app manages to differentiate between healthy individuals and asymptomatic patients through coughing pattern recordings alone.
Although such differences are impossible for the human ear to hear and understand, the artificial intelligence model can pick up on these differences and distinguish between them.
The paper that describes the development is published and out in the IEEE Journal of Engineering in Medicine and Biology. The team describes it as an AI model that can tell the difference between healthy people and those with the virus through self-induced cough recordings.
These recordings came from different volunteers coming through web browsers and devices such as laptops and cellphones, the researchers on the study explain.
The researchers then taught the model’s latest recordings of coughs, in doing so it began to accurately identify 98.5 % of the coughs from individuals that were diagnosed with COVID-19.
This also includes 100 % of cough recordings coming from asymptomatic people. These people although do not have any symptoms of the lethal virus, do test positive for the virus and have it in their bodies, according to the study members.
Currently, the team is spending time incorporating this useful model into a concise user-friendly application for smartphones. If it gets approved the tool could be taken on a large level and will not only be free to use but also will be an option that is noninvasive and convenient for a pre-screening tool. This can then be used to identify those that are most probably going to be asymptomatic sufferers for COVID-19, say the researchers.
Brian Subirana is a research scientist at a lab at MIT. He believes that this tool as a diagnostic instrument could greatly help reduce the infection rate for the virus if people use it each time before they enter a classroom, restaurant, or even a factory.
He means this way people would be able to understand whether or not they are one of the symptom-less sufferers and are a risk to the people around you. Once a person is diagnosed they will be able to effectively self-isolate and reduce the major harm they could cause by moving about in heavily populated areas.
Before the coronavirus pandemic shook the world, research groups have been successful in training algorithms to pick up on cough recordings for diagnosing conditions such as asthma and pneumonia.
The team of researchers has acquired more than 70,000 recordings, each of which has many coughs recorded in it. These are as many as 200,000 samples of forced coughs. According to Subirana, these could be the largest dataset for cough research to date.
To conclude Subirana says that the research believes that the way we sound could change with the infection of the coronavirus even if the person in question has no symptoms at all. This research offers hope for a non-invasive and more accessible way of diagnosis for a virus that has altered the way of living for the entire world we live in.