Artificial Intelligence Identifies Asymptomatic COVID-19 Infections

Researchers may be able to use artificial intelligence to distinguish asymptomatic individuals from healthy people, resulting in a noninvasive screening tool for providers, a study published in IEEE Journal of Engineering in Medicine and Biology revealed.

For more coronavirus updates, visit our resource page, updated twice daily by Xtelligent Healthcare Media.

Because asymptomatic people exhibit no clear signs of the virus, they may not seek out testing and could unknowingly spread the virus to others.

In past research, teams have trained AI algorithms on cell phone recordings of coughs to accurately diagnose conditions such as asthma or pneumonia.

Similarly, researchers from MIT were developing AI models to analyze forced-cough recordings to see if they could detect signs of Alzheimer’s disease, a condition that is associated with neuromuscular degeneration such as weakened vocal cords. The group found that the AI tool could identify Alzheimer’s samples better than existing models.

When the pandemic hit the US, the team set out to apply their AI method to identify individuals infected with the virus.

“The sounds of talking and coughing are both influenced by the vocal cords and surrounding organs. This means that when you talk, part of your talking is like coughing, and vice versa. It also means that things we easily derive from fluent speech, AI can pick up simply from coughs, including things like the person’s gender, mother tongue, or even emotional state,” said co-author Brian Subirana, a research scientist in MIT’s Auto-ID Laboratory.

“There’s in fact sentiment embedded in how you cough. So we thought, why don’t we try these Alzheimer’s biomarkers to see if they’re relevant for COVID.”

The team aimed to collect as many recordings of coughs as they could, including those from COVID-19 patients. The researchers set up a website where people can record a series of coughs through a cell phone or web-enabled device. Participants can also fill out a survey of symptoms they’re experiencing, whether or not they have COVID-19, and whether they were diagnosed through an official test.

To date, researchers have collected more than 70,000 recordings, each containing several coughs, amounting to some 200,000 forced-cough audio samples. About 2,500 recordings were submitted by people who were confirmed to have COVID-19, including those who were asymptomatic.

Researchers used these 2,500 recordings, along with 2,500 more recordings that they randomly selected from the collection to balance the dataset. They used 4,000 of these samples to train the model, and the remaining 1,000 recordings were then fed into the model to see if it could accurately distinguish coughs from COVID-19 patients versus healthy people.

The algorithm was able to pick up patterns in four biomarkers, including vocal cord strength, lung and respiratory performance, and muscular degradation, which are specific to COVID-19. The model was able to identify 98.5 percent of coughs from people with confirmed COVID-19, and of those, it accurately detected all the asymptomatic coughs.

“We think this shows that the way you produce sound, changes when you have COVID, even if you’re asymptomatic,” Subirana said.

Researchers noted that the AI model is not meant to diagnose symptomatic people, but instead to discern symptomatic coughs from healthy coughs. The team is currently working to develop a free pre-screening app based on the AI model.

Additionally, they’re partnering with several hospitals around the world to collect a larger, more diverse set of cough recordings, which will help train the model and increase its accuracy.

The ultimate goal is to incorporate audio AI models like this one into smart speakers and other listening devices so that people can conveniently get an initial assessment of their disease risk, maybe on a daily basis.

The team expects that use of their AI tool could significantly reduce the impact of the current pandemic, as well as deter others from occurring in the future.

“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,” said Subirana.