- Asymptomatic COVID-19 patients exhibit no symptoms of infection however, they can still infect others and spread the disease further.
- Researchers at MIT have designed an AI model that distinguishes between asymptomatic and symptomatic COVID-19 patients using forced cough recordings
- The AI model accurately identified 98.5% of coughs from symptomatic COVID-19 patients.
The World Health Organization (WHO) estimates that approximately 80% of SARS-CoV-2 infections result in mild or asymptomatic COVID-19. These individuals exhibit no physical symptoms of the disease. However, it seems they might have a cough distinguishable from healthy individuals. A change not easily discerned by the human ear.
Therefore, researchers at the Massachusetts Institute of Technology (MIT) have developed an AI (artificial intelligence) model that can differentiate healthy individuals from asymptomatic COVID-19 patients using forced-cough recordings.
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, co-author and a research scientist in MIT’s Auto-ID Laboratory
Asymptomatic COVID-19 Patients – How to tell them apart?
The team of researchers first collected forced-cough recordings from healthy individuals and COVID-19 patients. To do so, they set up a website in April where patients could send their recordings of a forced-cough. Furthermore, they could fill out a survey of their symptoms, COVID-19 status, and status of their COVID-19 test.
The AI model was trained using 4,000 of the collected samples along with samples of spoken words. They then tested the model using the remaining 1,000 recordings.
The team published their findings in the IEEE Journal of Engineering in Medicine and Biology.
The AI model accurately identified 98.5% of coughs from confirmed COVID-19 patients. Additionally, out of those it detected 100% of asymptomatic patients who did not have symptoms but tested positive for the virus.
AI Framework for Alzheimer’s Used for COVID-19
Before the pandemic, the researchers had been working on an AI framework that could detect muscular degradation using audio recordings and, diagnose Alzheimer’s.
When evidence appeared of COVID-19 patients experiencing similar neuromuscular impairment, Brian and his team decided to tweak the Alzheimer’s framework for diagnosing COVID-19.
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. 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.
Brian Subirana, co-author and a research scientist in MIT’s Auto-ID Laboratory
However, researchers argue the AI model is not meant for diagnosing COVID-19 but rather differentiating asymptomatic coughs from healthy ones.
Currently, the team is working on creating a user-friendly app that can serve as a prescreening tool to identify those who are asymptomatic.
Reference:
Laguarta, J., Hueto, F., & Subirana, B. (2020, September 30). COVID-19 Artificial Intelligence Diagnosis using only Cough Recordings. Retrieved October 29, 2020, from https://www.embs.org/ojemb/articles/covid-19-artificial-intelligence-diagnosis-using-only-cough-recordings/
