free counter
Health And Medical

Sound Sick? New AI Technology Might Tell If Its COVID

Sept. 19, 2022 — Imagine this: You imagine you may have COVID. You speak a few sentences into your phone. Then an app offers you reliable results within a minute.

You sound sick is what we humans might tell a pal. Artificial intelligence, or AI, could take that to new frontiers by analyzing your voice to detect a COVID infection.

A cheap and simple app could possibly be found in low-income countries or even to screen crowds at concerts along with other large gatherings, researchers say.

Its just the most recent example in a rising trend exploring voice as a diagnostic tool to detect or predict diseases.

In the last decade, AI speech analysis has been proven to greatly help detect Parkinsons disease, posttraumatic stress disorder, dementia, and cardiovascular disease. Research has been so promising that the National Institutes of Health just launched a new initiative to build up AI to utilize voice to diagnose several conditions. These range between such respiratory maladies as pneumonia and COPD to laryngeal cancer and also stroke, ALS, and psychiatric disorders like depression and schizophrenia. Software can detect nuances that the human ear cant, researchers say.

At the very least six studies took this process to COVID detection. In the newest advancement, researchers from Maastricht University in holland are reporting their AI model was accurate 89% of that time period, compared with typically 56% for various lateral flow tests. The voice test also was more accurate at detecting infection in people not showing symptoms.

One hitch: Lateral flow tests show false positives significantly less than 1% of that time period, weighed against 17% for the voice test. Still, because the test is virtually free, it could be practical to just have those that test positive take further tests, said researcher Wafaa Aljbawi, who presented the preliminary findings at the European Respiratory Societys International Congress in Barcelona, Spain.

I’m personally excited for the possible medical implications, says Visara Urovi, PhD, a researcher on the project and a co-employee professor at the Institute of Data Science at Maastricht University. If we better know how voice changes with different conditions, we’re able to potentially know whenever we are going to get sick or when to get more tests and/or treatment.

Developing the AI

A COVID infection can transform your voice. It affects the respiratory system, producing a insufficient speech energy and a lack of voice because of shortness of breath and upper airway congestion, says the preprint paper, which hasnt been peer reviewed yet. A COVID patients typical dry cough also causes changes in the vocal cords. And previous research discovered that lung and larynx dysfunction from COVID changes a voices acoustic characteristics.

Section of what makes the most recent research notable may be the size of the dataset. The researchers used a crowd-sourced database from the University of Cambridge that contained 893 audio samples from 4,352 people, of whom 308 tested positive for COVID.

It is possible to donate to this database its all anonymous — via Cambridges COVID-19 Sounds App, which asks one to cough 3 x, breathe deeply through the mouth 3 to 5 times, and read a brief sentence 3 x.

Because of their study, Maastricht University researchers only centered on the spoken sentences, explains Urovi. The signal parameters of the audio provide some info on the power of speech, she says. It really is those numbers which are found in the algorithm to create a decision.

Audiophiles could find it interesting that the researchers used mel spectrogram analysis to recognize characteristics of the sound wave (or timbre). Artificial intelligence enthusiasts will remember that the study discovered that long short-term memory (LSTM) was the sort of AI model that worked best. Its predicated on neural networks that mimic the mind and is particularly proficient at modeling signals collected as time passes.

For laypeople, its enough to learn that advancements in the field can lead to reliable, efficient, affordable, convenient, and simple-to-use technologies for detection and prediction of disease, the paper said.

Whats Next?

Building this research right into a meaningful app will demand an effective validation phase, says Urovi. Such external validation — testing the way the model works together with another dataset of sounds — could be a slow process.

A validation phase may take years prior to the app could be distributed around the broader public, Urovi says.

Urovi stresses that despite having the large Cambridge dataset, it really is hard to predict how well this model my work in the overall population. If speech testing is proven to work better when compared to a rapid antigen test, people might choose the cheap non-invasive option.

“But more research is necessary exploring which voice features are most readily useful in selecting COVID cases, also to make certain models can tell the difference between COVID along with other respiratory conditions,” the paper says.

So can be pre-concert app tests inside our future? Thatll be determined by cost-benefit analyses and several other considerations, Urovi says.

Nevertheless, It could still bring benefits if the test can be used in support or along with other well-established screening tools like a PCR test.

Read More

Related Articles

Leave a Reply

Your email address will not be published.

Back to top button

Adblock Detected

Please consider supporting us by disabling your ad blocker