Before You Swat That Mosquito, Record It on Your Cell Phone

The next time a mosquito buzzes in your ear, do Manu Prakash a favor: Don’t swat it, record it.

That’s right, instead of bemoaning your fate, take your phone and create an audio file of that sound you so hate to hear. Then send it off to Prakash. He’ll very much appreciate it.

In truth, Prakash and his team of bioengineers at Stanford hope millions of people around the world do the same, so that their odd collection of mosquito sounds keeps growing and growing.

This, he says, has “incredible value.”

The value comes from what all those tiny noises can be converted into—a global, continuously updated map of what kind of mosquitoes are where.

Pest sounds

That’s the purpose of an ambitious crowdsourcing project called Abuzz. Its built around the strategy of using the most common personal technology in the world—the cell phone—to not just track the movement of mosquitoes, but also to more precisely target the diseases they spread.

Mosquitoes are, in fact, one of Earth’s deadliest animals, with billions of humans living in regions where malaria, dengue fever, Zika, chikungunya or yellow fever remain a threat. By World Health Organization estimates, they’re responsible for several million deaths a year.

Since there aren’t vaccines for many mosquito-borne infections, the best option for preventing outbreaks is by targeting the pests themselves. But as Prakash points out, those efforts can be scattershot and reactive. Also, field research still pretty much comes down to catching a lot of mosquitoes and looking at them under a microscope, one at a time.

“This project was inspired by a work trip I had taken to a rainforest in Thailand, where I got a chance to meet medical entomologists doing backbreaking work—literally counting mosquitoes under microscopes,” he says. “This is very common in the field.”

Why not try a new approach to mosquito mapping, a more efficient and comprehensive type of evidence-gathering?

In this case, that meant collecting a lot of little clips of bug noise. Most people can’t distinguish one mosquito’s whine from another’s, but the sound of their beating wings does differ from species to species. While there are about 3,500 different species of mosquitoes, only about 20 to 25 are known to spread disease among humans. It was with those that the Abuzz team started building its sound library three years ago, ultimately capturing close to 1,000 hours of mosquito buzzing.

Cell phones for science

At first, they used expensive, “research-grade” audio equipment when they recorded mosquitoes in the lab. Then, one day while they were working, Prakash’s phone rang. He remembers it as the “Aha” moment.

“It made us think, ‘Why don’t we use cellphones to record.’”

The team spent the next six months recording as many different species of mosquitoes as possible on as many different types of phones as it could find, including some cheap flip phones. But the quality of the audio was surprisingly good, according to Prakash. To make sure the crowdsourcing concept could work in more remote areas, the researchers conducted a field test in a small village in Madagascar, described in a study published recently in eLife.

Ten volunteers were given only 10 to 15 minutes of training in using their phones to record mosquitoes. The next day they returned with 60 separate recordings, totaling three hours of buzzing.

Prakash realizes that the simpler the process can be, the more citizen scientists will participate.  Now, someone who wants to contribute a mosquito recording needs to upload the audio file to the Abuzz website. He says a mobile app being developed to simplify things should be available in a few months,  and is hopeful that eventually the sounds will be able to be sent through text messaging.

He also acknowledges that many people have a hard time imagining how they can record a sound that seems so fleeting. The good news is that it doesn’t have to be a very long clip to be useful. One second of buzzing is usually enough for the algorithm the Abuzz team created to match it with a particular species.

To encourage people to contribute, the Abuzz website provides a few how-to videos. One suggests that you put a paper cup over the mosquito, and poke a small hole in the cup’s top. Tap on the side and the bug will start buzzing. Even if you can’t hear it, your phone should be able to capture the sound once you place it near the opening. 

Stanford researchers create "Shazam" for mosquitoes

Watch on

Mapping mosquitoes 

The crowdsourcing aspect of Abuzz is still in its early stages, but submissions are already coming in from far afield. One recent one arrived from Guam, another from a high school student on a small island Prakash had never heard of. Because the audio files are time-stamped and marked with geolocation data, they can provide scientists with useful information about the movement and behavior of a particular mosquito species.

Prakash says this kind of data is increasingly important in tracking the impact of climate change, and also, natural disasters. For instance, time stamps on submissions could reflect an influx of a different mosquito species in a region after extensive flooding, such as what occurred in Houston during Hurricane Harvey.

“This can improve our ability to predict where mosquitoes will spread,” he says. “You can better identify hot spots where you should put your efforts. People are talking about eliminating diseases, but we still don’t really understand why are there certain buckets of density and then other places where there are no mosquitoes. There’s still a lot to be figured out and that requires data.

“When we have very little understanding of ecology, any data is good to have,” he adds. “Mosquitoes only fly a mile or two, so local ecology is as important to understand as global ecology. In fact, it’s more meaningful for policy for local neighborhoods to do high-density mapping.”

While mosquito-borne disease is ultimately a local matter, Prakash believes Abuzz also has value in providing a global perspective. “If we have 10,000 people around the world who record mosquitoes on their cell phones even once a month,” he says, “we would make some of the largest data-driven maps of mosquito ecology.”

Erin Mordecai, an assistant professor of biology at Stanford and an infectious disease expert who has not worked directly with the project, says she thinks Abuzz has much potential for helping scientists and public health officials truly zero in on pest populations.

“Surprisingly, one of the major unknowns in the risk of vector-borne diseases is when and where disease-transmitting mosquitoes are active,” she explains. “Since mosquitoes are excellent at finding human targets, allowing people to directly record and identify the mosquitoes that pester them is a great way to rapidly build knowledge about mosquito risks. 

“Traditional mosquito sampling methods are much more costly and time-consuming,” Mordecai adds, “and they often don’t capture mosquitoes where people are during their day.”

Prakash may be most excited about the prospect of giving people the opportunity to not only help science, but also their own communities. The idea of using simple tools to take on big science and health challenges has been a common thread in his career, whether it’s been in designing a cheap, paper microscope that can used in the field, or inventing an inexpensive centrifuge—based on a whirligig toy—that can be used to analyze blood to detect malaria.  

“It’s really about trying to engage people in the problem instead of it being, ‘Hey, I have a problem, come and fix it.’ What I’d like to see is that people who are in the heart of the problem, in places where the problem is most severe, that they get involved,” he says.

“One of the goals is to bring a much broader audience to this challenge. It’s very, very clear that we cannot do this project alone. No matter how many biologists we can put out in the field, it’s never enough.” 

Randy Rieland | | READ MORE

Randy Rieland is a digital media strategist and contributing writer in innovations for Smithsonian.com.