In 2014, Utkarsh Tandon, at the time a freshman at Cupertino High School in California, developed a machine learning model for his science fair project that collected and classified data on sufferers of Parkinson's disease. He won the fair, and, as part of his first place award, received a grant from the UCLA Brain Research Institute. A year later, the high school sophomore has turned his science experiment into a marketable product. The OneRing, now raising funds on Kickstarter, is a wearable device that monitors Parkinson's patients' tremors and delivers the data to an iOS app in the form of a digestible daily report.
The OneRing, named for the powerful ring at the center of J.R.R. Tolkien's Lord of the Rings, is a 3-D printed plastic ring topped with a flashdrive-like box that houses a Bluetooth microchip. Using an algorithm developed by Tandon, the device senses tremors commonly experienced by Parkinson's patients, classifies them based on severity, and generates a daily report that provides time-stamped analytics about the users' movements during each hour of the day. The movement patterns of the hand are divided into three categories: dyskinesia, bradykinesia, and tremor. "With these classifications it can be packaged in these very coherent patient reports that the physicians and the patients can read and interact with in a way that better recommends medication," Tandon says.
Tandon started learning the basics of artificial intelligence and programming in ninth grade, but the seeds of the OneRing project were planted years earlier. When Tandon was 10, he came across a YouTube video of Muhammad Ali lighting the 1996 Olympic torch. Curious what made Ali, who was diagnosed with Parkinson's in 1984, shake so badly, Tandon started looking into Parkinson's disease and wondering what he could do to help.
Four years later, Tandon found his chance in a computer science class where he was taught machine learning. In his research, he came across a data set online from a study that strapped a phone onto the hands of seven people with Parkinson's disease and recorded their movements. "This data was put online publicly for people who wanted to trace trends in parkinson’s disease," Tandon says. "At this time I was learning about machine learning so I thought, well, why don’t I test this out? I’ve always wanted to build something for Parkinson’s disease patients." His first prototype was a device that strapped onto users' wrists like a watch, but after testing it with patients during an internship at the Parkinson's Institute in nearby Sunnyvale, California, he decided to modify it to be smaller and less cumbersome. To print the devices, he sends the files to the online 3-D printing service Shapeways and they ship him the final product in small batches.
For now, this is how Tandon plans to continue production. With four days to go, the project has more than doubled its $1,500 fundraising goal on Kickstarter. The money will go toward producing OneRings for backers, as well as making the rings available for patients at the Parkinson's Institute so that Tandon can further develop the current model. While the technology embedded in the ring is impressive, the design of the ring itself isn't exactly fashionable—something Tandon is well aware of. "It has to be something people want to wear," he says. "I want to make it look good while it's doing the diagnosis in the background." His next step? Using a flexible polymer material to make the ring one size fits all.