Shwetak Patel and his colleagues at the University of Washington have already figured out how a smartphone can be used to measure blood pressure; to screen for pancreatic cancer; and to check for jaundice in newborns.
Now the innovators are developing smartphone-based technology that could have a huge impact for how concussions are assessed in sports, military, and many other scenarios.
PupilScreen is a new project out of the UW’s Ubicomplab that uses a combination of a smartphone’s camera and machine learning to analyze the pupil of someone suffering from a potential concussion.
The technology is still in its early phases, but researchers — who just published a scientific paper on their findings — want to release a commercial version within two years.
PupilScreen uses a smartphone’s video camera and flash to check the eye for its pupillary light reflex, which measures pupil size changes and can help objectively determine a concussion. The video is then processed by deep learning algorithms.
The idea is for PupilScreen to allow a football coach, for example, to quickly and accurately assess a head injury on the sidelines and determine whether a player is fit to return based on clinically-relevant metrics. It would be the first smartphone app to have this capability, according to the UW.
The project sparked after Dr. Tony Law and Dr. Lynn McGrath, resident doctors with UW Medicine, reached out to Patel, an endowed professor of computer science and electrical engineering at the UW, the founder of several technology startups, the CTO of the UW’s Global Innovation Exchange, and a certified genius.
They explained how clinicians in hospitals barely use a pupillometer — an expensive device found only in hospitals that measures pupil movement — and instead opt for a pen light that is less accurate but more convenient.
The doctors, also co-authors of the paper, wanted something in the middle — a tool easy and fast to use that provided reliable data. That’s when a UW team of computer scientists, electrical engineers and medical researchers got to work.
PupilScreen originally consisted of a smartphone embedded inside a 3D printed box to control the eye’s exposure to light. It was tested on a small pilot study that combined 48 results from both patients with traumatic brain injury and healthy people. The UW said it was able to “diagnose the brain injuries with almost perfect accuracy using the app’s output alone.”
Alex Mariakakis, a UW computer science doctoral student advised by Patel and a co-author of the published paper, told GeekWire that there is plenty of work to be done. Ultimately the technology won’t require the 3D box and only the smartphone. The team also plans to conduct larger-scale studies with coaches, emergency medical technicians, doctors and others to help refine the user experience and remove the possibility of human error when using the app to make important health-related decisions. Additional studies will also help aggregate more data that can improve the app’s algorithm accuracy through machine learning.
It’s early days, but PupilScreen seems to be on to something.
“There are lots of different reasons why we are excited about doing this kind of thing on a smartphone,” Mariakakis said.
A scientific paper from 2013 endorsed by the American College of Sports Medicine noted that 3.8 million concussions occur in the U.S. per year during “competitive sports and recreational activities,” though only 50 percent “may go reported.”
Some of those unreported injuries could be related to lack of resources on the sidelines, or because trained clinicians aren’t spotting something significant. There are also other ways teams assess concussions today, like the Sport Concussion Assessment Tool.
But the fact is, many athletes are at greater risk of head or brain injury without knowing it.
“Long story short, the status quo isn’t cutting it and people are missing concussions,” Mariakakis said.
The PupilScreen team is going public with their work in part to make connections that can help make their idea a reality.
“We definitely want to be on par with meeting standards of what clinicians look for,” Mariakakis noted.
An app like PupilScreen could make a big impact for a variety of use cases and particularly with football. The NFL, for example, had to shell out around $1 billion to former players who sued over how the league lied to them about potential head injuries suffered from big hits on the field. It may even be more important for younger players.
“Having an objective measure that a coach or parent or anyone on the sidelines of a game could use to screen for concussion would truly be a game-changer,” Patel said in a statement. “Right now the best screening protocols we have are still subjective, and a player who really wants to get back on the field can find ways to game the system.”
“PupilScreen aims to fill that gap by giving us the first capability to measure an objective biomarker of concussion in the field,” McGrath added in a statement. “After further testing, we think this device will empower everyone from Little League coaches to NFL doctors to emergency department physicians to rapidly detect and triage head injury.”
Other authors include Jacob Baudin, Eric Whitmire, Vardhman Mehta, and Megan A. Bank. The project was funded by the National Science Foundation, the Washington Research Foundation, and Amazon Catalyst.
A startup from Purdue University called brightlamp is working on something similar. It has a patent-pending technology that detects concussions by examining the eyes with a mobile device, though PupilScreen is first to publish data in scientific literature.
This also isn’t the first sports tech idea related to safety to start at the UW. More than three years ago, a small startup called Vicis spun out of the university to manufacture a high-tech football helmet that reduced the impact thought to cause concussions. Now NFL, college, and high school players are wearing a Vicis helmet for the first time during games this fall.
