Severe concussions where a person is visibly shaken or knocked unconscious are obviously cause for concern, but milder ones that go undetected can also lead to dangerous health impacts down the track. A new smartphone app could prove pivotal in this area, using a mobile device's camera to scan the pupil for telltale signs of injury on the spot.
We've seen a number of promising breakthroughs when it comes to detecting concussion, such as systems that track eye movements in search of symptoms and blood tests that can pick them up seven days after the incident. Knowing whether an athlete has suffered a concussion or not is important, because if left untreated it can not only lead to immediate consequences like headaches and memory loss, but further knocks to the head can pose even graver, long-term health risks.
So a smartphone app that can be used on the sideliness to immediately screen for concussion would be a game-changer, and that is exactly what researchers at the University of Washington (UW) have set out to develop. Called PupilScreen, the technology works by tracking the pupil's response to light, something known as the pupillary light reflex, which is a technique already used to gauge traumatic brain injuries.
While this pupillary light reflex has been used to assess serious brain injuries, recent research has shown that the technique can also reveal less severe ones, such as light concussions. So looking to take advantage of this new knowledge, the UW team has developed an app that uses the smartphone's camera, along with deep learning tools, to pick up changes in pupillary light reflex that are invisible to the naked eye.
The phone's flash is used to stimulate the eye and the camera than captures a three-second video. Deep learning algorithms then determine which pixels belong to the pupil in each frame, and in this way track changes in pupil size across the duration of the video. Putting the app to the test, the team tried it out in a pilot study involving 48 subjects, including both healthy patients and some with traumatic brain injuries. Using the app, clinicians were able to diagnose the brain injuries with almost perfect accuracy and almost as well as a pupilometer, an expensive machine found only in hospitals.
From here the team plans to conduct a broader clinical study in the coming months that will see the app put through its paces by coaches, medical technicians and doctors. They will also look to make improvements to the system, which currently requires the phone to be placed inside a box to control light exposure, and hope to produce a commercially available system within two years.
"The vision we're shooting for is having someone simply hold the phone up and use the flash," said lead author Alex Mariakakis. "We want every parent, coach, caregiver or EMT who is concerned about a brain injury to be able to use it on the spot without needing extra hardware."
The video below provides an overview of the technology, while the team will present its research the computing conference Ubicomp 2017 in Hawaii next week.
Source: University of Washington
