Office of Research & Development
Office of Research & Development
VA Research Currents archive
May 19, 2014
Dr. Uzma Samadani is developing protocols to detect mild brain injury based on eye tracking. (Photo by Lamel Hinton)
What do a music video by Colombian pop star Shakira and a 3,500-year-old Egyptian medical scroll have in common? They both relate to Dr. Uzma Samadani's research on something called dysconjugate gaze—the inability to move both eyes in the same direction at the same time, which often happens following brain trauma.
Think of Wile E. Coyote, of cartoon fame, shortly after an anvil drops on his head. Whenever a person experiences head trauma, the eyes tend to exhibit a dysconjugate gaze. The folks at Warner Brothers knew this in the 1940s and 1950s and used it to depict concussion in the oft-injured Coyote. The Egyptians knew it in 1,500 B.C.E., when they wrote what is now commonly known as the Edwin Smith Papyrus, which describes a deviation of the eye in a person with a head injury.
"It's been really widely accepted for a long time," says Samadani, chief neurosurgeon at the New York Harbor VA Healthcare System. "For centuries before radiographic imaging, brain trauma was identified by examining eye movements. It was a serious art and there were entire medical textbooks—such as Fred Plum and Jerome Posner's Diagnosis of Stupor and Coma—with multiple chapters about eye movements."
Newer technology such as CT scans and magnetic resonance imaging (MRI) later became the gold standard diagnostic tools for brain injury, but optometrists and ophthalmologists have known for years that eye movements can still provide valuable information, as did their ancient medical counterparts.
Now, with the widespread availability of commercial eye-tracking cameras, experts are taking a new look at this time-honored approach to TBI diagnosis.
In 2010, Samadani and her colleagues developed an algorithm to track the eye movements of people while they were watching television. "We were doing a clinical trial involving patients with very severe brain injury and were trying to identify measures of recovery. When you have a patient in a vegetative state it can be very difficult to determine if they're improving or not," says Samadani. "We decided to use their ability to watch TV as a potential outcome measure."
Researchers played videos on a computer monitor and simultaneously tracked the eye movements of the patients.
"We were hoping to get a percentage of time that they were watching—something quantifiable," says Samadani. "We were very surprised to find that swelling in the brain, or really any trauma at all, led to abnormal patterns of eye movements that we could detect and quantitate with our algorithm."
Then came Hurricane Sandy in 2012. The VA hospital was shut down for seven months. Samadani was forced to take her research to Bellevue Hospital in south Manhattan.
When she started up again, Samadani began looking at patients with relatively mild concussions. These are typically missed on CT scans because there is not the type of structural damage CT scans detect.
"We looked at three groups of people," says Samadani. "Some had a CT scan that showed a structural brain injury. Some experienced head injuries but had negative CT results, and a third group was trauma patients with no evidence of brain trauma."
Not only did eye tracking identify the patients with known brain trauma, it also identified milder concussions the CT scans had missed. Samadani presented the findings at the 82nd annual American Association of Neurological Surgeons meeting in April 2014.
In the study, patients were asked to watch 220 seconds of video while their eye movements were monitored. One of the first clips selected by the researchers? Shakira's music video for the 2010 FIFA World Cup.
"We just wanted something people would pay attention to," says Samadani. "We used cartoons, highlights, sports videos—anything eye-catching and not too boring."
As it turns out, Shakira was one of the most popular selections. Samadani says she has had excellent results showing sports highlights and Disney cartoons.
When the results came back—when Samadani realized how accurately she could diagnose brain injuries by tracking eye movements—she barely slept for two months, she recalls. She realized then that "this has the potential to be as important for diagnosing concussions and blast injuries as EEG was for seizures."
She is actively working with collaborators at other institutions to verify her results, which remain preliminary until published in a peer-reviewed journal.
She imagines the technology eventually being used in the field, on soldiers who may have concussions or been exposed to blast. "These are the people the technology will help the most," suggests Samadani, "the people who know they are not the same after trauma, but whose MRIs and CTs are normal. The ones who just want to be believed."
"We want to make the technology as universally accessible as possible—get it onto a tablet or iPad," she says, "but right now the portable cameras on those devices don't yet have the resolution we need. It's a matter of waiting for the technology to catch up," something Samadani says is only a matter of time, given the development of eye tracking technology for video games and other big markets.
"The disruption of eye movements as indicative of brain injury was described in the oldest known surgical paper," concludes Samadani. "It's been known for so long and now we've developed an easy way to quantify it. If you can quantify the damage associated with radiographically invisible brain injuries such as concussion and military blast, then you can treat these conditions and most importantly, you can figure out how to prevent them. That will be the greatest value of our technology —a chance that it will keep people from suffering from an otherwise invisible injury."