AI Spots Hidden Signs of Consciousness in Comatose Patients before Doctors Do

AI Spots Hidden Signs of Consciousness in Comatose Patients before Doctors Do

By Andrew Chapman edited by Allison Parshall

Imagine lying in a hospital bed, awake but unable to move your body to communicate with the people around you. This experience of “covert consciousness” is a reality for many people who have sustained traumatic brain injuries. In a new study published in Communications Medicine, researchers found that they could detect signs of consciousness in comatose patients by using artificial intelligence to analyze facial movements that were too small to be noticed by clinicians.

Covert consciousness was first detected in 2006, when researchers asked an unresponsive woman and healthy volunteers to imagine doing specific tasks while in a brain scanner. The team found that the woman showed brain activity in the same regions as the volunteers. Just last year researchers using similar brain imaging methods found that one in four behaviorally unresponsive patients was covertly conscious. Such tests aren’t routinely performed on people in an unresponsive state because this type of neuroimaging is time-consuming and its operation requires specialized skills. Instead, doctors typically rely on more subjective visual examinations to gauge a person’s level of consciousness, testing whether they open their eyes, respond to commands or startle at a loud noise.

“We were trying to find a way to quantify how conscious these patients are” using simple and readily available technology, says Sima Mofakham, a computational neuroscientist at Stony Brook University and senior author of the new study.

On supporting science journalism

If you’re enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.

Mofakham and her team recorded videos of 37 patients with recent brain injuries who outwardly appeared to be in a coma. They tracked the participants’ facial movements in extraordinary detail—on the level of individual pores—after they were given commands such as “open your eyes” or “stick out your tongue.” The researchers’ AI tracking tool, which they call SeeMe, detected any facial movements and then analyzed whether the movements were specific to the command given. SeeMe documented eye-opening responses in 30 of 36 patients and mouth movements in 16 of 17 patients with analyzable videos. Five of these patients did not go on to produce larger signs of movement that were visible to doctors, though most of the others did. On average, SeeMee detected participants trying to open their eyes and moving their mouths respectively 4.1 and 8.3 days before clinicians spotted these signs.

“What we found was: patients develop [small] movements before going to more obvious movements,” Mofakham says. The results suggest that, in some cases, people are conscious days before doctors notice. Patients with larger and more frequent facial movements also had better clinical outcomes, indicating that the technology may help predict prognoses.

This ability to detect consciousness earlier is clinically meaningful, says Jan Claassen, a neurologist at Columbia University, who wasn’t involved in the new research. Signs of consciousness can provide another layer of information for doctors and family members choosing between a range of treatments, from palliative care to more aggressive therapies. “Every day is potentially important” for those difficult decisions, Claassen says. Earlier detection could also allow care teams to start rehabilitation programs used to improve patients’ motor skills sooner. Separate research shows that starting rehab earlier is associated with greater improvements in physical function.

Consciousness recovery after a brain injury is often gradual and unpredictable. “When somebody recovers consciousness, it’s almost like a flickering light bulb,” Claassen says. “It doesn’t just come on or off.” The new study only followed participants until six months after they were discharged from the hospital. But it’s possible that some patients whose conditions have remained more static and who are currently presumed unconscious in long-term care facilities might also show signs of awareness that could be detected, whether with more advanced neuroimaging techniques or the simpler SeeMe technology. “We have to do the experiments and see,” Mofakham says. “There is a chance.”

Next, Mofakham plans to examine whether patients can answer yes-or-no questions using specific facial movements. “This has a big ethical implication” because people who cannot communicate “cannot participate in their care,” she says. “This study opens a way to communicate with these patients.”

It’s Time to Stand Up for Science

If you enjoyed this article, I’d like to ask for your support. Scientific American has served as an advocate for science and industry for 180 years, and right now may be the most critical moment in that two-century history.

I’ve been a Scientific American subscriber since I was 12 years old, and it helped shape the way I look at the world. SciAm always educates and delights me, and inspires a sense of awe for our vast, beautiful universe. I hope it does that for you, too.

If you subscribe to Scientific American, you help ensure that our coverage is centered on meaningful research and discovery; that we have the resources to report on the decisions that threaten labs across the U.S.; and that we support both budding and working scientists at a time when the value of science itself too often goes unrecognized.

In return, you get essential news, captivating podcasts, brilliant infographics, can’t-miss newsletters, must-watch videos, challenging games, and the science world’s best writing and reporting. You can even gift someone a subscription.

There has never been a more important time for us to stand up and show why science matters. I hope you’ll support us in that mission.