An unconscious person’s response to odors after a serious brain injury may be a simple yet powerful signal of how aware they are and how likely they are to survive and recover, a new study suggests, relying on responses to the scent of shampoo and the stench of rotting fish.

Patients who survive brain damage from trauma, stroke, or heart attack are plunged into forms of unconsciousness that vary from minimal consciousness to unresponsive wakefulness, sometimes called a vegetative state. Specialists trying to tell who is in which state have fared only a little better than a coin flip: About 4 in 10 people thought to be unconscious are actually aware.

That uncertainty makes decisions for families and clinicians supremely difficult, from weighing how to treat pain to whether to withdraw life support. Sophisticated imaging of unconscious patients’ brain activity can reveal hints of awareness that go beyond behavioral assessments, sometimes only to deepen the mystery of who will get better. Now Israeli scientists have turned to the sense of smell, evolutionarily speaking our most ancient sensory system, as a window into our brain. Their paper appears Wednesday in Nature.

It’s an idea so simple, study co-author and neurobiologist Noam Sobel of the Weizmann Institute said, it can be explained in a few sentences: If you don’t know if two people are conscious, give them an odor. If they sniff it, they’re conscious; if they don’t, they’re not. What’s more, if they do sniff it, they’ll probably live for at least three years. And if they don’t sniff it, then their chances of surviving are much lower.

“I think this was a very cleverly designed study that does add another tool to the consciousness-detection toolbox,” said Joseph Giacino of Harvard Medical School and Spaulding Rehabilitation Hospital. He was not involved in the study but he did write the American Academy of Neurology’s most recent guideline on managing people in a vegetative or minimally conscious state caused by brain injury.

Sobel and his team studied 43 brain-injured patients at a rehabilitation hospital for six years, using a small device attached to the nose to measure how much air they inhaled in response to jars containing a pleasant odor (shampoo), an unpleasant odor (rotting fish), or no odor.

It’s a technique the scientists had used before to ask people who could not tell them in words what they thought about an odor, either because they were asleep or they were nonverbal children with autism. A big sniff means the odor was mild. A short, abrupt sniff means the odor was unpleasant. Both responses are part of a sensory-motor feedback loop that is a sign of arousal and consciousness.

The scientists tested the patients repeatedly because consciousness can fluctuate after a brain injury. They noted whether a smell was detected at all, whether the patient could tell if it was pleasant or unpleasant, and whether the patient expected to smell something good or bad based on a previous odor.

Each patient’s consciousness was also gauged by standard behavioral tests immediately after the sniff test. All patients who sniffed eventually showed signs of consciousness, and all patients with unresponsive wakefulness syndrome (as assessed using standard measures), and who remained unconscious, did not have a sniff response. The test was accurate 100% of the time when patients did sniff, but it was accurate only 65% of the time when patients did not sniff, meaning 35% of patients did regain consciousness even though they did not sniff.

In one case, the sniff test indicated a patient was regaining consciousness two months before any other measure showed that, Sobel said.

“It seems like this approach is significantly more effective and potentially easier to implement in the long run than standard behavioral measures that are out there,” he said. “We’re providing valuable information.”

Giacino agrees that a bedside tool is very much needed, especially in hospitals not equipped to perform and interpret functional imaging studies. Like other promising proof-of-principle studies, this one needs to be replicated at multiple centers, he said, and compared to a single validated protocol, not the three different ones — including his own guideline — that were yardsticks in the Nature study. In his view, the tool is well-suited for diagnosis but it doesn’t offer the kind of detailed prognosis families want and the field has so far been unable to deliver.

“What we want to do better is to be able to say to the family what the likelihood is that [the patient] won’t be able to take care of themselves. Or be able to return to work and speak with you conversationally,” he said. “This test doesn’t help there.”

The sniff test could be most useful immediately after a brain injury, Giacino said, when the patient is in the intensive care unit and imaging looks so devastating that it leads families and physicians to withdraw life support, even though data have shown 1 in 5 patients will regain some measure of independent function in five years. The sniff test could slow down that rush to judgment.

“If we have three different tools and each one brings something a little bit unique to the problem, we put them all together and then we’re going to have a better chance of avoiding missing consciousness when it’s present,” he said. “We have our standardized behavioral assessment tool. We have our neuroimaging, and now maybe we have a sniff test. I think this gives us some more room for optimism that we can reduce this terribly high misdiagnosis rate.”

Jacobo Sitt, a researcher at the French Institute of Health and Medical Research and the Sorbonne who was not involved in the research, also sees a complementary role for the test, should larger studies back it up. “Followup work combining olfactory information together with cognitive tasks and neuroimaging will help unveil if these olfactory responses reflect conscious processing in these patients,” he said.

Neuroimaging has delivered some startling insights in the past. In a paper published in 2010 in the New England Journal of Medicine, Steven Laureys, a neurologist who leads the Coma Science Group at the University of Liege in Belgium, reported not only detecting awareness in a small number of unconscious patients who were prompted to imagine playing tennis or walking through their homes, but also inducing them to think of one activity or the other as a response to other questions. Since then, his team has also explored how unconscious patients react to odors, something that families have often noticed.

“I do think it is convincing. I do think it also makes sense based on the neuroanatomy,” Laureys said about the new study.

Sobel has spent his career studying olfaction and neuroanatomy.

“Your sense of smell is directly linked to your respiratory system in the deepest parts of your brain,” Sobel said. “You know, inhalation is the first and last thing you do.”