Consciousness isn’t easy to define, but we know it when we experience it. It’s not so simple to decide when someone else is conscious, however, as doctors must sometimes do with patients who have suffered traumatic brain injury. Now, researchers have come up with an approach that uses the brain’s response to magnetic stimulation to judge a person’s awareness, reducing it to a numerical score they call an index of consciousness. “You’re kind of banging on the brain and listening to the echo,” says Anil Seth, a neuroscientist at the Sackler Centre for Consciousness Science at the University of Sussex in the United Kingdom who was not involved in the work.
Faced with an unresponsive patient, clinicians do their best to determine whether the person is conscious. Through sound, touch, and other stimuli, they try to provoke verbal responses, slight finger movements, or just a shifting gaze. Yet some conscious patients simply can’t move or speak; an estimated 40% of those initially judged to be completely unaware are later found to have some level of consciousness.
Recently, physicians seeking to resolve a patient’s conscious state have gone right to the source, searching for signs of awareness using brain imaging or recording electrical activity of neurons. Most of these approaches define a conscious brain as an integrated brain, where groups of cells in many different regions activate to form a cohesive pattern, explains Marcello Massimini, a neurophysiologist at the University of Milan in Italy. “But that’s not enough,” he says. Sometimes even an unconscious brain looks highly integrated. For example, stimulating the brain of a sleeping person can create a huge wave of activity that “propagates like a ripple in water.” It’s a highly synchronized, widespread pattern, but it’s not consciousness, he says, and so this measure is often unreliable for diagnosis.
Recently, Massimini and colleagues have been exploring another possible criterion for consciousness. To create our information-rich, moment-to-moment experience, different groups of neurons must have their own timing, or their own unique firing patterns. They must work together, but keep their individuality. “It’s quite a slippery thing to [describe] verbally,” Seth says. The big question for neuroscientists developing computerized tests of consciousness, he says, is “can we make that quite slippery verbal intuition mathematically precise?”
That’s just what Massimini and his colleagues have tried to do with the perturbational complexity index (PCI), which they describe online today in Science Translational Medicine. PCI looks at the brain’s response to transcranial magnetic stimulation (TMS), where a magnetic coil is held up to the surface of the skull, generating a pulse to stimulate the neurons beneath and provoke a response that radiates through the brain.
Massimini and his colleagues record the brain’s “echo” with electroencephalography, a measure of electrical activity, and then turn that data into a numerical score between 0 and 1. Their calculations award higher scores to information-rich responses: those that are distributed across the brain, but also individualized. If two distant groups of neurons are active, but their activity is synchronized, the PCI equation compresses them, meaning they contribute less to the overall score. “The less we can compress the pattern, the more information is in it,” Massimini explains.
The researchers first calibrated their system using healthy patients. They recorded TMS responses in waking subjects, and then used the brain activity from people in deep sleep or under different types of anesthesia as a reference for unconsciousness. There was no overlap in scores between conscious and unconscious subjects, suggesting a potentially useful cutoff or threshold for consciousness, somewhere between the highest unconscious score (0.31) and the lowest conscious score (0.44).
Then the researchers tested the index with 20 people who had suffered different types of brain damage. Those who were believed to be in a vegetative state—awake but completely unconscious—got very low scores (between 0.19 to 0.31). Subjects who had emerged from a coma had varying degrees of awareness and intermediate scores. Two of the patients had a condition known as locked-in syndrome: Their cognitive abilities were normal, but they were unable to move. These patients, who could communicate by shifting their eyes, received PCI scores of 0.51 and 0.62—as high as the waking, healthy subjects. Without requiring any active participation from subjects, this index can reliably place them on a continuum between conscious and unconscious, the team concludes.
“This is a very well-conceived approach,” says Joseph Giacino, a neuropsychologist at Spaulding Rehabilitation Hospital in Boston. “It’s based on what we understand the nature of consciousness to be … to the extent that we can understand it at this point.”
Seth calls the work “a substantial contribution to the state of the art.” The attempt to set a threshold for consciousness with our limited definition is much like early attempts to establish freezing and boiling points without a thermometer or a good definition of temperature, he says. PCI is the most reliable test to date, and is “a first step” toward an absolute scale for consciousness, Seth says.
Giacino points out that there are still gray areas on the PCI scale. There was some overlap between the range of scores for patients considered to be in a “minimally conscious state,” showing intermittent signs of awareness, such as following simple movement commands, and those who had emerged from this state but were still disoriented and had limited communication. He says this distinction is important when doctors decide whether trying simple communication methods, such as indicating “yes” and “no” with the eyes, will be worthwhile or simply overwhelm the patient. Still, if the results can be replicated in a larger group, Giacino predicts the test could be a vast improvement over current tools for locating consciousness in the midst of damage.