The prospect of undergoing surgery while not fully "under" may sound like the stuff of horror movies. But one patient in a thousand remembers moments of awareness while under general anesthesia, physicians estimate. The memories are sometimes neutral images or sounds of the operating room, but occasionally patients report being fully aware of pain, terror, and immobility. Though surgeons scrupulously monitor vital signs such as pulse and blood pressure, anesthesiologists have no clear signal of whether the patient is conscious. But a new study finds that the brain may produce an early-warning signal that consciousness is returning—one that's detectable by electroencephalography (EEG), the recording of neural activity via electrodes on the skull.
"We've known since the 1930s that brain activity changes dramatically with increasing doses of anesthetic," says the study's corresponding author, anesthesiologist Patrick Purdon of Massachusetts General Hospital in Boston. "But monitoring a patient's brain with EEG has never become routine practice."
Beginning in the 1990s, some anesthesiologists began using an approach called the bispectral (BIS) index, in which readings from a single electrode are connected to a device that calculates, and displays, a single number indicating where the patient's brain activity falls on a scale of 100 (fully conscious) to zero (a "flatline" EEG). Anything between 40 and 60 is considered the target range for unconsciousness. But this index and other similar ones are only indirect measurements, Purdon explains. In 2011, a team led by anesthesiologist Michael Avidan at the Washington University School of Medicine in St. Louis, Missouri, found that monitoring with the BIS index was slightly less successful at preventing awareness during surgery than the nonbrain-based method of measuring exhaled anesthesia in the patient's breath. Of the 2861 patients monitored with the BIS index, seven had memories of the surgery, whereas only two of 2852 patients whose breath was analyzed remembered anything.
Despite that, Purdon and his co-workers were hopeful that an "unconsciousness signature" in the brain could be found. Last year, the team worked with three epilepsy patients who'd had electrodes implanted in their brains in preparation for surgery to reduce their seizures. Recording from single neurons in the cortex, where awareness is thought to reside, the researchers gave the patients an injection of the anesthetic propofol. They asked the volunteers to press a button whenever they heard a tone, recording the activity of the neurons. Loss of consciousness, defined as the point when the patients stopped pressing the button, was immediate—40 seconds after injection. Just as immediately, groups of neurons began to emit a characteristic slow oscillation, a kind of ripple in the cells' electrical field. The neurons weren't entirely inactive, but bursts of activity occurred only at specific points in this oscillation, resulting in inconsistent brain cell activity.
The next step was to see if the same signature could be detected externally, with an EEG. Purdon and his team recruited 10 healthy volunteers to "go under" extremely slowly with propofol: The anesthetic was delivered so gradually that the dropping-off process took not 40 seconds, but almost an hour. Every 4 seconds, the volunteers pressed a button in response to clicks or to words including their names, until they reached unconsciousness.
At that point, the researchers report today in the Proceedings of the National Academy of Sciences, EEG readings showed activity analogous to that seen in the study of epilepsy patients. Alpha waves, associated with relaxation and drowsiness, increased with loss of consciousness, as did the even slower, "low-frequency waves." Both patterns of activity began to decrease with returning consciousness.
The researchers also found a pattern unique to the transition period. During the transition in and out of unconsciousness, the waves almost cancel each other out: The high point or "peak" of the alpha waves occurred at the "trough" of the low-frequency waves. This conjunction, called the "trough-max" pattern, can be read on an EEG and may be the early warning signal that the patient is returning to consciousness. When the patient is deeply unconscious, the "peak-max" pattern, in which the high points of the two wave types occur together, may prove to be a reliable sign that the patient is out.
"It's a rigorous, elegant study," Avidan says. "The trough-max pattern may well prove to be an early warning signal." He cautions, however, that the volunteers were healthy and not undergoing any actual surgery, so further research in surgical patients is needed to confirm the findings.