Such studies find that the entire brain powers down on anesthesia, its activity dropping between 30 and 60 percent. The results are somewhat ambiguous, since brain regions respond differently to different drugs. But one region consistently becomes quieter than average: a grape-size cluster of neurons almost dead center in the brain known as the thalamus.

Is the thalamus the brain’s power switch? It certainly has the right stuff for the job. A thicket of neurons sprout from the thalamus and branch across the cortex, the outer layer of the brain where we interpret the information from our senses and make decisions, then back into the thalamus. As the brain’s sensory relay station, the thalamus is responsible for sending rousing signals to the cortex when we wake up from ordinary sleep. In 2007 Alkire and his collaborators probed the role of the thalamus by putting rats in a box flooded with anesthetics, which caused the animals to keel over. If Alkire and his colleagues then injected a tiny dose of nicotine into the thalamus, the rats immediately came to and stayed conscious even as they continued to inhale the anesthetics.

Yet studies on patients with Parkinson’s disease show that the thalamus cannot completely explain how anesthesia works. Surgeons can treat Parkinson’s by implanting electrodes deep inside the brain. These electrodes release pulses of current to tamp down the wild movements associated with the disease. Lionel Velly, an anesthesiologist at Mediterranean University in Marseille, France, ran an experiment in which he used the electrodes in the other direction, to record electrical activity in the brain.

In a second surgical procedure less than a week after the brain surgery, Velly and his colleagues took readings from the deep-brain electrodes in 25 patients while also collecting electrode readings from their scalp. The scalp recordings let the scientists monitor the cortex, while the deep-brain electrodes let them monitor the thalamus. Velly’s team found that the cortex started producing deep, slow waves as soon as patients became unresponsive. The thalamus, on the other hand, didn’t change for another 15 minutes. The pattern Velly saw was the reverse of what you would expect if the thalamus were the brain’s master switch.

The secret of anesthesia may lie not in any single clump of neurons but in the conversations taking place between many clumps in the brain.

Giulio Tononi, a University of Wisconsin neuroscientist, suggests that the secret of anesthesia may not in fact lie in any single clump of neurons. It may lie instead in the conversations that take place between many clumps in the brain. Normally information from our senses races from one region of the cortex to another, getting processed in different ways in each place. Some regions help us recognize faces in a scene, for example, while other regions help us figure out what emotions those faces are expressing. The sensory signals travel through a mass transit system made up of long branches of neurons that crisscross the brain. This system has a few hubs through which many connections pass. One is the thalamus, but certain parts of the cortex also serve as hubs.

Although the brain may become less active under anesthesia, it usually doesn’t shut down completely (if it did, we would die). In fact, when scientists played a tone into the ears of an anesthetized cat, its cortex still produced strong bursts of electricity. But its responses were different from those of a waking cat. In an anesthetized cat, the brain responds the same way to any sound, with a noisy crackle of neurons. In a waking cat, the response is complex: One brain region after another responds as the animal processes the sound, and different sounds produce different responses. It’s as if the waking brain produces a unique melody, whereas the anesthetized brain can produce only a blast of sound or no sound at all.

Tononi suggests that this change happens because anesthesia interferes with the brain’s mass transit system. Individual parts of the cortex can still respond to a stimulus. But the brain can’t move these signals around to other parts to create a single unified experience.

Tononi argues that the difference between brain music and brain noise defines the very nature of consciousness. Consciousness is the brain’s ability to be in a complex state, even in response to a simple stimulus like a tone. The vast number of different states our brains can enter when we are aware gives consciousness its marvelously rich feeling. In order to produce those states, the brain needs lots of neural elements that are active and able to respond, as well as the mass transit system that links them all together.

Working from this hypothesis, Tononi and his colleagues are trying to develop tools that can monitor levels of consciousness in anesthetized patients. They are also developing software to measure the complexity of the brain’s responses to stimuli. If Tononi’s idea is correct, anesthesiologists may be moving toward being able to gauge consciousness much as doctors gauge a patient’s temperature with a thermometer. Perhaps some of the mystery of consciousness itself—a question that has vexed philosophers for centuries—will be solved on the operating table.