Your brain is a bit like a concert hall. To drive our cognitive processes, several groups of neurons need to become active—and, like the various sections of an orchestra, work in harmony to produce the symphony of computations that allow us to perceive and interact with our surroundings.

As with an orchestra, the brain likely requires a conductor to keep all its active parts in sync. There are neuroscientists who think that gamma rhythms, fast brain waves that fluctuate at a frequency of approximately 40 cycles per second, play this role. By ticking at regular intervals, these oscillations are thought to act like a clock that coordinates information transfer from one group of neurons to another. There is ample evidence suggesting that gamma waves are important for the brain's computations: decades of studies in humans and other animals have found these patterns in many parts of brain and have associated them with a range of cognitive processes, such as attention and the mental scratchpad of working memory. Some studies have even linked disturbances in these oscillations to various neurological diseases, including schizophrenia and Alzheimer's.

But a consensus does not exist. Some neuroscientists think that these gamma waves may not do much at all. Rather than a relevant physiological signal, one camp believes that these rhythms are simply “an exhaust fume of computation,” says Chris Moore, a neuroscientist at the Carney Institute for Brain Science at Brown University. In the same way your car releases emissions each time you drive it—the gamma signal could be perfectly correlated with brain activity, but not provide any meaningful contribution to the actual function of the car, he explains.

This viewpoint is also supported by scientific evidence. Researchers have reported the exact frequency of gamma oscillations can depend on the strength of a stimulus. That runs counter to the idea that these waves are a timekeeper or conductor, Moore says, “because the conductor shouldn't change the rhythm depending on how loud the music is.”

This seesawing may have tipped in favor of gamma in a study published July 18 in Neuron. Moore and a doctoral student in his lab, Hyeyoung Shin, reported their discovery of a previously unknown set of neurons that act like metronomes in mice. These “metronome” cells fire rhythmically at gamma frequencies (30 to 55 cycles per second) regardless of what's happening in the outside world—and the probability of an animal detecting a sensory stimulus is associated with the cells' ability to keep time.

Moore and Shin began their investigation as a general search for brain activity linked to perceiving touch. To do so, they implanted electrodes into the barrel cortex in the somatosensory area of the brain, which processes inputs from the senses, in mice. The duo then measured neural activity while observing the rodents' ability to notice subtle taps on their whiskers.

The team homed in on gamma oscillations, and they decided to look specifically at a group of brain cells called fast-spiking interneurons because prior studies from Moore's group and others had suggested that they may be involved in generating these fast rhythms. That analysis revealed that, as expected, the degree to which those cells were firing at gamma frequencies predicted how well the mice would be able to detect touch.

But when the pair dug a bit deeper, they found something strange. Initially, they had expected cells that would be activated in response to a sensory stimulus to show the strongest links to perceptual accuracy—but when they examined those cells, this association grew weaker. “Then we realized, maybe it's the non-sensory ones [that matter],” Moore says. “I remember that moment distinctly, because it made a lot of sense.” After all, he adds, if these cells act as timekeepers, you want them to remain consistent, regardless of what's happening in the surroundings.

They re-ran the analysis with only the cells that didn't respond to a sensory stimulus—and, lo and behold, the link with perceptual success grew stronger. In addition to remaining unperturbed by the outside world, this specific subset of cells tended to spike regularly at gamma-range intervals, like a metronome. In addition, the more rhythmic these cells were, the better the animals seemed to be at detecting whisker taps. If you go back to the analogy of the brain as a concert hall, what appears to be happening, according to Shin, is that “the better the conductor is at keeping time, the better the orchestra will do.”

Vikaas Sohal, a neuroscientist at the University of California, San Francisco, who did not take part in the study, says that this paper “changes the way we think about brain rhythms.” The field is based largely on observations originally made recoding electrical signals from the scalp, which represent an amalgamation of many neurons in various parts of the brain, but “here we find a different type of brain rhythm that is discerned not by looking at electrical signals driven by many cells but rather by signals in a specific type of cell,” Sohal writes in an email.

There have been previous reports of neurons that spike at gamma intervals. Back in 1996, scientists reported so-called chattering cells, which spontaneously produce bursts at gamma frequencies, in the brains of anaesthetized cats. “The idea that there are these cells that generate this organized temporal rhythm by themselves has been out there for a while,” says Jess Cardin, a neuroscientist at Yale University and a former postdoc in Moore's lab who was not involved in the study. “This is the first time someone has identified a set of self-generating gamma neurons and linked them to behavior.”

But how these newly discovered cells coordinate neural activity remains an open question. Supratim Ray, a neuroscientist at the Indian Institute of Science, says that one puzzling result from this study was that, when the authors looked at the local field potential—the summed activity of a population of neurons—they were unable to find an increase in gamma oscillations linked to the activity in the metronome cells. One possible explanation for this finding, according to Moore, is that these gamma-generating cells are only influencing a small subset of neurons that are important for the brain's computations—and thus not showing up in the population signal. But if that is the case, Ray says that the clock might not be very useful. “If everyone wears a different watch and they're not synchronized, then there's no point of having a common time.”

So the jury is still out. When it comes to whether gamma oscillations play a functional role, these findings are another piece of evidence that pushes toward that idea, Cardin says. But like many other studies in this field, the experiments could only identify a correlation between gamma rhythms and behavior, not a causal one, she adds. “This is an area where the smoking gun experiment is really hard—and currently impossible—to do.”