Neurons may function more solo than thought

Neurons vote independently.

Brain cells hardly ever follow their neighbors in spouting electrically charged opinions despite being wired together, two studies published January 29 in Science suggest. This fierce independence could change the view of how the brain codes information.

Researchers led by neuroscientist Andreas Tolias of the Baylor College of Medicine in Houston measured electrical activity in pairs of neurons in the V1 visual processing centers of macaques. The researchers found that the neurons almost never synchronize their firing patterns with those of neighboring cells, even when the cells are very close together and detect the same sort of visual information. While previous studies have also detected low levels of coordination, even those studies showed 10 times higher levels of coordination in neuronal activity patterns than Tolias and his colleagues report.

“It’s possible that no two cells in the brain are doing exactly the same thing,” says Tolias, who is also affiliated with the Michael E. DeBakey Veterans Affairs Medical Center in Houston.

The degree of coordination Tolias’s team measures is surprisingly low. “He’s seeing something so close to zero that you could almost call it zero,” says Michael Shadlen, a neuroscientist and Howard Hughes Medical Institute investigator at the University of Washington in Seattle.

If that’s true, then the brain’s processing should be much more accurate than it actually is. Scientists think that low coordination among neurons actually boosts the brain’s computing power. In theory, if every neuron really acts on its own and the brain’s computing power is optimized, then people and animals should never make mistakes.

Scientists may need to look for other reasons to explain why movements and perception aren’t always precise, says Hendrikje Nienborg, a neuroscientist at the Salk Institute for Biological Studies in La Jolla, Calif. “It needs to be understood why the measurements are so different,” she says.

Synchrony garners high scores for pairs of figure skaters, but coordinated neuronal activity is thought to be a source of errors. It’s a matter of statistics, scientists say. The brain seems to average the activity of many neurons to create a picture of what is going on in the world. If each neuron offers an independent assessment (by sending out jolts of electricity at a particular rate that varies depending on the stimulus) then averaging many neurons’ activity gives the brain a fairly accurate picture.

But if many neurons fire signals in similar patterns, those neurons just repeat the same message, like pundits spouting talking points on cable news shows, and the neurons’ group assessment will be no more accurate than that of just one neuron.

That might be all right, but “Neurons are unreliable,” says Alfonso Renart, a neuroscientist at Rutgers University in Newark, N.J. Coordination could cause a slightly skewed impression to be overrepresented in the sample, leading to mistakes, he says. Indeed, scientists have blamed many perception mistakes that monkeys make on correlated neuronal activity. Too much synchrony in neuronal activity may lead to imprecise movements and other inaccuracies. Taken to an extreme, neuronal coordination can produce epileptic seizures.

One reason scientists have offered for occasional coordination in neurons is that the cells are wired together in vast networks with many neurons sharing common wiring partners. It has been thought that when several cells receive the same signal, that shared input can make all the neurons act similarly.

But, Renart says, “We believe firmly that [correlations] are not inevitable.”

In the second paper appearing in Science, Renart and his colleagues show that neurons can remain independent despite many shared connections. The team measured spontaneous activity of neurons in the brains of anesthetized rats and, like Tolias, found little coordination between neurons despite the fact that the neurons share many of the same sources of input.

The researchers then created a mathematical model to explain how neurons fight peer pressure. Common inputs might excite neurons to coordinate firing patterns, but messages from inhibitory neurons cancel out the excitatory influences, allowing each neuron to have its own voice, the team found.

“I totally embrace the ideas in the Renart paper,” says Shadlen. “It’s beautiful.” While Renart and his colleagues may have ruled out common input as a source of coordinated activity, they haven’t yet explained what does lead to the synchronicity between two neurons that has been seen in previous experiments, he says.