Voluntary movements are one of the brain’s main “outputs,” yet science still knows very little about how networks of neurons plan, initiate and execute them. Now, researchers from Columbia University and the Champalimaud Center for the Unknown in Lisbon, Portugal, say they have discovered an “activity map” that the brain uses to guide animals’ movements. The findings, published Wednesday in Neuron, could advance our understanding of how the brain learns new movements—and of what goes wrong in related disorders such as Parkinson's disease.

Movements are controlled and coordinated by multiple brain structures including the primary motor cortex. Located at the back of the frontal lobe, it contains cells whose long fibers extend down through the spinal cord, where they contact “secondary” motor neurons that signal the body muscles. A set of deep brain structures called the basal ganglia are also critical for movement, as evidenced by their degeneration in conditions such as Parkinson’s. One component of the basal ganglia, called the striatum, receives information about possible actions from the motor cortex and is thought to be involved in selecting, preparing and executing the appropriate commands before they are sent to the body. Earlier research had shown that signals leave the striatum along one of two distinct pathways: one that facilitates movement, and another that suppresses it. A number of more recent studies show that both pathways are active during motion, however, suggesting that they do not act by simply sending “stop” and “go” signals. And although it has long been suspected that different groups of neurons in the striatum represent distinct actions, exactly how they might do so has remained unclear. 

To investigate further, neuroscientist Rui Costa of the Zuckerman Institute at Columbia University and his colleagues created a strain of mice carrying a genetically encoded calcium sensor in neurons in the striatum. This sensor, a protein, emits fluorescent light in response to the increases in calcium ion concentrations that occur in cells when they become active. By combining this with a recently developed imaging technique called one-photon microendoscopy, the researchers were able to visualize the activity of up to 300 individual neurons in the striata of freely moving mice—using miniaturized microscopes attached to the animals' heads—and to capture the dynamics of the cells' firing patterns in time and space.  

“We found a 'local bias' in the cellular activity, [such that] neurons that are closer together are more likely to be active together,” says Costa, the senior author of the study. “In addition, we saw that many neurons are active specifically during one movement, while others are active during more than one movement, so there is some kind of map of actions.”

The researchers could predict what movements the animals were making based on which neurons lit up, Costa adds. “Similar actions had similar patterns of neuronal activity, and dissimilar actions had less-similar patterns,” he explains. “So our predictions weren't as accurate when we looked at the patterns for similar movements.”

The study authors noted that these activity patterns were independent of the animals’ speed of movement, as measured by accelerometers attached to their heads. This, Costa says, suggests that movement-related activity in the striatum is far more complex than we thought, and that the precise pattern of activity within the “stop” and “go” pathways is more important than overall levels of activity within each. 

“This is a great paper using complex methodology to help resolve a conceptually simple problem that has been a source of significant debate,” says John Reynolds, a professor of physiology at the University of Otago in New Zealand, who studies how the basal ganglia generate movement. “It takes us one step forward to resolving the conundrum of whether the two major classes of neurons in the striatum that independently either activate or inhibit movement exist in discrete functional clusters or coexist and work in concert during behavior.”

Joshua Dudman, a neuroscientist at the Janelia Research Campus of the Howard Hughes Medical Institute who studies the neural circuits controlling movement,  says the study is useful, but stresses its limitations. “We've known for several decades about such weakly clustered, functionally similar activity patterns in the striatum, so it's good to confirm that in freely moving animals,” Dudman says. “They [the researchers] seem to focus on evaluating a model that no one has proposed—that movement speed is encoded in the striatum independently of action identity.” But, he says, the researchers’ conclusions rest on “inferring what simultaneous measurements of [the entire 'stop' and 'go' cell] populations might look like, rather than making the important step of actually performing them.”*

Dudman adds that the study is nevertheless useful because it shows distinct populations of striatal neurons being active for a variety of behaviors. “In the past, such observations have primarily been made between a few, relatively similar actions, such as left/right turns,” he says, “so providing evidence across a range of behaviors is valuable. Other recent work has highlighted that the 'stop' and 'go' populations act in concert to control actions, and so this study further emphasizes the critical importance of simultaneous measurements across diverse behaviors as a goal for future work.”

Costa and his colleagues are now trying to decipher the logic of the activity patterns they observed in the striatum. Cells in the primary motor cortex—which sends movement commands to the muscles via the spinal cord—are organized such that cells controlling adjacent body parts and muscle groups are located next to each other, and Costa believes that this so-called “somatotopic” organization may contribute to the patterns in the striatum.

“We think there's something very similar going on [in the striatum], and we already have some preliminary data on that,” Costa says. “But turning right and left had very different patterns, although both involve the head, so it's not somatotopy alone. It could have something to do with the muscles, or a combination of both.”   

“We'd also like to know what happens to these action maps in diseases like Parkinson's,” he adds. “We predict they'll be greatly altered.”

*Editor's Note (9/12/17): The last sentence of this quote was removed after posting at the request of the speaker, who objected to the substance of how he was quoted.