The famously distorted motor homunculus cartoon, popularized in the 1950s, suggests that the motor cortex sports an orderly organization, with distinct regions controlling single body parts from the foot to the face. But even Wilder Penfield, one of the neurosurgeons who created the homunculus after briefly electrically stimulating each part of the cortex in awake participants, noted that some regions’ functions overlapped.
This intermixing in the human motor cortex is seen at the level of single neurons, too, according to a study published today in Nature. Neurons along a region called the crown of the precentral gyrus, long thought to house the primary motor cortex, not only represent specific areas of the body that match the motor homunculus, but also carry information about many different body parts, the study shows.
The motor cortex is “a blurry map of the body,” says Michael Graziano, professor of neuroscience and psychology at Princeton University, who was not involved with the work.
Instead of being divided into regions dedicated to the arms, legs, trunk and so on, the precentral gyrus is organized into functional zones dedicated to speech and to facial and whole-body movements more broadly, the researchers found by using data from brain-computer interfaces that had been implanted in eight people for clinical reasons.
The findings uphold the idea of a motor cortex that may have some specialization, “but it’s still integrating parts of the body in a way that probably promotes meaningful, ethologically relevant behavior patterns,” Graziano says. “The brain evolves and learns over a lifetime to produce actual meaningful movement. Its goal is not to control separate muscle movements.”
In addition, because the crown of the precentral gyrus encodes more abstract information about movement, this brain area may not really be a primary motor area, says Darrel Deo, who conducted the work as a postdoctoral scholar in the lab of Jaimie Henderson and Francis Willet at Stanford University’s Wu Tsai Neurosciences Institute and is now instructor of neurosurgery at Stanford. “We, as a lab, consider it more of a premotor cortex.”
T
For example, using long pulses of current to stimulate a motor cortex region in monkeys results in complex motor movements, earlier findings from Graziano show. And functional MRI studies in people suggest that the information for different body parts is not limited to one region and that regions can connect strongly with other areas of the brain for goal-driven action planning.
But techniques such as fMRI “still measure brain activity at the scale of tens of thousands or hundreds of thousands of neurons, and they haven’t been able to resolve what happens at the level of single neurons,” Deo says.
Deo and his colleagues found that they could accurately decode movements from the neural activity recorded by different electrodes in the implanted brain-computer interfaces. There were “whole-body representations everywhere that we sampled along the [precentral gyrus],” Deo says. “When you zoom in, then you’re around this intermixed jumble of tiles.”
Although previous research had already shown overlap and intermixing, “it’s great to have it at this resolution,” says Tamar Makin, professor of cognitive neuroscience at University of Cambridge, who was not involved with the new work.
The ventral side of the precentral gyrus has an arm-preferring area and a broadly-tuned face and mouth area, flanked by two regions tuned to speech, the study shows. “This was really exciting, because this was not necessarily captured by the original homunculus model,” Deo says.
It makes sense that the human motor cortex would evolve not one but two functional regions to enable speech, Graziano says. “Speech is obviously the quintessential ethological action that humans produce.”
There was also functional linking across limbs, the researchers found. Homologous movements by different limbs, such as wrist and ankle flexion, share similar neural activity patterns, the study shows, which suggests that “there’s something shared in the neural code,” Makin says.
T
Makin agrees. “This is not primary motor cortex.”
The participants all have limited mobility—four have spinal cord injury, three have ALS and one is paralyzed due to a stroke—so it’s still unclear whether these findings would generalize to the general population, Deo says.
But seeing the same results across eight different people is convincing, Graziano says. “Lots of different conditions, one consistent result—and that makes me think this result is real,” he says.
At the same time, just because movement is decodable from neural activity does not mean that it is relevant to the behavior, Makin adds. Next, “we need to better understand why the information is there,” she says. “What makes it go away, and what makes it change? Is it something that you can learn to adapt, or is it something that is innate?”
