The rat stood on its hind limbs at one end of a narrow runway. It wore a tiny black vest attached to a robotic arm that hovered above its head. Without such mechanical support, the rat would have fallen over—its spinal cord had two deep cuts, rendering its back legs useless. Rubia van den Brand, then a doctoral candidate at the University of Zurich, stood at the other end of the runway, urging the animal to walk. Although the robotic arm kept the rat upright, it could not help the creature move; if the rodent were ever to walk again, it would have to will its feet forward. For the first time since van den Brand began her experiments, the rat moved one of its back legs on its own—a small, effortful step. She ran to her boss's office with the news and a crowd immediately gathered in the lab to watch what many had deemed impossible.

Van den Brand and Grégoire Courtine, now at the École Polytechnique Fédérale de Lausanne (E.P.F.L.), along with their colleagues, have trained rats with nearly severed spinal cords to walk again. One week after being injured, the rats could not move their hind limbs at all. Six weeks later they could walk, run, climb stairs and even sprint—but only with the support of the robotic arm accompanied by electrical and chemical stimulation of the spinal cord. Rats that trained on a moving treadmill instead of on a stationary runway moved their feet reflexively but never learned to walk voluntarily. Only conscious participation in walking encouraged new connections between the rodents' brains, spinal cords and limbs, which they needed to take those first deliberate steps. "It's kind of like how a toddler learns to walk," Courtine says. "Their spinal cord is full of activity and the brain needs to learn to take control of the spinal cord. As long as the brain has something to control it can learn progressively to communicate again with these cells."
To mimic the kind of severe spinal cord injury that paralyzes many people, the researchers anesthetized the animals and sliced deep into two spots on either side of the rats' spinal cords, severing many neural connections running from the brain down the back, but sparing a small bridge of tissue between the lesions. As an analogy, imagine carving two notches out of opposite sides of a thick rope, dividing many of the individual strings in the bundle, but not quite cutting the rope in half.

After a week's recovery time the rats could not budge their hind limbs—they only dragged themselves around by their front legs. Van den Brand and her colleagues settled on the vest-and-robotic-arm apparatus as a way to force the rats to use their back legs. First, however, the researchers began to wake up dormant neurons below the lesions in the rats' spinal cords—cells that had been cut off from the upper spinal cord and brain. Courtine and his colleagues stimulated the lower spinal cord with electrodes and a cocktail of chemicals that act as neurotransmitters, molecules neurons use to communicate with one another. With the assistance of such stimulation—and the robotic arm—rats placed on their hind legs on a treadmill went through the motions of walking. But none of the movement was voluntary. The spinal cord can orchestrate most of the motions involved in walking without help from the brain, as long as it gets sensory input from the environment—such as the ceaseless motion of the treadmill. Courtine and his team wanted to restore conscious, deliberate movement to the rats.

The researchers divided the rats into three groups: One group learned to walk on a treadmill; another learned to walk along a stationary runway and a set of steps; and a third received no training. Rats in each of the first two groups trained for about 30 minutes every day for several weeks as they received electric and chemical stimulation. All the while researchers provided encouragement, either by cheering or offering some chocolate. Two to three weeks later, the rodents training on the runway took their first voluntary steps. By six weeks the same rats had learned to sprint and walk up a small staircase. The treadmill-trained rats never learned to take voluntary steps.

When Courtine and his colleagues dyed the rats' spinal cords and brains—specifically the motor cortex, a band of tissue that helps plan and direct voluntary movement—they observed extensive new networks of neurons in rats that trained on the runway but not in those that trained on the treadmill. An electrophysiological test provided similar evidence. Stimulating the motor cortex with electrodes before the surgeries made the rats' leg muscles twitch. Immediately after surgery their legs did not move in response to the same stimulation. But when Courtine electrically stimulated the brains of rats that learned to walk across the runway, their legs twitched once again, indicating that the active training had reestablished some of the neural connections between brain, spinal cord and legs. Courtine thinks that surviving neurons may have grown new links with one another across the remaining sliver of tissue in the spinal cord, bypassing the lesions. In summary: Forcing the rats to get their brains involved in rehabilitation restored voluntary movement of their hind limbs; removing the need for the brain's involvement with a treadmill ruined any chance of regaining conscious control. The results appear in the June 1 issue of Science.

"This is not a cure for spinal cord injuries," Courtine says, "but what we are working on is quite surprising and encouraging. It's a very different concept from what has been done before, clearly showing that what is really important is to promote a highly functional state during training."

Reggie Edgerton of the University of California, Los Angeles, who has conducted similar research with both animals and people, was impressed with what he called "a very important study." Edgerton was part of a team that helped 25-year-old Rob Summers, who was paralyzed from the chest down after a hit-and-run accident, learn to stand, flex his limbs on command and walk on a treadmill with the help of harness and electrical stimulation of the lower spine. In all likelihood, Summers's determination and willpower contributed to his achievements, just as the rats in the new study only regained voluntary control when they were forced to actively think about walking.

"This new work helps provide some insights into what the mechanisms might be for what we have previously observed in humans," Edgerton says. "The most important information is that you have to engage the brain and make the brain try to regain function. Even if a connection does not yet exist, you can still consciously try. You've got to activate the circuits so they can figure out how to get to the lower spinal cord. This group has put a lot of pieces together from earlier research—we're beginning to get a better understanding of some new possibilities. It's a new ball game for rehabilitation. We now know there is significant plasticity even years after injury."