A combination of virtual reality, tactile feedback and exoskeleton training has allowed paralyzed individuals to regain both voluntary motion and sensation below the level of their spinal cord injury, scientists reported today.

The recovery is not such that it allows the trained patients to walk, or grasp objects, independently. But several patients were able to walk with the help of therapists and either an exoskeleton or a walker.

The patients also regained control over their visceral functions. Miguel Nicolelis, who described the study to reporters in a press conference earlier this week, said the patients had regained “some level of bladder control, which is very important, and also bowel movement . . . They were able, for the first time in many years, to control their bathroom routine.”

Patients’ sexuality also benefited. Male patients reported better erections, and “one of the ladies decided to deliver a baby . . . she had a second baby because she recovered visceral sensation, too,” Nicolelis, who is a professor of neuroscience at Duke University, said. “She could feel the contractions.”

Overall improvement was sufficient to reclassify four of eight trained individuals from complete paraplegia, meaning they had no sensation or motor function below the level to partial paraplegia, after 12 months of training.

Furthermore, 28 months after training started, recovery “has not plateaued yet.” Individuals continue to train, and continue to show functional improvement.

Nicolelis has been working on brain-machine interfaces for nearly two decades, described them as a prime example of how basic science can lead to unexpected applications. Originally, his goal in working on brain machine interfaces was to develop “a tool to probe the brain,” he told reporters.

At some point, he realized that his work had “a lot of potential” for rehabilitation medicine.

But “I think I speak for the whole field – nobody ever imagined that one day we would be talking about using brain-machine interfaces to induce partial neurological recovery in individuals who . . . [are] totally paralyzed.”

The findings overturn basic ideas about the amount of functional recovery that is possible after neural damage.

“It was assumed until now that after a year or so if there was no significant improvement of any sort, you were basically going to be a chronic paraplegic for life,” and so rehabilitation was aimed at helping patients gain independence with the help of assistive devices such as wheelchairs.


The original goal of the project, which is called the Walk Again Project and headquartered in Brazil (where a paraplegic in a brain-controlled exoskeleton kicked off the 2014 world cup in men’s soccer), was to develop one such assistive technology, a brain-controlled assistive neuroprosthetic exoskeleton.

But the report, which Nicolelis and his colleagues reported in the Aug. 11, 2016, issue of Scientific Reports, “may upgrade this technology from just an assistive technology to a potential rehab therapy for patients with severe spinal cord injuries,” he said. “If you invest in these patients with the correct type of rehabilitation therapy, you may be able to give them back some neurological functions that may be useful for improving their quality of life dramatically.”

In terms of the experimental protocol, Nicolelis suspects that the unexpected recovery was due to the presence of “very rich tactile feedback” in addition to mental imagery and movement.

Because the patients in the study had been paralyzed for anywhere from three to 13 years, the first step was to re-establish brain activity when they imagined themselves walking.

“When they got to us, we couldn’t detect any signal” related to imaginary walking, Nicolelis said. “It was almost as if the brain had erased the concept of movement by walking.”

To re-establish that brain activity, patients first learned “to use their brain activity to control an avatar body of themselves on a virtual soccer field,” he explained.

The avatar would walk when the subjects generated neural activity by imagining themselves walking.

But the neural activity was not just transmitted to the avatar – it was also fed back to the patients.

“As this avatar touched the ground, a tactile signal, a pressure signal would be delivered to the arm of these patients,” Nicolelis explained, essentially as though their feet were on their arms.

Patients developed “phantom sensations – even though they were in a virtual environment, they reported that they felt themselves walk.”

As a result of that virtual training and feedback, “by the time they got to the [exoskeleton] and started walking, we already could detect the reappearance of the representation of their legs in their brains. So the concept of walking was back.”

Anatomically, the combination of imagery, motion, and tactile feedback may have “rekindled” remaining nerves to be able to send messages from the brain of the patients to the periphery by triggering “reorganization in the cortex and spinal cord of these patients…if there were any nerves that survived the original injury, which is possible, we think that we may have taken advantage of this.”

Ultimately, the patients were able to regain varying levels of both feeling and muscle control in the part of the body that was cut off by their injuries.


For now, the team is focused on spinal cord injury, but Nicolelis said he believes the protocol could be adapted to other types of neurological damage, including stroke.

He also suggested the brain-machine interfaces could work in parallel with other approaches such as cell therapy or more invasive implants that use deep electrical stimulation.

“They are not mutually exclusive; they are synergistic in some cases,” Nicolelis said. “We may see in our lifetimes patients receiving a stem cell implant and using a brain-machine interface postsurgery to improve or potentiate the plasticity they may get from these implants.”

The team is also working on making its technology cheap enough for broad use. Nicolelis was optimistic that could happen.

“A lot of the progress can be achieved initially with a virtual reality system,” he said, which is “very simple to render.”

Nicolelis said there are roughly 25 million individuals who are paralyzed due to spinal cord injury – certainly a large enough group for economies of scale.

“Technology development process starts with things that are very expensive,” he said. But “we are already working on a version of the entire protocol that will not require all the sophisticated gizmos that we used here.”