Until a few years ago, reversing paralysis was the stuff of movie miracles.
Yet according to Dr. Andrew Jackson, a neuroscientist at Newcastle University in the UK, as early as the end of this decade, we may witness patients with spinal cord injuries regain control of their own two legs and walk again.
By implanting a wireless neural prosthetic into the spinal cord of paralyzed monkeys, a team led by Dr. Grégoire Courtine at the Swiss Federal institute of Technology (EPFL) in Lausanna, Switzerland achieved the seemingly impossible: the monkeys regained use of a paralyzed lower limb, a mere six days after their initial injury without requiring any training.
The close-looped system directly reads signals from the brain in real-time and works on the patients’ own limbs, which means it doesn’t require expensive exoskeletons or external stimulation of the patient’s leg muscles to induce the contractions necessary for walking. That’s huge: it means the system could be readily used by patients in their own homes without doctor supervision.
“When we turned on the brain-spine interface for the very first time … and the animal was showing stepping movement using its paralyzed leg, I remember a lot of screaming in the room; it seemed incredible,” says Dr. Courtine in an interview with Nature.
“The study represents a major step towards restoring lost motor function using neural interfaces,” agrees Jackson, who was not involved in the study.
Bridging the Gap
Every time we decide to move, the brain sends a cascade of signals down the spinal cord to instruct our muscles to contract accordingly. Severing this information relay, as in the case of spinal cord injuries from sports or car accidents, often results in irreversible paralysis.
Rewiring a damaged spinal cord is incredibly tough. The nerves don’t regenerate, even after careful coaxing with different cocktails of regenerative drugs. The injury sites are often hostile to stem cell transplants, making it difficult for foreign cells to grow and integrate.
To get around the issue, Courtine and other brain-machine interface (BMI) experts are turning to neural interfaces to manually reconnect brain and muscle. And Courtine’s system is exceedingly clever.
To start off, his team designed two implants: one to receive incoming signals from the brain and another to replace the damaged spinal cord.
The first, a neural interface, is made up of arrays of 96 microelectrodes that hook into the parts of the brain that controls leg movement. Once implanted, it automatically captures signals coming from multiple neurons that usually work together to give out a certain command — for example, flexing your foot, bending your leg or stop walking altogether.
The matchbox-sized device then sends the signal to an external computer, which uses an algorithm to figure out what movement the neural signals were encoding.
Fine-tuning the input signal took a lot of effort. Figuring out how different sets of electrical signals represent different aspects of movement was just the first step. The scientists also had to map out how the signals cyclically changed with time as a monkey walked, to ensure they could reproduce the smooth, gliding gait when working with a paralyzed monkey.
Once the neural signals were decoded, the computer used the information to wirelessly operate a second electrode array sitting over the lower part of the spinal cord of a paralyzed monkey, below the level of injury. This second implant, the “pulse generator,” acts as an electrical stimulator that takes in messages from the brain implant and delivers them to undamaged parts of the spinal cord that normally control leg muscle movement with a series of zaps.
The results were astonishing.
In two monkeys that each lost the use of a hind leg, the wireless brain-spine interface allowed them to walk normally within the first week after their injuries — without any training. As time went on, the quality and quantity of the steps improved, suggesting the system had triggered neuroplasticity in the brain and damaged spinal cord.
The field of brain-machine interfaces is moving so fast that blink, and you might miss the latest breakthrough. Within the past year or so, BMIs have allowed paralyzed patients to Google on a tablet with brain waves, grasp objects using robotic surrogates and control a variety of prosthetic hands and other devices. And just a few months ago, a surprisingly study showed that implants that directly stimulate the spinal cord helps paraplegic patients recover some voluntary movement of their own legs.
Yet even amongst this slew of incredible advances, Courtine’s study stands out. For one, walking is genuinely hard. So far, most neural interface studies have focused on reanimating the upper body.
“There’s quite a lot to locomotion, so when we’re walking we’re not just moving our legs to step. We’re also controlling balance and coordinating activity across both sides of the legs,” says Jackson, “So restoring movement to the legs brings with it a different set of challenges to restoring grasping movement in the hand.”
Courtine puts it even more bluntly. “Walking is all-or-nothing,” he says.
If patient can’t walk normally with a neural prosthetic they may instead prefer to remain in a wheelchair as the more pragmatic solution. There’s much more pressure to deliver something that works right off the bat.
Which leads to the second thing that makes Courtine’s system impressive. With wireless, closed-loop stimulation, patients would no longer be tethered to expensive equipment by dozens of wires. Even the computer may soon be out of the picture.
“The only reason why the computer is there is because it allows some flexibility in how we change different algorithms that we use to control new activity and that we use to control the stimulation,” says study author Tomislav Milekovic to Motherboard. In humans, he envisions direct communication between the brain and spinal cord implants, without having to bring the signal outside the body.
Going from monkeys to humans requires a bit more work. Unlike monkeys that use all four limbs to walk, we’re bipedal. And real-world injuries are often messier than the surgically induced lesions in the study, which only damaged one leg.
That said, in humans with incomplete spinal cord damage, there’s a lot of circuitry that survives and could be commandeered for the system to work on.
“It seems quite feasible that a similar technique could be used to generate walking in an individual who has both legs paralyzed,” says Jackson.
And according to him, that day may come much faster than you’d imagine.
We’re seeing a remarkably fast translation from first demonstrations in monkeys to clinical trials, usually around four to five years, he says. If the trend continuous, we may be seeing the first human brain-spinal prosthetic trials as early as 2020.
Courtine and his team have already begun testing the spinal stimulator in eight wheelchair-bound patients. Once that part’s refined, he’s moving on to optimizing the brain implant and hooking the two parts together.
“There are many, many challenges that we are going to face in the coming decade to optimize all these interventions, but we are really committed to making this step forward,” he says.
Banner image credit: EPFL/YouTube
A Neural Bridge. The brain-spine interface developed for this study uses a brain implant like this one to detect spiking activity in the brain’s motor cortex. Seen here, a microelectrode array and a silicon model of a primate’s brain, as well as a pulse generator used to stimulate electrodes implanted on the spinal cord.
Credit: Alain Herzog / EPFL