Cathal Harte of NeuroRestore is developing software with the aim of interpreting the brain’s intentions to help restore motor function.
Researchers in Switzerland are working on groundbreaking trials that could give people with paralysis the ability to walk again by sending electrical impulses to targeted nerves.
NeuroRestore is a research, innovation and treatment centre focused on ‘electroceuticals’ – a type of neurotherapy that uses electrical stimulation to help neurological functions. There are various potential applications for this, such as improving motor function in people who have had a stroke or have Parkinson’s disease.
But one key area of research is restoring motor functions for people with chronic spinal cord injuries. NeuroRestore researchers have been working on a clinical feasibility study called Stimo, which has shown very positive results in experimental treatments helping people with full paralysis to walk.
One of the people working on this technology is Cathal Harte, an Irish software engineer who joined NeuroRestore in 2020. He leads the development efforts behind the centre’s clinical software.
The NeuroRestore centre was set up by the Defitech Foundation, Lausanne University Hospital, the University of Lausanne’s Faculty of Biology and Medicine, and the Swiss Federal Institute of Technology (EPFL).
Some of the team behind NeuroRestore made headlines in 2018 after helping David Mzee, a man who had lost the use of his legs following a spinal cord injury, to walk again through continuous electrical stimulation and a pacemaker-like implant.
“I was taken on to bring the next-generation software, which really opened the door of possibilities to do things in an automatic way,” Harte told SiliconRepublic.com.
Harte and the team at NeuroRestore aim for this software to eventually interpret motor intentions from the brain and translate them into targeted epidural spinal stimulation that can restore movement for people with paralysis.
Developing the software
Harte said the first generation of the NeuroRestore software was more simplistic. It allowed users to make “eye-based judgements” on the electrical stimulations using a separate app.
An important part of developing the software involves testing different stimulation settings on different nerves. The team can examine this electromyography, or EMG, data to see which muscles are being activated by stimulating specific nerves.
“The data is a vital part of the mapping process, it’s for the sake of the patient right there in front of you,” Harte said.
This data is then used to understand exactly how to stimulate the nerves to create a specific movement, such as a walking motion. Thanks to everything being integrated in the latest version of the software, Harte said it’s possible to integrate machine learning.
“We know exactly what the configuration we’re sending is, we know exactly for how long we’re stimulating,” he explained. “We can annotate all of the muscle signals with that information, send this all to a machine learning program and then it can make the calls.”
Harte explained that machine learning is an important step to bring this sort of technology to a product stage.
“When you’re a dedicated research team, you have one patient every six months or so, then you can do it with a bunch of very smart people in the room, manually looking at the data,” Harte said. “But if you want to bring this to hundreds of thousands of people, you have to automate these kinds of things.”
One participant in the Stimo study is Michel Roccati, an Italian man who became paralysed after a motorcycle accident.
Harte said earlier patients in Stimo trials had some ability to move their legs, so the focus was in aiding the rehabilitation process and giving them “the missing parts of their walking patterns”. Roccati could not move his legs at all, but Harte said the results have been “amazingly successful” and the man has been able to walk again.
“Firstly, it’s just because of his attitude. He’s an incredibly dedicated guy. He’s really motivated to do this,” Harte said.
After inserting a surgical implant on his spinal cord, the NeuroRestore team attached two remote controls to Roccati’s walker.
These controls were connected wirelessly via a tablet that forwards the signals to a pacemaker in his abdomen, which then relays the signals to the implant that stimulates specific neurons.
“That triggers a left step or a right step. And if it’s not triggering a left separate step, it’s engaging a standing stimulation. So essentially, this gives him autonomy to walk.”
NeuroRestore said the three patients involved in this trial followed a training regimen based on stimulation programmes and were able to regain muscle mass, move more independently and take part in social activities like having a drink standing at a bar.
Some of the eventual products that could be created from these studies are shown from Onward Medical, a medtech company that has a research partnership with NeuroRestore and was involved in the Stimo study.
Onward is developing two technology platforms based on targeted nerve stimulation to try solve multiple medical issues.
For example, the company’s ArcEx platform aims to improve the physical movement of hand and arm function for people with spinal cord injuries. This is an external device that includes a wearable stimulator and wireless programmer.
The other device, ArcIm, is an implantable pulse generator and lead that is placed near the spinal cord. This device was used in the Stimo study, but Onward believes it can help in other areas such as improved sexual function, bladder, and bowel control.
Brain Spine Interface
One trial that Harte said he is excited about is NeuroRestore’s Brain Spine Interface study, which aims to establish a digital bridge between the brain and the spinal cord to restore motor control of paralysed limbs.
Using various brain recording techniques such as electroencephalography, electrocorticography and intracortical recordings, the centre said it is possible to decode motor intentions and translate them into targeted stimulations.
“There’s an implanted brain decoder, what you do is you remove parts of the skull, and you replace it with a sensor,” Harte explained. “Then you train pattern recognition algorithms to say, OK, think about moving your left leg, think about moving your toe, think about various things and it learns those brain states.
“And so it’s able to match up then what it’s sensing with intentions, and then we turn those intentions into spinal cord stimulation.”
10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.