“Any sufficiently advanced technology is indistinguishable from magic,” the late science-fiction visionary Arthur C. Clarke wrote in 1961. Watching patients with limb paralysis grasp with their affected hands by the power of their thoughts veers into magic territory. However, it’s happening by way of a brain-computer interface (BCI) with functional electrical stimulation (FES) of the muscles.
Here’s the problem, though: “BCI/FES therapy involves a lot of complex equipment and takes a while to set up,” says Nina Petric-Gray, a PhD candidate in biomedical engineering at the University of Glasgow, Scotland. “Patients have to arrange appointments with trained professionals.”
But what if getting a reliable EEG (electroencephalography) signal could be simpler, so the patient (or a minimally trained caregiver) could manage it at home? That would put rehabilitation of limb after stroke or spinal-cord injury within reach of many more patients, at a drastically lower cost. Petric-Gray has taken on that challenge, designing a portable BCI headset for home use.
Petric-Gray has researched spinal-cord injury and stroke rehabilitation extensively. Her master’s degree is in product design engineering, a program managed jointly by the Glasgow School of Art’s School of Design and the University of Glasgow James Watt School of Engineering. Her principal PhD supervisor is Dr. Aleksandra Vuckovic, senior lecturer in rehabilitation engineering at the University of Glasgow, who has been working on BCIs for 15 years.
“In Britain, there are 50,000 cases, some with complete paralysis and others with partial loss of movement,” Dr. Vuckovic says. “About 1.2 million people a year in the UK experience a stroke. When they leave hospital, about one third can’t properly move or use their arms and would benefit from further rehabilitation at home.”
Victims of spinal-cord injuries or strokes often experience paralysis—partial or complete loss of the use of their limbs—because the brain can no longer deliver instructions to the arms or legs. For at least a decade, rehabilitation specialists have been able to restore some limb mobility by rerouting the signals through neuroelectronic gadgetry that uses brain signals to stimulate the muscles. Some patients have even trained themselves to move again unaided—after months or years of technological therapy. With a BCI, some eventually may recover the ability to use their hands.
The BCI headset records brain activity. “You can measure brain activity, show it graphically on the screen, and teach people how to self-regulate their brain activity,” Dr. Vuckovic says. “When you think about moving, you produce changes in brain activity similar to those that are produced when you actually move. You think of moving your hand, and the computer detects your intention and activates a stimulator. The stimulator sends a signal to the hand muscles, and you get movement.”
Spinal cord–injured patients with incomplete loss of hand function make prime candidates for BCI/FES. “We’re happy if they can get back reach, flex, and extension—basically the dexterity to grab something,” Dr. Vuckovic says. “We might expect better outcomes with stroke patients, as typically they have more preserved control of their hands, plus their injury is in the brain, so we directly target it with BCI. We can certainly help people become more independent and functional.”
Dr. Vuckovic’s lab had tested BCI using a commercially marketed portable headset from Emotiv. “That headset was designed for gaming,” Petric-Gray says. “It didn’t cover key areas of the brain that we were interested in focusing on. The headset didn’t fit different users’ heads in a consistent way. This is essential to recording a good EEG signal—having a good contact between the electrodes and the head.”
Ultimately, Petric-Gray decided to develop her own headsets, using the generative-design capabilities within Autodesk Fusion 360. “The original headset was designed using traditional computer-aided design,” says Craig Whittet, head of product-design engineering at the Glasgow School of Art and Petric-Gray’s PhD co-supervisor. “In late 2018, the concept development work was moved over into Fusion 360 to further enhance that work by applying generative design.”
Generative design lets the design engineer enter a set of requirements from which the software creates multiple iterations. The designer typically will remove prospective designs likely to be rejected by end users for aesthetic or practical reasons. Valid designs might, for example, be rejected if they looked intimidating or would be difficult to clean and maintain.
“I could use 3D scans of heads, import them into the generative-design software, and use them to create a personalized headset design,” Petric-Gray says. Her process generated about 15 different renderings, using various plastic materials.
“In the initial design, I sort of guessed where the electrodes should go,” Petric-Gray says. “But I wanted to be able to map the electrodes more accurately onto the head. That was the main reason for using generative design.”
Most commercial, standardized EEG headsets record only from the forehead, where there is no hair to interfere with electrode-to-skin contact, Dr. Vuckovic says. “If you don’t have a specialist to determine whether you are getting a good signal—whether this is really EEG or just noise—it becomes really tricky.”
Unfortunately, the best places on the scalp for useful EEG signals are normally covered with hair. It can be done, but getting a good contact is much easier if the headset is made to fit the individual patient’s head.
Another obstacle is that EEG electrodes typically require a conductive gel to make a good contact point. There are dry electrodes available for EEG recording applications—such as tracking the progress of dementia or monitoring microseizures in epileptics—that could be adapted. This reduces setup time and eliminates the need for the patient to shampoo after using the headset. If the patient undergoes the procedure daily, or several times a day, the hygiene issue can be significant.
Initially, Petric-Gray tried to develop a one-size-fits-all headset with adjustable bands but found that a single design wouldn’t work for everyone. She needed the ability to design customized devices that could be fabricated quickly and cheaply for individuals.
The intent was to use an additive-manufacturing process, essentially 3D printing, to minimize the cost of custom fabrication. Petric-Gray also wanted to minimize the number of moving parts used when fitting the headset to the patient, which add to the complexity and cost of manufacturing.
So far, Petric-Gray’s bespoke headsets have been tried only with healthy test subjects. The system has not yet undergone formal medical-device testing; Additional refinement and funding are needed to take the device into actual clinical trials.
If the team can show regulators at the UK National Health Service (NHS) and the Medicines & Healthcare products Regulatory Agency (MHRA) that the device has both therapeutic and cost advantages over hospital-based rehabilitation, it could be commercialized for rehabilitation, for which the NHS might provide reimbursement.
“We’re working on another application in neuropathic pain, which affects about 8% of the general population,” Dr. Vuckovic says. “So the headset really is not limited to movement rehabilitation. It can be used for any kind of neuromodulatory intervention. You also could use it for improving peak performance in gaming. We use it for that here to test how healthy people play multiuser games by using solely their brain waves.”