Computer-connected brains: science fiction or science future?
Newcastle University’s Thomas Hall listens to the chatter between neurons to find signals which could help restore movement to people paralysed by strokes or spinal injuries. He describes his research in his commended entry for the 2014 Max Perutz Science Writing Award.
I visit Charlotte on a Saturday morning, arriving to the smell of fresh baking. After seeing her grandchildren, we head to the village hall for a surprisingly competitive monthly bake-off. But I’m not here just for tea and cake. A year ago, aged 73, Charlotte suffered a stroke, leaving her wheelchair-bound and with her right arm almost completely paralysed. One day she was working as a freelance architect; the next, she was unable to even write or dress herself.
But six months later, in 2034, Charlotte became one of around 200 patients worldwide fitted with a revolutionary new medical device called a ‘brain-computer interface’, or BCI.
Back at home, she shows me the scar on her scalp where doctors implanted thousands of microscopic electrodes in the part of her brain that controls her right arm — the part that was ‘disconnected’ by the stroke.
A tiny cable runs from her scalp, under her skin, to the BCI, which appears as a bump on her chest, like a pacemaker. From there, a cable runs to another set of electrodes implanted in Charlotte’s spinal cord.
The device charges overnight, but it’s switched off now, and Charlotte’s paralysed right arm lies awkwardly across her lap. She switches it on, and suddenly her arm comes to life. It’s jerky for a moment, but the movement soon becomes quite natural.
She reaches for her tea, and explains how it works. The BCI records the tiny electrical signals produced by her brain when she ‘thinks’ about moving her arm. It translates this information into electrical impulses that are delivered, painlessly, to her spinal cord, activating the nerves to her arm and making her muscles contract.
The BCI has been life-changing. She’s not yet back to work, but can dress herself, use a keyboard and bake again. She can’t imagine what life would be like without it.
OK, so that was science fiction! But it’s very possible that within 20 years Charlotte’s scenario will be reality. Across the world, scientists are working hard to solve the remaining challenges.
As a neuroscientist, my research focuses on one of the key challenges: studying which brain signals would be best for controlling such a device.
The technology already exists to record brain signals from the motor cortex (the part of the brain controlling dexterous arm movements) over many months. But for most patients, a BCI needs to work for decades, otherwise the benefits don’t justify the risks of surgery.
As well as having a long lifespan, the brain signals also need to be stable. If they constantly changed, Charlotte would wake up each morning not knowing how to control her arm.
The problem is that the brain reacts to having electrodes implanted in it. The ‘foreign’ material leads to microscopic scarring (called gliosis). Over time, this causes neurons to die, or be pushed away from the electrode. Currently, we can’t record from individual neurons indefinitely.
With my supervisor, Andrew Jackson, I am studying a different type of brain signal, called the ‘local field potential’, or LFP, which includes very low-frequency patterns of brain activity.
Recording individual neurons is a bit like being at a noisy football stadium, trying to use microphones to record the voices of individual fans during the chaos of the match. Recording LFPs is like using the same microphones to record the chanting of the crowd, or the swell of the ‘Mexican wave’.
Our research shows that we can make a reasonable estimate of what an individual neuron (football fan) is saying based on what these slower LFP signals (sounds of the crowd) are doing.
Importantly, these LFP signals appear to be stable over time. And they may have another advantage. BCIs need dramatic miniaturisation, but the major barriers are processing power and battery life. Our low-frequency LFP signals may help, because they can be processed efficiently with low-power electronics.
We made these discoveries in rhesus monkeys, who controlled an experimental BCI device using LFP signals. Monkeys are irreplaceable models in our research, because the neurons that control their arm movements are so similar to ours (whereas in rats, for example, they’re very different).
In other work from our lab, Andrew Jackson and Jonas Zimmermann have shown that temporary arm paralysis in monkeys can be partially reversed by spinal cord stimulation — suggesting that a BCI like Charlotte’s is achievable.
Our research matters because arm paralysis — from common diseases like stroke and spinal injury — places an enormous burden on individuals and on society. Arm and hand function is critical to people’s independence and sense of identity.
I hope our work will contribute towards the development of a BCI like Charlotte’s within 20 years: a fully implantable device that can reanimate a patient’s own arm, and restore independence to many thousands of people.
Thomas M. Hall