Scientists at Duke University have demonstrated a wireless brain-machine interface (BMI) that allows monkeys to navigate a robotic wheelchair using their thoughts. This is the first long-term wireless BMI implant that has given high-quality signals to precisely control a wheelchair’s movements in real time.
“This is the first wireless brain-machine interface for whole-body locomotion,” says Miguel Nicolelis, professor of neuroscience at Duke who led the work published in the journal Scientific Reports. “Even severely disabled patients who cannot move any part of their body could be placed on a wheelchair and be able to use this device for mobility.”
Nicolelis and his colleagues pioneered brain-machine interfaces in a 1999 study on rats. Since then, researchers have done several demonstrations of primates using brain signals to control prosthetic arms, advanced devices, and computers, and even receive haptic signals.
Despite those exciting advances, reliable, long-lasting implants that give high-quality signals have been lacking for human trials and use.[shortcode ieee-pullquote quote=""I was very critical ten years ago of attempts to use rigid electrodes and probes in humans and of using tethered humans in clinical trials. What I saw were mediocre results in implants that couldn't last more than a few weeks"" float="left" expand=1]
This new BMI finally gives that hope, Nicolelis says. It is groundbreaking for a few reasons. Until now, researchers have mainly used noninvasive EEG electrodes attached to the scalp for thought-control of wheelchairs. But the low-frequency signals don’t contain enough information to allow continuous, real-time control of machines. More sophisticated brain implants have relied on wires connecting implanted electrode arrays to external computers, which increases infection risks and is simply impractical.
Another drawback of these past approaches: they have all produced wheelchair movements by recording and reproducing brain signals for joystick or arm movements. “I was very critical ten years ago of attempts to use rigid electrodes and probes in humans and of using tethered humans in clinical trials,” Nicolelis says. “What I saw were mediocre results in implants that couldn’t last more than a few weeks.”
In the new study, the researchers were able to decode neural signals for whole-body movement through 2-D space. They converted that into the translational and rotational velocities for a wheelchair so that a monkey with the BMI implant could navigate the wheelchair toward a reward. This approach offers much more nuanced control of the wheelchair.
The team began their experiments in 2012. Two monkeys were implanted with hair-thin microelectrodes to monitor about 300 neurons in each animal. At first, the monkeys sat in wheelchairs that were moved—first in straight lines and then more distorted paths—towards a bowl of grapes. The BMI device, affixed on the animal’s head, has a 512-channel wireless interface that sends signals to a computer. The researchers recorded the animals’ neuronal activity as they perceived the wheelchair’s trajectory. The signals were used to train a decoder program.
Then the monkeys attempted to control the wheelchair by thinking. Again the wireless BMI sent signals to the computer and the decoder translated brain signals into motor commands for the wheelchair. Over time, the monkeys gained better control on the chair’s movements and were able to navigate to the grapes faster.
The team hopes to record even more neural signals to get more accurate and reliable control. They’ve already developed a wireless interface that can transmit 1000 channels of data.
Also exciting is the nearing prospect of clinical trials. The researchers have been recording high-quality neural recordings from a monkey in their laboratory for seven years now. “We’re assuming that this could go for a decade, which for me would be a benchmark that we need for human clinical trials,” Nicolelis says.
Prachi Patel is a freelance journalist based in Pittsburgh. She writes about energy, biotechnology, materials science, nanotechnology, and computing.