Brain-Computer Interface for Spinal Cord Injury

Sensors use the brain’s electrical signals to power a robotic arm

Loading the podcast player...

This segment is part of the IEEE Spectrum series “The New Medicine

Susan Hassler: Every time you move your arm or even think about doing it, your brain generates electrical signals. Scientists are now trying to decode those signals and use them to move artificial limbs. A technology like this could make a world of difference to amputees or those who are paralyzed. And it has been tested for the first time on someone with a spinal cord injury. Prachi Patel visited the University of Pittsburgh Medical Center to find out more.

[ambient sound; wheelchair whirring; Tim Hemmes talking to girlfriend]

Prachi Patel: Tim Hemmes was 23 when he broke his spinal cord in a motorcycle accident. He was paralyzed from the neck down. That was eight years ago. Last fall, surgeons at the University of Pittsburgh Medical Center placed a small sensor on the surface of his brain. Four weeks later, Hemmes could move a robotic arm with his thoughts.

[ambient sound; robotic arm moving]

Prachi Patel: A video shows him concentrating intensely. The metal hand moves with erratic bursts and finally touches a researcher’s palm.

[ambient sound; “All right!” “There you go” “Yay!” “Nice!”]

Tim Hemmes: To have Wei standing there and to reach out to him, that was what I’ve been working for seven years. Whether it was robotic, whether it was metal and plastic—my mind, my thought process, put that there.

Prachi Patel: In 2008, the Pittsburgh team had shown that monkeys could feed themselves treats by controlling a robotic arm with their minds. Hemmes is the first human to have tried the technology. Michael Boninger, the lead physician on the research trial, shows me the sensor that was used to read Hemmes’s brain signals.

Michael Boninger: You can see there’s a bunch of tiny, like 1-millimeter silver spots. Those are the electrodes. And it’s through this small pad that’s the size of one of those designer postage stamps we’re able to record the electrical signals. The only thing we’re doing is recording the electrical signal that the brain normally produces when someone thinks.

Prachi Patel: People with spinal cord injuries can think about moving their arms or clenching a fist, but their brain signals can’t reach their arm. So the researchers direct those signals to a prosthetic arm through what’s called a brain-computer interface. Specialized software decodes the brain signals.

Michael Boninger: Right now, the electrode sits on the surface of the brain. The wires then come out, and we connect them to a computer. The computer then does the high-level analysis that says, “Okay, these electrical impulses mean that the subject wants to do this.”

Prachi Patel: Hemmes’s trial started with the researchers’ imaging his brain as he imagined moving his right arm. That helped them pinpoint which area of his brain was firing. Neurosurgeons placed electrodes on that area. Then the researchers recorded his brain signals as he watched a figure on the computer move its arm. Software analyzed the up and down spikes of those signals and learned what they looked like as Hemmes watched the arm move in a certain direction.

Michael Boninger: The way we figure out the code is by having the subject watch the arm move. We see this electrical activity; that’s the code. And then we say, “Okay, now imagine the arm’s moving,” and we use that same code to drive the arm to move.

Prachi Patel: Once the computer had learned Hemmes’ brain signals, it was his turn to train. Six days a week for four weeks, Hemmes practiced working with the brain-computer interface. He first tried to move a red ball on a computer screen to the left, right, up, or down.

[audio from two-dimensional computer training video]

Michael Boninger: Tim was able to generate a really strong signal when he thought about flexing his wrist or extending his wrist. We linked that thought to a specific movement of the cursor to a specific spot in space. So Tim knew that whenever he thought about flexing his wrist, the ball would move to the right.

Prachi Patel: The thought of bending his elbow or flexing his thumb generated other signals. The researchers linked those to different cursor movements. Tim Hemmes progressed to moving the ball diagonally and then on a three-dimensional screen.

[whirring sounds; “bing” sound]

Prachi Patel: Finally, he went on to the robotic arm, controlling it to reach out and touch targets arranged on a panel.

[whirring sounds]

Prachi Patel: Slowly but surely, Hemmes could take the arm where he wanted it to go. A dramatic video shows him touching his girlfriend’s hand.

[ambient sound; “aww”; “baby”]

Tim Hemmes: [laughs] Yeah, the feeling of that was, obviously the adventure isn’t over yet. We’ve still got lots of stuff to do here, but all my hard work for the last seven years was put into that one moment, and it wasn’t my arm, but I was able to do it, and that was really touching to me.

Prachi Patel: The Pittsburgh research team wants to test the technology on at least five more patients. They also plan to make it wireless. Neurobiologist Andrew Schwartz, who has been working on this technology for years, says moving a prosthetic arm is just the beginning.

Andrew Schwartz: And the next step, then, is to put sensors on that arm—for instance, tactile sensors on the fingertips—so when the hand encounters an object, it can feel that object and convey that back to the brain and stimulate [the] sensory region of the cortex so [the] subject has some sort of idea what they are touching.

Prachi Patel: It might also be possible to use the brain-computer interface to stimulate muscle fibers so that patients could use their own arms. But that’ll come in the future. For now, Hemmes says he did his part in furthering the technology and looks forward to the day 5 or 10 years from now when it will help him.

Tim Hemmes: Seeing what I was able to do, I believe in this wholeheartedly. And I believe this is one avenue for the future for spinal cord injury. And reaching out and grabbing a door, you know, being able to grasp a drink instead of asking for something...I mean, I’m solely dependent on everything, you know. If I have a runny nose, to wipe my nose, you know, to grab a drink, to open a door—anything. What would this type of technology do? It would completely change my world. I don’t—I mean in every aspect of it, it would completely change my world for the positive.

Prachi Patel: This is Prachi Patel.

Photo: UPMC

Advertisement
Advertisement