Above: Matthew Fifer (right), the technical lead on the brain-computer interface study for the Johns Hopkins University Applied Physics Laboratory, preps study volunteer Buz Chmielewski for a lab assessment.
Johns Hopkins researchers are studying brain-machine interfaces (BMIs) designed to control movement of prosthetic limbs and induce perceived touch in a volunteer with tetraplegia. The goal: to one day help patients with paralysis function more independently.
The researchers are from the Department of Physical Medicine and Rehabilitation, the Department of Neurology and Neurosurgery, and the Johns Hopkins University Applied Physics Laboratory. In January 2019, the group implanted six microelectrode arrays, each the size of a large ant, into both brain hemispheres (half in the motor cortex and half in the sensory cortex) of 50-year-old volunteer Buz Chmielewski, who has a C6 incomplete spinal cord injury and is paralyzed from the shoulders down, with some residual function in his shoulders, biceps and wrists.
The arrays are designed to either read neural signals from the motor and sensory parts of the brain or directly stimulate sensory parts of the brain, in essence bypassing the damaged spinal cord. The recorded neural signals can then be converted into motor commands to control a robotic arm or other devices, or the brain can be stimulated to evoke percepts of touch. This allows the participant to control these devices with only his thoughts or feel percepts of touch when the devices interact with their environment, such as when touching objects.
In laboratory studies, the arrays are connected through wires to a computer that can read the neural signals coming from Chmielewski’s motor cortex, conveying the intention of movement. These signals are then decoded into movement commands that are transmitted to robotic arms, cursors or virtual limbs, allowing the participant to move these devices using “mind control.” The team also has been tracking the volunteer’s accuracy in reaching toward targets, moving individual fingers, and producing various type of grasps using real or virtual prosthetic limbs. They also are studying his percepts of what and where he can “feel” sensations when his brain is directly stimulated in response to object interactions picked up by sensors on the prosthetic limbs. For example, if the participant squeezes a ball with the prosthetic arm, he will feel as if his own fingers are touching the ball.
“At this point, we’ve found that Mr. Chmielewski can control both limbs to perform some simple reaching or grasping movements in a coordinated manner,” says physical medicine and rehabilitation scientist Gabriela Cantarero, who with department chair Pablo Celnik and others has been engaging the patient in three testing sessions each week. Although this research subject’s journey has been “purely altruistic,” Cantarero says, repetition of attempting motor tasks during testing sessions has led to a slight increase in his ability to move his fingers. While these movements are not yet clinically useful, Cantarero says, “It’s cool to see.” Early work was presented last fall in four poster presentations at the Society for Neuroscience annual conference.
Previous BMI studies in humans have been done recording from only one side of the brain, Cantarero says, which may limit movement control of two limbs. The brain has a contralateral organization of control, she says, so the right side of the brain predominantly controls the left side of the body, and the left side of the brain controls the right side of the body. Although some simple movements of one limb can be decoded or controlled from either side of the brain, using both sides of the brain likely becomes necessary for more complex tasks that require simultaneous coordinated movements of both arms, such as opening a bottle or cutting food. The team is studying which types of movements do or do not require bilateral input from the brain to achieve control.
Next steps for research include testing progressively more complex bimanual movements and determining methods of stimulating the brain so that the percept of touch feels more natural and useful, Cantarero says: “One of our main goals is to ultimately build a clinically relevant bimanual task that can improve a user’s independence, such as grabbing a bottle, opening it and drinking from it.”
Engineering a BMI control of that task is more difficult than it might seem, she says, with some components of the movements requiring each hand to work independently toward distinct goals, and other components of the movements requiring both hands to work together in a coordinated fashion. “This is because the way our brain represents the movement of one hand might largely differ depending on what the other hand is doing,” Cantarero says. But the team is determined.
“Most of our everyday activities in life consist of bimanual movements,” says Cantarero. “We want to build something that goes beyond what can be achieved with a single arm, and explore how we can use BMI to simultaneously control two arms in a coordinated manner to truly be of value to a spinal cord injury patient.”
The work is supported by the Defense Advanced Research Projects Agency.
Learn more about the Brain-Computer Interface Study.