A whole-body exoskeleton controlled by brain signals recently helped a tetraplegic man move his arms and walk. He used a ceiling-mounted harness for balance. While the early results are promising, the four-limbed robotic system and brain-machine interface will need refinement before clinical application and before it can become widely available, said the research team behind the exoskeleton.
The tetraplegic patient was able to move all four of his paralyzed limbs using a system that recorded and decoded brain signals to operate the whole-body exoskeleton, according to results of a trial published in The Lancet Neurology journal.
The researchers conducted a two-year study for the robotic system, which they said could help improve patients’ quality of life and autonomy. For now, however, the treatment is experimental, and major improvements still need to be made, they said.
Related content: The Robot Report issue on exoskeletons
Whole-body exoskeleton controlled by brain waves
“Ours is the first semi-invasive wireless brain-computer system designed for long-term use to activate all four limbs,” said Alim-Louis Benabid, president of the Clinatec executive board and professor emeritus at the University of Grenoble, France. Clinatec is a biomedical research laboratory of the CEA, or French Alternative Energies and Atomic Energy Commission.
“Previous brain-computer studies have used more invasive recording devices implanted beneath the outermost membrane of the brain, where they eventually stop working,” Benabid said. “They have also been connected to wires, limited to creating movement in just one limb, or have focused on restoring movement to patients’ own muscles.”
In cervical spinal cord injuries, the most severe of spinal cord injuries, around 20% of patients are left tetraplegic, with all four limbs partially or completely paralyzed. The 28-year-old patient in the new trial was paralyzed from the shoulders down, with only some movement in his biceps and left wrist. He was able to operate a wheelchair using a joystick controlled with his left arm.
Two recording devices were implanted, one either side of his head between the brain and the skin, to span the sensorimotor cortex — the area of the brain that controls sensation and motor function. Each recorder contained a grid of 64 electrodes that collected brain signals and then transmitted them to a decoding algorithm. This system translated the brain signals into the movements the patient thought about, and sent commands to the whole-body exoskeleton to complete them.
Training the algorithm by playing a video game
Previously, another patient was recruited to the study but was excluded because the brain implants stopped communicating with the algorithm and had to be removed.
Throughout the 24 months of the study, the remaining patient did various mental tasks to train the algorithm to understand his thoughts and to progressively increase the number of movements he could make. This included controlling a virtual avatar to play a video game similar to Pong, reach for targets with an avatar and in the whole-body exoskeleton, and walk.
The patient’s progress was measured in terms of how many degrees of freedom he was able to achieve during tasks, from operating a brain-powered switch to start walking, to reaching out to touch 2D and 3D objects. The whole-body exoskeleton has 14 joints and 14 degrees of freedom. The operator spent a total of 45 days operating it in the lab, and the skills he acquired were reinforced with 95 days spent training at home with a researcher using an avatar and the video game.
The simplest task was to turn the brain switch on to start a walking in video game where he made an avatar walk. That was followed by making the exoskeleton start walking while attached to the suspended harness.
Seeking success, stability for whole-body exoskeleton
The patient’s success was measured in terms of how many times he managed to activate the switch. Two months after surgery, he was successful 73% of the time during six sessions using the exoskeleton. Using the avatar, video game, and exoskeleton combined, he covered a total of 145m (475 ft.) with 480 steps over 39 sessions.
“Our patient already considers his rapidly increasing prosthetic mobility to be rewarding, but his progress has not changed his clinical status,” said Benabid.
Using both the avatar and the exoskeleton for more complex tasks, the tester progressed from reaching out to targets on cubes with one hand at a time (moving in three dimensions) five months after surgery to using both hands to touch targets on the cubes 16 months after surgery (moving in eight dimensions, including rotating both wrists). He completed five eight-dimensional tasks with a success rate of 71%.
“Innovative adaptive algorithms based on artificial intelligence methods [machine learning] have been developed to decode a large number of degrees of freedom,” said Dr. Tetiana Aksenova, brain-computer interface signal-processing research director at the University of Grenoble. “The exceptional quality of the collected neural signals allowed a stable and robust decoding.”
Over the 24 months of the trial, the system did not need to be recalibrated for up to seven weeks, demonstrating that it may be suitable for day-to-day use over a long period. The quality of the recordings from the implants remained stable, the algorithm continued to decode the signals, and the patient experienced no post-surgical complications.
The Robot Report is launching the Healthcare Robotics Engineering Forum, which will be on Dec. 9-10 in Santa Clara, Calif. The conference and expo will focus on improving the design, development, and manufacture of next-generation healthcare robots. Learn more about the Healthcare Robotics Engineering Forum, and registration is now open.
Next steps and managing expectations
“Our findings could move us a step closer to helping tetraplegic patients to drive computers using brain signals alone, perhaps starting with driving wheelchairs using brain activity instead of joysticks and progressing to developing an exoskeleton for increased mobility,” said Prof. Stephan Chabardes, neurosurgeon at the Centre Hosptalier Universitaire of Grenoble-Alpes, France.
Three further patients have been recruited, and the trial is ongoing. The next goal of the researchers is to solve the problem of allowing a patient to walk and balance autonomously without using a ceiling suspension system.
Further studies will also shed more light on brain function, providing more information on how the sensorimotor cortices generate the signals needed to achieve real and virtual movements. The authors found that the patient was able to perform tasks with an average of 10% to 20% greater success using the exoskeleton than with the avatar. This may be because feedback with the avatar is purely visual, whereas with the exoskeleton the patient gets richer feedback from the real world.
“An originality of this study is showing the control of four limbs, whereas in most previous studies, only one limb was controlled,” commented Tom Shakespeare, a professor at the London School of Hygiene and Tropical Medicine. “However, autonomous walking with equilibrium is not so far possible.”
“Although this study presents a welcome and exciting advance, we must remember that proof of concept is a long way from usable clinical possibility. A danger of hype always exists in this field,” he added. “Even if ever workable, cost constraints mean that high-tech options are never going to be available to most people in the world with spinal cord injury. One analysis suggests that only 15% of the world’s disabled population have access to the wheelchairs or other assistive technologies that they need.”