Researchers at UC Irvine, Caltech and USC reported a bidirectional brain-computer interface that let a participant control a robotic gait exoskeleton using brain signals and receive artificial leg sensation via direct electrical stimulation of the sensory cortex. The results appear in a paper published recently in Brain Stimulation.
The study used bilateral interhemispheric subdural electrocorticography (ECoG) implants to read motor intent from the leg motor cortex and to deliver targeted stimulation to the somatosensory cortex. A 50-year-old woman undergoing epilepsy evaluation operated an Ekso GT powered exoskeleton across 10 exercises, the team said.
Performance was high in sensory validation tasks. In a blind step-counting test the participant identified steps with almost 93% accuracy. In a sensory discrimination task she reported right-leg, left-leg and no-stimulation correctly at 96%, 84% and 100%, respectively. The researchers reported no adverse events.
Lead author Jeffrey Lim described the system as a compact, embedded platform that runs neural signal acquisition, real-time decoding, electrical stimulation and wireless communications on three 48-megahertz microcontrollers. The team positions the setup as a benchtop prototype for a future fully implantable system.
Co-author An Do said, "Recovering the ability to walk ranks among the highest rehabilitation priorities for paralyzed individuals." Do and others noted that most commercial exoskeletons require manual control and provide no direct sensory feedback, which the bidirectional BCI aims to address.
Co-author Payam Heydari outlined a possible clinical path: a skull-mounted unit linked by a subcutaneous cable to a chest-implanted processing and stimulation module, eliminating transdermal components that raise infection risk. The paper also cites prior safety data for long-term cortical stimulation in approved devices.
The work was approved by the institutional review boards at UC Irvine and Rancho Los Amigos National Rehabilitation Center and was funded by the National Science Foundation. The authors call this the first implementation combining bilateral interhemispheric leg sensorimotor recording with artificial bilateral leg sensation in a single BDBCI–exoskeleton system.
Photo credit: www.universityofcalifornia.edu
Tags: brain–computer interface, electrocorticography (ECoG), exoskeleton, somatosensory cortex, spinal cord injury
Topics: Brain–computer interfaces, Neuroprosthetics & neural implants, Neuromodulation