Most brain-controlled walking systems have a strange gap at their center. A person can command a robotic exoskeleton to take a step, but they cannot feel the step happen. The signal runs one way, brain to machine, and stops there. Walking, for the rest of us, is not like that. It is a constant loop of moving and sensing, the ground pushing back, the leg reporting where it is.
A study in Brain Stimulation tries to close that loop. Researchers led by teams at UC Irvine, USC, and Caltech built what they call a bidirectional brain-computer interface, a system that reads movement intent from the brain and writes sensation back into it at the same time. In their proof-of-concept test, one person used it to walk a robotic exoskeleton by thought while feeling artificial sensations in both legs.
Reading intent, writing sensation
The participant was an epilepsy patient who already had electrodes placed on the surface of the brain for seizure monitoring. That surgery, called electrocorticography, gave the researchers a rare window. They mapped which electrodes lit up over the leg motor cortex when the person imagined stepping, and separately found spots in the somatosensory cortex where a small jolt of current produced a felt sensation in the legs.
Then they ran both channels together. A custom embedded system decoded stepping intent from the motor electrodes in real time and used it to drive a robotic gait exoskeleton. As the legs swung, the system delivered pulses of electrical stimulation to the sensory cortex, producing artificial percepts timed to the movement. Command going out, feeling coming back, in one device.
The decoding held up well. Across ten runs, the match between what the person intended and what the system read hit a correlation of 0.92, with a lag of about three and a half seconds. To check that the sensation was real and usable, the researchers ran a blind step-counting task. The participant, without watching, had to report leg sensations, and got them right 92.8 percent of the time. The authors also confirmed the electrical stimulation was not secretly corrupting the decoding, a real worry when you stimulate and record from the brain at once. No adverse events were reported.
Why the middle of the brain matters
One quieter detail may end up mattering most. The electrodes sat along the interhemispheric fissure, the deep groove between the brain's two halves where the leg region of the sensorimotor cortex is tucked away. Reaching it is harder than tapping the more accessible outer surface of the brain, which is where earlier gait interfaces mostly worked. But the leg territory lives down in that groove, and the researchers report that placing electrodes there gave better decoding than the older lateral approaches. For a system meant to control walking, going where the legs actually live seems to pay off.
That is the practical thread running through the work. Restoring movement after spinal cord injury is the headline goal, but the group frames sensation as part of the same problem, not a nice extra. Walking without feedback is unstable and effortful. A leg you can feel is a leg you can trust your weight on.
One person, and a long way to go
The limits here are large, and the authors do not hide them. This is a single subject. The person was not paralyzed. They were an epilepsy patient with temporary electrodes, and the whole test rode on hardware placed for an entirely different medical reason. So the study shows the pieces can work together, not that the approach restores walking to someone living with paralysis. There is no long-term data, no durability testing, nothing about how the artificial sensations feel over days or weeks rather than minutes.
A three-and-a-half-second lag between intent and motion is also a long time when you are trying to walk. And a percentage that high on a step-counting task in one motivated participant tells you the signal is there, not that it will hold in a crowded clinic with a different brain.
What the study does establish is a direction. Fully implantable two-way interfaces, the kind that could one day let a person with paraplegia both move and feel a limb, need a foundation to build on. This is one early brick. The researchers are clear that translating it into a system people can actually live with is the work still ahead, and that work has not been done yet.
Comments