What if imagining movement were enough to control a machine? Building a neural interface that works without an implant is difficult. Brain signals weaken by the time they reach a headset and become mixed with electrical activity from facial movements and nearby electronics. Implants get a cleaner reading because they sit closer to the brain. Synaptrix uses an EEG headset to capture brain activity linked to imagined movement. Its large motor models decode intended movement from those signals and turn it into commands for a device. The first application is wheelchair control. Synaptrix says users can steer by imagining where they want to go. It also says its signal-processing accuracy and denoising performance have held up on busy New York streets. For non-invasive control to be useful in daily life, it has to work beyond a shielded lab. Synaptrix says it has already brought the same system to freely moving cursor control. Steering a wheelchair and moving a cursor require different commands, but both begin with decoding intended movement. If the models can handle that intent across more people and settings, prosthetics and other devices could follow. Wheelchair use gives Synaptrix real world data on how intended movement appears across people and conditions. As more headsets and applications come online, that dataset can widen and the models trained on it can improve. Better decoding could make the system useful on still more devices. The wheelchair shows what the system can do today. The data and models built through real world use could determine how far it goes. (Original content provided by @august_wstein & @DrewAHenderson at Delphi Ventures)