PORTLAND, Ore.—The Tools for Brain-computer Interaction (TOBI) project recently reported progress in allowing paralyzed patients to control a computer with their mind, to control a mobile robot that becomes it eyes-and-ears, and to regain some functionality for paralyzed limbs at The École Polytechnique Fédérale de Lausanne (EPFL, Switzerland).
For four years TOBI has been developing in-brain machine interfaces with funding provided by the European Commission under the Seventh Framework Program for Research, which is coordinated by EPFL. More 100 patients with severe motor impairments have participated in the development of core sensor and actuator technologies. The group recently displayed three of their most successful projects, but cautioned that they are not yet ready for prime time.
"Each brain has its own way of transmitting its signals requiring significant calibration," said professor Jose´ del R. Milla´n at the Centre for Neuroprosthetics at EPFL. "A long road is still ahead before turnkey products will be available."
The three showcase demonstrations by TOBI included:
Braintree—electrodes attached to a cap worn by the patient transmit brain signals to electronics that integrates residual muscular activity to allow a cursor to be controlled for texting and surfing the Internet. A proprietary tree-structure maximized the speed and accuracy of the system by allowing only context-aware choices to be made.
Robotino—a small robot with video, audio and obstacle-detection sensors was controlled by the brain-cap, allowing patients to take a virtual walk around the hospital, or even at remote locations where the bed-ridden patient was able to meet with family and loved ones located in other cities.
Regeneration—by combining the brain-cap with electrical stimulation, patients were able to regain control of their paralyzed limbs, and in some cases, were stimulated to regrow nerves that, after intensive training, allowed the patients to retain some of that control even after the electronics was removed (see video below).