Brain-Machine Interfaces

Brain-machine interfaces (BMIs; also called brain-computer interfaces or neuroprostheses) are devices that connect neural circuits to machines. BMIs hold promise to cure sensory and motor deficits that result from neurological diseases or trauma. Researchers also hope that BMIs will be used to treat cognitive deficits. Currently, the major classes of BMIs are: motor, sensory and bidirectional. Motor BMIs decode motor intentions from brain activity and convert them into movements of artificial actuators, such as computer cursors and robotic limbs. Sensory BMIs deliver artificial sensations from machines to sensory areas of the brain, commonly using intracortical microstimulation. Bidirectional BMIs, also called brain-machine brain interfaces (BMBIs), both extract motor information from the brain and enable artificial sensations.

History

In the 1950s NIH researcher John Lilly implanted multiple electrodes in the cortex of rhesus macaques (Lilly 1956) . 25 to 610 electrodes were packed at intervals from one to two millimeters apart. Lilly stimulated cortical tissue through these electrodes to study the distribution of motor functions across cortical areas.

In the 1960s and 1970s several groups (Kamiya, Black, Sterman and others) provided subject with a biofeedback of their cortical potentials.

The first demonstration of a BMI was probably done by Grey Walter in 1963. Walter recorded electrical activity in the motor cortex of patients undergoing neural surgery. The patients were asked to push a button to advance a slide projector. Walter detected cortical potentials related to that motor act. He then disconnected the button from the projector and used cortical readiness potential advance the slides. The patients continued to push the button. Cortical signals often led the actual hand movements.