Brain Computer Interface

What is a Brain-Computer Interface?

A brain-computer interface (BCI) translates measured brain activity into commands for an external device. Invasive systems use implanted electrodes; non-invasive systems use sensors outside the head, such as scalp electrodes for electroencephalography (EEG). A BCI decodes particular signals for a defined task rather than reading all of a person's thoughts.

Examples of Brain-Computer Interfaces

Robotic Limb Control

In a 2012 study of neurally controlled robotic arms, Hochberg and colleagues showed that two people with long-standing tetraplegia could use motor-cortex signals to perform reaching and grasping actions. One participant used the arm to drink from a bottle. Performance was slower and less accurate than typical able-bodied movement, but the study demonstrated that an implanted interface could translate brain signals into useful actions.

Communication

A BCI can also translate signals into text. In a 2021 handwriting study, researchers decoded attempted handwriting from the motor cortex of one person whose hand was paralyzed. The system converted those signals into text in real time.

These are specific demonstrations of assistive technology.

Non-invasive EEG spelling

In Farwell and Donchin’s 1988 experiment, four healthy volunteers attended to a chosen character in a grid while its rows and columns flashed. Electrodes on the head recorded EEG. The system used a P300 response associated with a relevant flash to identify the intended row and column, then select the character.

This required repeated flashes and attention to the chosen character. The authors estimated a rate of about 2.3 characters per minute under their study conditions; that historical result does not establish how fast other EEG systems can operate. The study tested healthy volunteers, so it did not by itself establish reliable daily communication for people with paralysis.