Accuracy of a brain-computer interface in subjects with minimal training

A Brain Computer Interface (BCI) allows individuals with severe disabilities to communicate with their external environment, by translating changes in their electroencephalogram (EEG) to control signals. Practical applications of BCIs are limited by the long training time required for achieving high levels of control. We present a pilot study to assess the level of BCI control achieved within a short training period. Using an adaptive algorithm, five healthy subjects were trained to operate a sensorimotor rhythm based BCI, for achieving a simple cursor control task. After an initial screening, users were able to attain an average accuracy of 72% for a two target task and 45% for a four target task, within 10 minutes. The results suggest that practical applications of BCIs might be possible without the need for extensive training.