|Title||Electroencephalographic(EEG)-based communication: EEG control versus system performance in humans.|
|Publication Type||Journal Article|
|Year of Publication||2003|
|Authors||Sheikh, H, McFarland, DJ, Sarnacki, WA, Wolpaw, J|
|Keywords||augmentative communication, brain-computer interface, brain-machine interface, Electroencephalography, mu and beta rhythms, neuroprosthesis, Rehabilitation|
People can learn to control electroencephalographic (EEG) sensorimotor rhythm amplitude so as to move a cursor to select among choices on a computer screen. We explored the dependence of system performance on EEG control. Users moved the cursor to reach a target at one of four possible locations. EEG control was measured as the correlation (r(2)) between rhythm amplitude and target location. Performance was measured as accuracy (% of targets hit) and as information transfer rate (bits/trial). The relationship between EEG control and accuracy can be approximated by a linear function that is constant for all users. The results facilitate offline predictions of the effects on performance of using different EEG features or combinations of features to control cursor movement.