Multitasking with Noninvasive Neuroprosthetics

Most non-invasive brain-machine interfaces (BMI) rely on electroencephalographic (EEG) activity because of its excellent time resolution. Recent progress has shown that online analysis of EEG, if used in combination with advanced robotics and machine learning techniques, is sufficient for humans to control mobile robots, wheelchairs, and hand ortheses. Moreover, smart interaction designs makes it possible for subjects to split their attention between BMI operation and a primary task such as driving or handwriting, thus naturally multitasking.