The non-invasive Berlin Brain-Computer Interface: fast acquisition of effective performance in untrained subjects.

摘要: Brain-Computer Interface (BCI) systems establish a direct communication channel from the brain to an output device. These use signals recorded scalp, surface of cortex, or inside enable users control variety applications. BCI that bypass conventional motor pathways nerves and muscles can provide novel options for paralyzed patients. One classical approach EEG-based is set up system controlled by specific EEG feature which known be susceptible conditioning let subjects learn voluntary feature. In contrast, Berlin (BBCI) uses well established competencies its machine learning extract subject-specific patterns high-dimensional features optimized detecting user's intent. Thus long subject training replaced short calibration measurement (20 min) (1 min). We report results study in 10 subjects, who had no little experience with feedback, computer applications imagination limb movements: these intentions led modulations spontaneous activity specifically, somatotopically matched sensorimotor 7-30 Hz rhythms were diminished over pericentral cortices. The peak information transfer rate was above 35 bits per minute (bpm) 3 23 bpm two, 12 while one could achieve control. Compared other need longer comparable results, we propose key quick efficiency BBCI flexibility due complex but physiologically meaningful adaptivity respects enormous inter-subject variability.