Categorizing imagined right and left motor imagery BCI tasks for low-cost robotic neuroprosthesis

Focusing on low-cost articulation control for neuroprosthesis, electroencephalography (EEG)-based brain computer interfaces require rapid and reliable discrimination of EEG patterns associated with motor imagery generated via imagined or real movement. The objective of this study was to characterize EEG signals of two different motor imagery tasks used to control a robotic articulator. With one-sided hand movement imagination resulting in EEG changes located contra and ipsilateral areas, time-courses of two different imagery tasks were investigated via instantaneous band power changes. We compared the features extracted from the EEG patterns with standard machine learning algorithms. We report frequency-based categorization of visualized imagery more relevant than machine learning methods.