Real-time target selection based on electroencephalogram (EEG) signal

Abstract

Electroencephalogram (EEG)-based mobile system are getting mainstream although the true potential of EEG signals are yet to be discovered. With the help of a specific control paradigm, the success rate of a mobile system to reach to a certain location could be increased. To further extend the existed control paradigm of EEG-Based mobile robotic system, this paper demonstrated the real time target selection of a wireless mobile robot using only human mind. A unique protocol was developed to mimic a scanning process while at the same time allowing subject to make a selection. Our system utilized only single EEG channel with no subject training. We statistically verify that it is feasible to select a target by manipulating only alpha content of EEG. We also show that it is hard to achieve a stable high performance of synchronous EEG-Based BCI application in one trial with a single frequency band. However, we found that the BCI's performance in term of sensitivity is getting more stable with increase in trial.