A study of EEG signals associated with intended cursor movement using asymmetry ratio

In this paper, an experiment was done to investigate the EEG signals associated with direction of cursor movement, namely ´up´, ´down´, ´right´ and ´left´ without involving motor imagery. A subject was asked to move the arrow on the computer screen virtually. The arrows were consisted of ´up* ´down´, ´left´ and ´right´. Power spectral density (PSD) asymmetry ratio from the EEG signals was used as input features. The backpropagation neural network (BPN) was used as classifier. Results show that the ´direction´ task can be discriminated from baseline task with over 90% of accuracy. However, they cannot be discriminated from each other.