EEG-based Discriminative Features During Hand Movement Execution and Imagination

Electroencephalogram (EEG) is one the most commonly used brain activity measurement tools in Brain-Computer Interface (BCI) framework to establish “thought” based human-machine interaction. On account of the fact that mental rehearsal of motor movement termed motor imagery, is a robust candidate for developing BCI systems, it is of great importance to find the distinction between the neural correlates of executed motor movement and its imagination, to enhance command signal generation and utilization in BCI-based communication and control. This paper investigates the EEG-based parameters associated with execution and imagination of left and right hand movement from a set of 10 healthy subjects. The study explores the distinguishable power and phase features in theta (4–8 Hz), alpha/mu (8–12 Hz), lower beta (12–20 Hz), higher beta (20–30 Hz) and gamma (30–40 Hz) bands of EEG. It is observed that the most informative channels that effectively discriminates motor execution and imagination varies across bands and tasks. Frontal and parietal channels are found to be more discriminative compared to those in central region in both tasks. Combining the task-specific discriminative features along with Common Spatial Pattern (CSP) based spatial features, an average classification accuracy of 81% is achieved in $5 \times 5$ -fold cross validation, over 10 subjects. Further study is essential to investigate the temporal structure of the distinguishable features, to generalize the discriminative spectral and spatial characteristics and to improve the classification accuracy.