Motor Performance Index for Evaluation of Distributed Pattern in Multi-channel EEG

Recently, motor imagery (MI) is widely used for brain-computer interface (BCI) for rehabilitation, but it is an ambiguous task and difficult to perform for patients with stroke. To overcome this limitation, MI guidance systems were proposed in previous studies. However, the performance index to evaluate motor performance enhancement of MI guidance system and the motor ability for different motor tasks was focused on magnitude even though they used many kinds of the feature vector. The present study proposes the motor performance index considering the distributed pattern in multi-channel EEG by using the direction for multi-dimensional feature vectors. For validation of the proposed motor performance index, eleven healthy participants' data from our previous study were analyzed through cross-validation. Offline analysis results showed proposed direction-based indexation can be applicable for the evaluation of motor performance to compare different motor tasks. In addition, the present study showed that the MI guidance system can induce a similar activation pattern to motor execution, compared to pure MI.