Comparative analysis of the optimal performance evaluation for motor imagery based EEG-brain computer interface

Abstract

The purpose of this paper is to evaluate the performance of EEG-BCI(Brain Computer Interface) based on motor imagery characteristics for a ubiquitous health service, brain computer interface technology referring to a technique to control an external device using the brain signals without other expressions. The EEG-BCI algorithm used in this paper is composed of a common spatial pattern (CSP) and a least square linear classifier. The CSP is used to obtain the characteristics of event-related desynchronization, and the least square linear classifier classifies the motor imagery EEG data of the left hand or right hand. The effect of a performance factor is important to evaluate the optimal performance for a motor-imagery-based EEG-BCI algorithm. There are five performance factors; the EEG mode, feature calculation, selected CSP channel number, selected classifier and window size. © 2011 Springer-Verlag.