Abstract
Motor Imagery signal is one of the brain signals that generated in the brain during moving status or when imagine a movement. It is one of the famous and most challenging research area in Brain Computer Interaction field. So, exploring a new combination of algorithms yields in improving this area more. In this paper, we try to find an effective classification method to classify Motor Imagery signals into two classes 'left hand' and 'right hand' movements with high accuracy. In this work we used Least Square Support Vector Machine classifier after choose its optimal parameters using Chaotic Particle Swarm Optimization search algorithm. The proposed algorithm has been tested on Graz data set III (Motor Imagery signals). The results indicate that the proposed approach produced good classification algorithm with high performance and accuracy up to 90%. The results show that it is a competitive classification method compared with other studies.