A novel pattern with high-level commands for encoding motor imagery-based brain computer interface

Abstract

Motor Imagery (MI)-based Brain Computer Interface (BCI) has received increasing attention from academicians for its practicability and convenience. With the development of classification algorithms, many traditional patterns based on low-level commands, such as the movement of hands, feet, tongue, or shoulders, can be decoded quite well. However, when encoded by low-level commands, each body part can perform only one imaginary command. This not only limits the number of tasks that a user's body can encode, but also makes the system less user-centered. In this paper, we proposed a novel pattern based on high-level commands for encoding MI BCI. This pattern combines clockwise and anticlockwise movements of both hands. Compared with a typical traditional pattern based on low-level commands, this pattern shows insignificant differences with regard to distinctiveness and stability for some of the tested subjects (p > 0.05). Thus, the proposed pattern can potentially allow additional tasks by human body without reducing their distinctiveness and stability. The results of this study also provide insights into the creation of high-level commands and improvements toward more intelligent and user-centered systems.