Individual Differences in Motor Imagery BCIs: a Study of Gender, Mental States and Mu Suppression

Abstract

Performance variation among users is a well-known issue in the Brain-Computer Interface (BCI) field. Although extensive research has been conducted, there is yet no clear agreement on the specific factors that influence the BCI performance. Using a large sample size of 54 subjects, this study investigated the impact of different demographic and psychological variables including gender, attention, relaxation, motivation, visual fatigue, and somnolence on motor imagery BCI (MI-BCI) performance. A comprehensive approach was taken in evaluating the MI-BCI performance, by both looking at online classification accuracies as well as the mu rhythm suppression from the recorded electroencephalography (EEG) signals. In line with previous studies, our results suggest that there is a significant difference of BCI performance between users, depending on their relaxation level and (visual) fatigue. Additionally, gender was found as a predictor showing an advantage for female participants in their MI-BCI control abilities. Our findings suggest the need for BCI training and instruction protocols that take into consideration the psychological states and background information of the users.