Home > Publications > Doctoral dissertation & Master thesis
Domestic Conference
Title Hybrid Motor and Speech Imagery BCI Gaming using Minimum Dry Electrodes
Degree MS
Author Roya Sadat Babaei
Advisor Kiseon Kim
Graduation Date 2015.02.25 File
    Date 2017-03-06 10:49

Over the past several decades, research on brain computer interface (BCI) has

focused on developing ways through which disabled people can express themselves and communicate with the physical world. While finding wide aplications in medical field, BCI technology has recently been extended to recreational use such as gaming which allows a wider group of users to experience rich and new means of interaction with a computer or game console. Such an application also serves as a promising and entertaining therapy tool to help improve cognitive focus in people with impairments.

The contribution of this thesis is three-fold. Firstly, we develop a real-time BCI game driven by both, motor (MI) and speech imagery (SI) using dry electrodes. This can be considered as the first BCI-based application controlled byMI and SI combinedly without manual intervention. Furthermore, taking into account the electroencephalographic (EEG) synchrony between brain signals, we propose a paradigm that uses channel pariwise cross-correlation (CC) in parallel with principal component analysis (PCA) or common spatial pattern (CSP) to improve feature extraction. This proposed approach is evaluated and compared for offline and online scenarios. Finally, to enhance

user convenience and reduce time complexity incurred in the game, we then find

the minimal set of electrodes that result in acceptable accuracy of MI and SI-related signals. Experimental results indicate the better performance of CSP along with CC in terms of mean classification accuracy compared to using PCA with CC.

광주과학기술원 한·러 MT-IT 융합기술연구센터 광주과학기술원정보통신공학부