A Passive EEG-BCI for Single-Trial Detection of Changes in Mental State

作者：Myrden Andrew; Chau Tom^*

来源：IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2017, 25(4): 345-356.

DOI：10.1109/TNSRE.2016.2641956

• Brain-computer interfaces
• classification algorithms
• electroencephalography
• pattern analysis
• pattern clustering

Traditional brain-computer interfaces often exhibit unstable performance over time. It has recently been proposed that passive brain-computer interfaces may provide a way to complement and stabilize these traditional systems. In this study, we investigated the feasibility of a passive brain-computer interface that uses electroen-cephalography to monitor changes in mental state on a single-trial basis. We recorded cortical activity from 15 locations while 11 able-bodied adults completed a series of challenging mental tasks. Using a feature clustering algorithm to account for redundancy in EEG signal features, we classified self-reported changes in fatigue, frustration, and attention levels with 74.8 +/- 9.1%, 71.6 +/- 5.6%, and 84.8 +/- 7.4% accuracy, respectively. Based on the most frequently-selected features across all participants, we note the importance of the frontal and central electrodes for fatigue detection, posterior alpha band and frontal beta band activity for frustration detection, and posterior alpha band activity for attention detection. Future work will focus on integrating these results with an active brain-computer interface.

• 出版日期2017-4