Robust filter bank common spatial pattern (RFBCSP) in motor-imagery-based brain-computer interface

The filter bank common spatial pattern (FBCSP) algorithm performs autonomous selection of key temporal-spatial discriminative EEG characteristics in motor imagery-based brain computer interfaces (MI-BCI). However, FBCSP is sensitive to outliers because it involves multiple estimations of covariance...

Full description

Saved in:
Bibliographic Details
Published inConference proceedings (IEEE Engineering in Medicine and Biology Society. Conf.) Vol. 2009; pp. 578 - 581
Main Authors Ang, Kai Keng, Chin, Zheng Yang, Zhang, Haihong, Guan, Cuntai
Format Conference Proceeding Journal Article
LanguageEnglish
Published United States IEEE 01.01.2009
Subjects
Algorithms
Brain computer interfaces
Brain Mapping - methods
Covariance matrix
Electroencephalography
Electroencephalography - methods
Evoked Potentials, Motor - physiology
Filter bank
Humans
Imagination - physiology
Machine learning algorithms
Motor Cortex - physiology
Pattern Recognition, Automated - methods
Pollution measurement
Reproducibility of Results
Robustness
Sensitivity and Specificity
Signal processing algorithms
Signal Processing, Computer-Assisted
Testing
Training data
User-Computer Interface
Online AccessGet full text

Cover

More Information
Summary:The filter bank common spatial pattern (FBCSP) algorithm performs autonomous selection of key temporal-spatial discriminative EEG characteristics in motor imagery-based brain computer interfaces (MI-BCI). However, FBCSP is sensitive to outliers because it involves multiple estimations of covariance matrices from EEG measurements. This paper proposes a Robust FBCSP (RFBCSP) algorithm whereby the estimates of the covariance matrices are replaced with the robust minimum covariance determinant (MCD) estimator. The performance of RFBCSP is investigated on a publicly available dataset and compared against FBCSP using 10times10-fold cross-validation accuracies on training data, and session-to-session transfer kappa values on independent test data. The results showed that RFBCSP yielded improvements in certain subjects and slight improvement in overall performance across subjects. Analysis on one subject who improved suggested that outliers were excluded from the robust covariance matrices estimation. These results revealed a promising direction of RFBCSP for robust classifications of EEG measurements in MI-BCI.
ISSN:1094-687X
1557-170X
DOI:10.1109/IEMBS.2009.5332817