Fig 1.
Block-diagram of the proposed method.
Fig 2.
Comparison between classification accuracy of methods for two-class problem in dataset IIIa, BCI competition III.
Fig 3.
Comparison between classification accuracy of methods for two-class problem in dataset IIa, BCI competition IV.
Fig 4.
Comparison between classification accuracy of methods for two-class problem in dataset IVa, BCI competition III.
Fig 5.
Comparison between classification accuracy of methods for multi-class problem in dataset IIIa, BCI competition III.
Fig 6.
Comparison between classification accuracy of methods for multi-class problem in dataset IIa, BCI competition IV.
Fig 7.
Boxplot of all subject for CSP, TRCSP and CCSP in two-class problem.
Fig 8.
Boxplot of all subject for CSP, TRCSP and CCSP in multi-class problem.
Fig 9.
The scalp topography for the first spatial filter using CSP, TRCSP and CCSP methods for the subjects aw, k3, A06 and A08.
Table 1.
Mean, median and standard deviation (std.) for two and multi-class problem (Best values are in boldface).
Table 2.
Specificity and sensitivity for all of datasets in two-class problem (Best values are in boldface).
Table 3.
Kappa coefficient for dataset IIIa, BCI competition III and dataset IIa, BCI competition IV in multi-class problem (Best values are in boldface).
Table 4.
Running time for two-class and multi-class problems.