Fig 1.
Three types of ‘prime-word—image’ pairs.
Fig 2.
Experimental sequence for a single-trial consisting of a blank screen, fixation mark, prime stimulus, mask, main stimulus, and response box.
Table 1.
The p-values obtained from paired-samples t-test performed over the average response scores corresponding to positive-negative (Pos-Neg), positive-neutral (Pos-Neu), and negative-neutral (Neg-Neu) pairs.
Fig 3.
Grand ERP average for positive (in red), negative (in green), and neutral (in blue) prime affect types.
Table 2.
The one-way repeated measures ANOVA test results for the windowed average ERPs with 25 ms analysis window that yield lowest p-values.
Table 3.
The one-way repeated measures ANOVA test results for the windowed average ERPs with 25 ms analysis window that yield lowest p-values.
Table 4.
Confusion matrix for a multiclass classifier.
Table 5.
Confusion matrix for a positive-negative binary classifier.
Table 6.
Binary SVM classifier performance for average-ERP data.
Table 7.
Binary AdaBoost classifier performance for average-ERP data.
Table 8.
Performance of the multiclass classifiers for average ERP data.
Fig 4.
Multiclass SVM and AdaBoost performance on randomly permuted data.
Table 9.
Performance of subject-independent classifiers for single-trial ERP data.
Table 10.
Performance of the subject-dependent classifiers for single-trial ERP data.
Fig 5.
Single-trial ERP classification results (SVM (left) and AdaBoost (right)) of individual subjects (subject#1 to subject#40).
Table 11.
Binary SVM and AdaBoost classifier performance for average-ERP data using identical input features.
Table 12.
Performance of the multiclass classifiers for average ERP data using identical input features.
Table 13.
Single-trial subject dependent and subject independent SVM classifier performance on identical input features.
Table 14.
Single-trial subject dependent and subject independent AdaBoost classifier performance on identical input features.