Table 6.
A comparative analysis of the previous research and the proposed work.
| Reference | Study Setting | Classifier | Accuracy |
|---|---|---|---|
| Liu et al. [28] | 0-, 1-, 2- N-back | LDA | 64.4% (EEG) 55.6% (fNIRS) 68.1% (EEG+fNIRS) |
| Aghajani et al. [10] | 0-, 1-, 2-, 3- N-back | SVM | 85.9% (EEG) 74.8% (fNIRS) 90.9% (EEG+fNIRS) |
| Nguyen et al. [38] | Simulated driving system | FLDA | 73.7% (EEG) 70.5% (fNIRS) 79.2% (EEG+fNIRS) |
| Saadati et al. [29] | N-back DSR Word generation LHand vs. RHand |
DNN, SVM | 67.0% (EEG-DNN) 80.0% (fNIRS-DNN) 87.0% (EEG+fNIRS-DNN) 82% (EEG+fNIRS-SVM) |
| Chu et al. [39] | Mental workload | SVM, RF, DT | 55.4% (EEG-RF) 69.2% (fNIRS-RF) 78.3% (EEG+fNIRS-RF) |
| Proposed study | 0-, 2-, 3-back | SVM | 77% (0-back vs. 2-back) 83% (0-back vs. 3-back) 59% (2-back vs. 3-back) |
Abbreviations: LDA—Linear discriminant analysis; SVM—Support Vector Machine; FLDA—Fisher Linear Discriminant Analysis; DNN—Deep Neural Network; RF—Random Forest; DT—Decision Tree; DSR—discrimination/selection response task.