Table 2.
Comparisons with different methods.
| Reference | Approach | Signal | Task | Classifier | Accuracy |
|---|---|---|---|---|---|
| Zhang et al. [12] | Binarized brain network (based on time series during task and phase synchronization and Pearson correlation algorithms) metric features | EEG | Ug-vs-Tg | SVM | # |
| Ug-vs-Sc | 55.00% | ||||
| Tg-vs-Sc | # | ||||
| fNIRS | Ug-vs-Tg | SVM | # | ||
| Ug-vs-Sc | 51.60% | ||||
| Tg-vs-Sc | # | ||||
| EEG + fNIRS | Ug-vs-Tg | SVM | # | ||
| Ug-vs-Sc | 58.2% | ||||
| Tg-vs-Sc | # | ||||
|
| |||||
| Zhang et al. [14] | Binarized brain network (based on ERP components and WPLI algorithm) metric features | EEG | Ug-vs-Tg | SVM | 50.00% |
| Ug-vs-Sc | 64.83% | ||||
| Tg-vs-Sc | 69.67% | ||||
|
| |||||
| This paper | Improved DSP | EEG | Ug-vs-Tg | KNN | 63.80% |
| Ug-vs-Sc | 64.67% | ||||
| Tg-vs-Sc | 59.54% | ||||
The Ug-vs-Tg, Ug-vs-Sc, and Tg-vs-Sc denote binary classification. The symbol “#” represents there is no classification task. fNIRS is the functional near-infrared spectroscopy.