Table 1 Classification performance of the support vector machine across different frequency bands. All results are reported as the mean (standard deviation) over 10 repetitions. Concat indicates the combination of features across frequency bands and feature types. Frequency bands: δ = Delta, θ = Theta, α = Alpha, β = Beta, γ = Gamma, High-γ = High Gamma. Graph-theoretical measures: CC = clustering coefficient, Eloc = local efficiency, BC = betweenness centrality, D = node degree, EC = eigenvector centrality.
Feature | Frequency band | |||||||
|---|---|---|---|---|---|---|---|---|
Metric | δ | θ | α | β | γ | High-γ | Concat | |
Graph-theoretical measures | Accuracy | 0.5277 (0.0331) | 0.6021 (0.0347) | 0.6007 (0.0401) | 0.6163 (0.0477) | 0.6716 (0.0177) | 0.6851 (0.0180) | 0.7078 (0.0402) |
F1-score | 0.5051 (0.0489) | 0.6018 (0.0310) | 0.6184 (0.0397) | 0.6440 (0.0460) | 0.6865 (0.0193) | 0.6941 (0.0230) | 0.7129 (0.0395) | |
Recall | 0.4831 (0.0735) | 0.5972 (0.0384) | 0.6437 (0.0531) | 0.6901 (0.0556) | 0.7155 (0.0440) | 0.7113 (0.0455) | 0.7211 (0.0503) | |
Precision | 0.5345 (0.0361) | 0.6086 (0.0403) | 0.5967 (0.0395) | 0.6040 (0.0402) | 0.6619 (0.0231) | 0.6795 (0.0198) | 0.7068 (0.0448) | |
AUC | 0.5280 (0.0331) | 0.6022 (0.0348) | 0.6004 (0.0402) | 0.6158 (0.0477) | 0.6713 (0.0177) | 0.6849 (0.0180) | 0.7077 (0.0403) | |
Connectivity weights | Accuracy | 0.5943 (0.0402) | 0.6894 (0.0241) | 0.6482 (0.0327) | 0.6582 (0.0366) | 0.6894 (0.0363) | 0.7695 (0.0188) | 0.7674 (0.0378) |
F1-score | 0.5806 (0.0497) | 0.6894 (0.0342) | 0.6387 (0.0374) | 0.6513 (0.0383) | 0.6948 (0.0360) | 0.7667 (0.0211) | 0.7612 (0.0424) | |
Recall | 0.5634 (0.0843) | 0.6887 (0.0641) | 0.6211 (0.0708) | 0.6352 (0.0506) | 0.7028 (0.0456) | 0.7535 (0.0363) | 0.7394 (0.0578) | |
Precision | 0.6098 (0.0520) | 0.6930 (0.0208) | 0.6663 (0.0521) | 0.6704 (0.0411) | 0.6884 (0.0386) | 0.7819 (0.0235) | 0.7865 (0.0375) | |
AUC | 0.5945 (0.0404) | 0.6894 (0.0239) | 0.6484 (0.0328) | 0.6583 (0.0366) | 0.6893 (0.0363) | 0.7696 (0.0188) | 0.7676 (0.0377) | |
Concat | Accuracy | 0.5688 (0.0320) | 0.6993 (0.0352) | 0.6262 (0.0352) | 0.6582 (0.0263) | 0.7050 (0.0410) | 0.7553 (0.0373) | 0.7901 (0.0298) |
F1-score | 0.5561 (0.0383) | 0.6997 (0.0457) | 0.6202 (0.0374) | 0.6869 (0.0279) | 0.6982 (0.0449) | 0.7568 (0.0304) | 0.7878 (0.0303) | |
Recall | 0.5394 (0.0630) | 0.7014 (0.0763) | 0.6085 (0.0645) | 0.7465 (0.0504) | 0.6803 (0.0633) | 0.7535 (0.0323) | 0.7746 (0.0441) | |
Precision | 0.5795 (0.0382) | 0.7027 (0.0368) | 0.6383 (0.0486) | 0.6375 (0.0235) | 0.7219 (0.0499) | 0.7633 (0.0563) | 0.8036 (0.0398) | |
AUC | 0.5690 (0.0321) | 0.6993 (0.0351) | 0.6264 (0.0353) | 0.6575 (0.0263) | 0.7051 (0.0411) | 0.7553 (0.0375) | 0.7902 (0.0298) | |
