Table 4.

Mean classification performance of multimodal features.

Classifier	Extreme learning machine								SVM-L	SVM-RBF	LDA	RF
Feature type	Train Acc	Test Acc	p-value	Sn	Sp	F1-Score	PPV	NPV	Test Acc	Test Acc	Test Acc	Test Acc
Hybrid WC	0.9954	0.9929	0.0001	1.000	0.9857	0.9933	1.000	0.9875	0.7780	0.763	0.6810	0.7080
Simple WC	0.9977	0.9804	0.0001	0.9732	0.9875	0.9790	0.9778	0.9875	0.7500	0.750	0.6880	0.7010
Simple concatenation	0.9931	0.9724	0.0001	0.9607	0.9857	0.9723	0.9625	0.9875	0.7710	0.750	0.6880	0.7500
Cortical WC	0.9899	0.9367	0.0001	0.9429	0.9321	0.9359	0.9500	0.9375	0.5960	0.611	0.5830	0.5490
Subcortical WC	0.9931	0.9248	0.0001	0.9571	0.8857	0.9296	0.9639	0.9128	0.6670	0.667	0.6250	0.6040
Cortical concatenation	0.9938	0.9381	0.0001	0.9303	0.9464	0.9367	0.9403	0.9528	0.6320	0.638	0.6110	0.6110
Subcortical concatenation	0.9908	0.9314	0.0001	0.9196	0.9446	0.9309	0.9260	0.9514	0.6390	0.681	0.6180	0.6460
Functional concatenation	0.9961	0.9452	0.0001	0.9304	0.9625	0.9426	0.9375	0.9607	0.6740	0.729	0.6880	0.6740
Structural concatenation	0.9907	0.9319	0.0001	0.9304	0.9357	0.9308	0.9389	0.9417	0.6460	0.625	0.6600	0.6460

Sn, Sensitivity; Sp, Specificity; SVM-L, linear support vector machine; SVM-RBF, support vector machine with radial basis function kernel, LDA, linear discriminant analysis; WC, Weighted Concatenation; Acc, Accuracy; PPV, positive predictive value; NPV, Negative predictive value; RF, Random Forest.