TABLE 2.

Classification results in terms of “mean(standard deviation)” achieved by 10 methods in two cross‐site MDD versus HC classification tasks on the REST‐MDD (i.e., Site 20→Site 21, and Site 20→Site 1), with best results shown in bold.

Method	Site 20→Site 21						Site 20→Site 1
	ACC	AUC	REC	PRE	F1	p‐value	ACC	AUC	REC	PRE	F1	p‐value
PCA + SVM	0.53	0.52	0.64	0.56	0.60	*	0.54	0.56	0.59	0.54	0.56	*
DC + SVM	0.53	0.53	0.65	0.57	0.61	*	0.57	0.61	0.59	0.57	0.58	*
EC + SVM	0.55	0.53	0.69	0.57	0.62	*	0.53	0.53	0.55	0.53	0.54	*
CC + SVM	0.54	0.52	0.59	0.58	0.59	*	0.55	0.55	0.53	0.56	0.54	*
BC + SVM	0.53	0.50	0.66	0.56	0.61	*	0.52	0.52	0.59	0.52	0.55	*
CNN	0.58(0.09)	0.62(0.10)	0.62(0.23)	0.65(0.11)	0.60(0.15)	*	0.58(0.09)	0.59(0.12)	0.55(0.15)	0.61(0.11)	0.56(0.09)
GCN	0.57(0.06)	0.63(0.09)	0.50(0.14)	0.64(0.03)	0.55(0.10)	*	0.58(0.10)	0.66(0.13)	0.54(0.24)	0.61(0.12)	0.54(0.17)	*
ST‐GCN	0.60(0.16)	0.64(0.18)	0.61(0.13)	0.65(0.17)	0.63(0.15)		0.61(0.09)	0.68(0.11)	0.56(0.15)	0.61(0.10)	0.58(0.13)	*
NO PRETEXT	0.58(0.11)	0.60(0.09)	0.66(0.14)	0.62(0.08)	0.63(0.09)		0.55(0.17)	0.57(0.21)	0.56(0.28)	0.51(0.18)	0.53(0.23)
UCGL (Ours)	0.63(0.11)	0.65(0.15)	0.63(0.15)	0.68(0.08)	0.65(0.11)	‐	0.62(0.14)	0.66(0.16)	0.60(0.17)	0.62(0.13)	0.61(0.15)	‐

Note: “*” indicates that the UCGL is statistically significantly different from a competing method. The term “S→T" denotes that a model is pre‐trained/trained on a source domain S and fine‐tuned/tested on a target domain T.