FIGURE 3.
ROC curves and AUC values obtained by different methods in (a) cross‐site (i.e., Site 20→Site 21) MDD versus HC classification on the REST‐MDD and (b) cross‐dataset (i.e., ABIDE‐Site NYU→ADNI) AD versus HC classification. 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.