Figure 4.

System calibration and association with self-test score of depression scale

(A) The ROC of drug-naive MDD vs. HC and treated MDD vs. HC on the REST-meta-MDD dataset.

(B) Boxplot of prediction probabilities output by LGMF-GNN for drug-naive and treated patients. The “ns” notation in the plot signifies that the p-value derived from the t-test is greater than or equal to 0.05, indicating no statistically significant difference of model prediction probabilities between the two patient groups.

(C) Density plots of prediction probabilities output by LGMF-GNN for drug-naive and treated patients.

(D) Calibration curve of two-stage and end-to-end trained LGMF-GNN.

(E) Boxplot of system prediction probability and participants’ Beck Depression Inventory (BDI) self-rating scale scores.