Extended Data Fig. 5: Spearman coefficients of the CPM prediction for a range of p values (from 0.0001 to 0.05 with an interval of 0.0001).
From: Multimodal brain connectome-based prediction of suicide risk in people with late-life depression
Spearman correlations between the true and predicted values were calculated (two-tailed, unadjusted). This analysis was used to select models with optimal performance using a range of p values and the select models with selected p values were used for the final CPM analysis. Therefore, these p values were not unadjusted. For the functional connectivity positive network, the optimal p thresholds for BSS, SPS and TSII are 0.0013, 0.0001,0.0013; for the negative network, the optimal p thresholds for BSS, SPS and TSII are 0.0057, 0.0014, 0.0004. For the structural connectivity positive network, the optimal p thresholds for BSS, SPS and TSII are 0.0069, 0.0067, 0.0179; for the negative network, the optimal p thresholds for BSS, SPS and TSII are 0.0441, 0.0067, 0.0373. BSS: Beck Scale for Suicidal Ideation; SPS: SAD PERSONS Scale; TSII: Triggers of Suicidal Ideation Inventory. Y axis denotes the Spearman coefficients between the predicted scores and the observed behavioral scores. X axis denotes the different p values.
