Fig. 4

Correlation between the predicted values generated by the models and the actual values for each cognitive domain under examination. The predictions from the model with the lowest mean absolute error (MAE) were represented. The values depicted in the figure correspond to the predictions made on the test set from a ten-fold cross-validation. To minimize the influence of outliers in the representation, the scales of both the X and Y axes were adjusted considering the 95th percentile of the true values. Abbreviations: \(\rho\), correlation between model predictions (\(\hat{Y}\)) and true values (Y); R\(^2\), coefficient of determination; RF, random forest; XGB, extreme gradient boosting