Figure 4.

Feature selection and performance of machine learning in atlas method. (A) Mean square error (MSE) of each feature set in atlas method. The minimum MSE was 401.5 when the first two features (according to absolute value of ${\beta }_i$) were used in the training set via feature selection. B, Correlation between actual and predictive VHI scores. Acceptable results were achieved with an r value of .2721 and an R² value of .0741, which were lower than the values of the ROI method. C, Distribution of R² via permutation test in ROI method. A significant P value (P < .05) was achieved via permutation test