FIGURE 1.

Schematic of the machine learning algorithm used in this study. (A) A 10-fold cross validation (CV) approach was used. The training dataset (N = 140) was randomly split into ten “folds.” During each CV run, data from ninefold were combined to form the training subsample while the remaining fold was used as the holdout sample. Operations in the training sub-sample are coded in blue, and those in the holdout sample in green. The 10-fold CV was repeated 10 times for improved stability. (B) Details of the five steps employed in each CV run while predicting a child’s cognitive score. Color coding same as in (A).