Percentage values of each classifier for each criterion: accuracy (ACC), which represents the percentage of correct classification; recall (TPR), which represents the probability that an actual positive will test positive (positives are PD in PD vs. HC, OFF in OFF vs. HC and ON in ON vs. HC and OFF vs. ON); precision (PPV), which represents the percentage of true positive with respect to all positive; goodness index (G), which represents the Euclidean distance between the evaluated point in the ROC space and the point [0 1], representing the perfect classifier. Values exceeding the threshold and passing to the next selection are shown in green, those not exceeding the threshold are shown in white and values not passed are shown with “/”.