Figure 5.

Classification performance on distinguishing ataxic subjects from control subjects. From left to right and top to bottom: normalized confusion matrix (rows show true classification, columns show prediction); ROC curve; test–retest correlation of classification scores, with scores closer to 1 representing the prediction of ataxia and scores closer to 0 the prediction of control; Total BARS score vs. predicted probability of ataxia (>0.5 corresponds to a prediction of ataxia, $\mathrm{x}$ = incorrect prediction).