Fig. 5

Classification of ALS patients based on FA in the corticospinal tracts. The Tract Profiles in the two CSTs are submitted to a Principal Components Analysis—the first two PCs form the dimensions of this plot (accounting for about 50% of the variance in the data). The data are separately used to train a support vector machine classifier, with a polynomial kernel. The classification boundary is shown here in the space of first two PCs. This classifier performs at 88% accuracy (cross-validated) in discriminating patients from controls. The Jupyter Notebook containing all steps of the analysis is shared here: https://github.com/YeatmanLab/AFQ-Browser_data/blob/master/AFQ-Browser_ALSexample/Figure6.ipynb