Both the differential diagnosis of parkinsonian syndromes and the prediction of gait dysfunction may eventually be improved by automated machine learning tools, two studies suggest.