Automatic, unsupervised classification of dyskinesia in patients with Parkinsons Disease.

Abstract — One of the characteristic symptoms of patients with Parkinson Disease (PD) is a rigidity of movement. These symptoms disappear after administration of Levodopa. However, the long-term use of levodopa causes involuntary movements (dyskinesia). A proper diagnosis requires an automatic, unsu-pervised method for the detection and classification of levodopa induced dyskinesia. The main problem, however, is to distinguish dyskinesia from voluntary movements. The aim of this study was to train a neural network for the classification and rating of dyskinesia and to use the trained neural networks to extract parameters, which are important to distinguish between dyski-nesia and normal voluntary movements. The neural network has a performance near 97 % correct, which greatly improves upon previous methods (typically 75 % correct). I