Estimation ofUPDRSFinger Tapping Scorebyusing Artifi'cial Neural Network forQuantitative Diagnosis ofParkinson's disease

Thepurposeofthisstudywastoestimate UPDRSfinger tapping score withParkinson's disease patients by usingan artificial neuralnetwork. The measurement system wascomposed ofapairof3-axis accelerometers, apair oftouchsensors, ananalog todigital converter andapersonal computer. Firstly, theaccelerations during thefinger tapping weremeasured with44normalsubjects and17Parkinson's diseases subjects byusingthissystem. Secondly, thefour features wereextracted fromtheobtained accelerations. Finally, theUPDRS finger tapping scorewasestimated by using athree-layer artificial neural network model. KeywordsArtificial NeuralNetwork, finger tapping, UPDRS,Parkinson's disease, acceleration