Application of Artificial Neural Network on Speech Signal Features for Parkinson’s Disease Classification

Since speech is affected in up to 90% of Parkinson’s Disease patients, it has drawn interest as a target symptom for diagnostic testing. Using a dataset of speech sample features, multilayer perceptron was trained to accurately classify between Parkinson’s Disease patients and healthy individuals. By optimizing the number of neurons in a single hidden layer network, a 0.937 accuracy model was identified. Using the single hidden layer results to inform a search for deep learning models with two hidden layers, accuracy improved to 0.952. Finding high accuracy and often 100% sensitivity, this method could be a very useful screening test for Parkinson’s Disease as it is also non-invasive, quick, and can be done remotely.