Parkinson’s Disease Detection Using Multiple Speech Signals

Parkinson's disease (PD)is the most common progressive neurodegenerative disorder. The symptoms of PD are classified into motor and non-motor. The non-motor symptoms are difficult to recognize in the early stages. So, the research is started to recognizing motor symptoms. One of the motor symptoms is a speech disorder. This paper deals with the detection of parkinson's disease using speech signals. The approach evaluates the use of two classifiers to classify PD from Healthy data obtained from two vocal tasks using acoustic cardioid (AC) and smartphone (SP) microphone channels. The dimensionality reduction is used to select the best features among all the features that are extracted from the speech. The performance of the classification is evaluated by calculating accuracy, precision, recall, and area under curve (AUC) measures from receiver operating characteristic curves. The best performance AC channel recorded phonation sample is giving with an approximate accuracy of 86%, AUC of 0.8436.