A Machine Learning Method to Process Voice Samples for Identification of Parkinson’s Disease

Abstract Machine learning approaches have been used to develop methods for the automatic detection of Parkinson’s Disease with voice recordings being the most used data type due to the simple and non-invasive nature of acquiring data. While most studies used voice samples recorded under controlled conditions, a few studies have used voice samples acquired via telephone. Although voice recordings captured via telephone or mobile devices allow much easier and wider access for data collection, current conflicting performance results limit their clinical applicability. The contribution of this study is two-fold: First, we show the reliability of telephone-collected voice recordings of the sustained vowel /a/ by collecting samples from 50 people with Parkinson’s Disease and 50 healthy controls and applying machine learning classification with voice features related to phonation. Second, we utilize a novel application of a convolutional neural network with transfer learning to analyze the spectrograms of the sustained vowel from these samples. This approach considers speech intensity estimates across time and frequency scales rather than collapsing measurements across time., We show the superiority of this pre-trained Inception V3 convolutional neural network model with transfer-learning for the task of classifying people with Parkinson’s Disease as distinct from healthy controls.