A Novel Approach to Parkinson's Disease Progression Evaluation Using Convolutional Neural Networks

Abstract

Parkinson's disease (PD) is a devastating disorder with serious impacts on the health and quality of life for a wide group of patients. While the early diagnosis of PD is a critical step in managing its symptoms, measuring its progression would be the cornerstone for the development of treatment protocols suitable for each patient. This paper proposes a novel approach to digital PPMI measures and its combination with spirals drawings to increase the accuracy rate of a neural network to the maximum possible. The results show a well performing CNN model with an accuracy of 1(100%). Thus, the end-users of the proposed approach could be more confident when evaluating the progression of PD. The trained, validated, and tested model was able to classify the PD's progression as High, Medium, or Low, with high sureness.