Visualização e classificação de características para a discriminação entre indivíduos saudáveis e com doença de Parkinson sob tratamento com levodopa

Abstract

Parkinson's disease (PD) is one of the most common pathologies among neurodegenerative diseases, its correct and anticipated diagnosis is an important condition for the treatment and reduction of patient costs. To achieve this goal, new strategies for rapid and objective assessment are needed. This work presents an objective method, using inertial sensors and EMG, able to visualize, characterize and differentiate the motor capacities between individuals with PD treated with levodopa and healthy individuals. Data from 6 patients with PD users of levodopa and 5 healthy subjects were analyzed. Individuals performed four tasks (finger claw movement, forefinger movement to the nose, forearm supination and pronation movement, and elbow extension movement). Characteristics of the data were extracted and the Sammon method was used to reduce the dimensionality of the obtained characteristics. To validate and classify the data, the K-fold cross validation method was used. The presented results indicate visual difference between the groups for all the tasks and the proposed methods. The task that obtained the best mean of sensitivity was Task 4 with 86% ± 5% and Task 2 was the one that reached the best specificity average with 83% ± 5%. The methods used in this study may be able to objectively classify individuals with PD and healthy individuals, contributing to and assisting in clinical diagnosis