Analysis of Reaching Movements at Different Speeds using Recurrence Quantification Analysis and Nonlinear Quantifiers

Using nonlinear signal processing methods is critical in processing biological signals due to their nonlinear dynamics. Recurrence plots are one of these nonlinear methods that provide qualitative and graphical representation of inherent dynamic of signal. Reaching movement is one of the important skill movements during human life. Despite of nonlinear methods capability to analyze the electromyogram signals during reaching movement, these methods are less considered. Therefore, the current manuscript investigates the classification of reaching movements at different speeds in horizontal plane. To achieve this, some quantitative indicators of recurrence plot analysis and nonlinear quantifiers including Lyapunov exponent and Higuchi fractal dimension are used. Based on multivariate analysis of variance, most discriminative features in the separation of different speeds of reaching movement are selected. Results show Recurrence rate, determinism, laminarity and Higuchi fractal dimension are best indicators to describe the recorded signals. The accuracy of KNN is 96.67%, SVM is 100%, linear discriminant analysis is 100%, and decision tree is 90%.