A comparison of sEMG temporal and spatial information in the analysis of continuous movements. (2017)

• Record Type: Journal Article
• Authors: Di Girolamo, Michela; Favetto, Alain; Paleari, Marco; Celadon, Nicolò; Ariano, Paolo
• Is Part Of: Informatics in medicine unlocked. Volume 9(2017)
• Page Start: 255
• Page End: 263
• Publication Date: 2017
• Subjects: EMG analysis -- Time -- Movement discrimination -- Continuous movement -- Template matching; TMM Template Making and Matching -- CCDF Complementary Cumulative Distribution Function -- GCA Global Classification Accuracy -- LCA Local Classification Accuracy -- OP Output Percentage; Medical informatics -- Periodicals; 610.285
• DOI: 10.1016/j.imu.2017.10.006
• Languages: English
• ISSNs: 2352-9148

Abstract

Abstract: Much effort has recently been devoted to the analysis of continuous movements with the aim of promoting EMG signal acceptance in several fields of application. Moreover, several studies have been performed to optimize the temporal and spatial parameters in order to obtain a robust interpretation of EMG signals. Resulting from these perspectives, the investigation of the contribution of EMG temporal and spatial information has become a relevant aspect for signal interpretation. This paper aims to evaluate the effects of the two types of information on continuous motions analysis. In order to achieve this goal, the spatial and temporal information of EMG signals were separated and applied as input for an offline Template Making and Matching algorithm. Movement recognition was performed testing three different methods. In the first case (the Temporal approach) the RMS time series generated during movements was the only information employed. In the second case (the Spatial approach) the mean RMS amplitude measured on each channel was considered. Finally, in the third case (the Spatio-Temporal approach) a combination of the information from both the previous approaches was applied. The experimental protocol included 14 movements, which were different from each other in the muscular activation and the execution timing. Results show that the recognition of continuous movements cannot disregard the temporal information. Moreover, the temporal patterns seem to be relevant also for distinguishing movements which differ only in the muscular areas they activate.