Authors

• Ignacio Rodríguez Carreño(irodriguezc@unav.es)
• L. Gila Useros, A. Malanda Trigueros, J. Navallas Irujo, J. Rodríguez Falces, S. Gómez Elvira ( )

Abstract
The aim of this work is to analyze the influence of a method for baseline fluctuation (BLF) cancellation for electromyographic (EMG) signals on automatic methods for measurement of the motor unit action potential (MUAP) duration. These methods include four conventional automatic methods (CAMs) and a recently published wavelet transform method (WTM). A set of 182 MUAPs from 170 EMG recordings were studied. The CAMs and the WTM were applied to the MUAPs before and after applying BLF cancellation to the recordings. A gold standard of duration marker positions (GSP) was manually established. The accuracy of each algorithm was estimated as the dfference between its positions and the GSP. Accuracies were compared for the 5 methods and for each method before and after BLF cancellation. A significant difference between accuracy pre- and post-BLF removal was found in two CAMs; markers were closer to the GSP after BLF removal. For all MUAPs, the differences between WTM markers and the GSP were the smallest, and significant differences were not found for the WTM before and after BLF cancellation. The management of BLF is an important issue in EMG signal processing and BLF removal must be considered in extraction and analyse of MUAP waveforms. The BLF removal method improved the performance of two CAMs for MUAP duration measurement. The WTM was the most accurate and was not affected by BLF.

Number of Pages:19

Creation Date:2008-12-19

Number:null/08