A bi-dimensional index for the selective assessment of myoelectric manifestations of peripheral and central muscle fatigue

Two physiological factors are assumed in this paper to mainly determine the myoelectric manifestations of fatigue: (1) the decrease of the conduction velocity (CV) of motor unit action potentials (MUAP) (peripheral fatigue), and (2) the increase of MU synchronization by the central nervous system (central fatigue). To describe separately the peripheral and central components of the myoclectric manifestations of fatigue, we investigated the following indexes: (1) mean spectral frequency - MNF, (2) median spectral frequency - MDF, (3) root mean square - RMS, (4) average rectified value - ARV, (5) estimation of muscle fiber conduction velocity - ECV, (6) percentage of determinism - %DET, (7) spectral indexes defined as the ratio between signal spectral moments - FIk, (8) MNF estimated by auto-regressive analysis - MNFAR, (9) MNF estimated by Choi-Williams time-frequency representation - MNFCWD, (10) MNF estimated by continuous wavelet transform - MNFCWT, (11) Signal entropy - S, (12) fractal dimension - FD. The indexes were tested with a set of synthetic EMG signals, with different CV distribution and level of MU synchronization. The indexes were calculated on epochs of 0.5 s. It was observed that ECV is uncorrelated with the level of simulated synchronization (promising index of peripheral fatigue). On the other hand FD was the index least affected by CV changes and most related to the level of synchronism (promising index of central fatigue). A representative application to some experimental signals from vastus lateralis muscle during an isometric endurance test supported the results of the simulations. The vector (ECV, FD) is suggested to provide selective indications of peripheral and central fatigue. The description of EMG fatigue by a bi-dimensional vector opens new perspectives in the assessment of muscle properties, with potential application in both clinical and sport sciences. (C) 2008 Elsevier Ltd. All rights reserved.