Author(s): FENNELL, C., MAUGER, A., HOPKER, J., Institution: UNIVERSITY OF KENT, Country: UNITED KINGDOM, Abstract-ID: 859

INTRODUCTION:
Torque complexity (the torque signal’s nonlinear temporal and spatial structure) is indicative of the neuromuscular system’s ability to meet the imposed motor task, and knee extensor (KE) torque complexity has been shown to decline with fatigue (1). However, the mechanisms underlying the fatigue-induced changes in complexity are yet to be fully elucidated. It has previously been demonstrated that common synaptic inputs to the motoneuron pool, estimated by intra-muscle coherence, largely determine the structure of oscillations in the torque signal (2). This study investigated whether the proportion of common synaptic input to the vastus lateralis muscle (VL) could explain the effect of fatigue on KE torque complexity.
METHODS:
Fifty-seven participants completed three fresh 20 s isometric KE contractions at 20% MVC, followed by a series of repeated 3 s isometric KE contractions at 60% MVC to task failure. Immediately after failure a 20 s isometric KE contraction at 20% MVC was performed. High-density surface EMG signals were recorded from the VL during all 20% MVCs and decomposed into individual motor unit (MU) spike trains. Thirty-seven participants provided enough MUs (>6) during all 20% MVCs for coherence analysis, as such only these participants data were analysed. Intra-muscle coherence was estimated from the MU data. KE torque variability and complexity was assessed using the coefficient of variation of torque (CVT) and multiscale sample entropy (MSE; across 28 coarse-grained scales) respectively.
RESULTS:
The MSE analysis revealed a condition (fresh vs. fatigue) by coarse-grained scale interaction (i.e., a MSE curve cross-over; P<0.001); the fatigued 20% MVCs presented with a significantly lower sample entropy (SampEn; P<0.05) at shorter scales (<7 scales) but a significantly higher SampEn (P<0.05) at longer scales (>13 scales) when compared to the fresh 20% MVCs. Alpha coherence (5-15Hz) was significantly higher after task failure (P=0.001; fresh=1.30±0.09 vs. fatigued=1.43±0.22). The difference in Alpha coherence from fresh to fatigue was predictive of the difference in SampEn at scales 10 to 28 (all P<0.05; R2=0.11-0.19). Alpha coherence was also predictive of SampEn calculated at coarse-grained scales 7 to 28 during fresh (all P<0.05; r=0.34-0.60; R2=0.12-0.36) and fatigued (all P<0.05; r=0.34-0.52; R2=0.11-0.27) 20% MVCs. Fatigue resulted in a significantly higher CVT (P<0.001). The increase in alpha coherence with fatigue was predictive of the decline in the CVT (P<0.001; r=0.56; R2=0.31).
CONCLUSION:
The MSE curve cross-over captures a fatigue-related change in torque regularity at the different scales, suggestive of an alteration in structure across the oscillatory frequencies present within the torque signal. Common synaptic input in the alpha band may be able to explain some of the fatigue-related alterations in the structure of lower-frequency oscillations of the KE torque signals.

1)Pethick et al. J Physiol 2015
2)Farina & Negro. Exerc Sport Sci Rev 2015