Task-dependent recruitment across ankle extensor muscles and between mechanical demands is driven by the metabolic cost of muscle contraction

Abstract

The nervous system is faced with numerous strategies for recruiting a large number of motor units within and among muscle synergists to produce and control body movement. This is challenging, considering multiple combinations of motor unit recruitment may result in the same movement. Yet vertebrates are capable of performing a wide range of movement tasks with different mechanical demands. In this study, we used an experimental human cycling paradigm and musculoskeletal simulations to test the theory that a strategy of prioritizing the minimization of the metabolic cost of muscle contraction, which improves mechanical efficiency, governs the recruitment of motor units within a muscle and the coordination among synergist muscles within the limb. Our results support our hypothesis, for which measured muscle activity and model-predicted muscle forces in soleus-the slower but stronger ankle plantarflexor-is favoured over the weaker but faster medial gastrocnemius (MG) to produce plantarflexor force to meet increased load demands. However, for faster-contracting speeds induced by faster-pedalling cadence, the faster MG is favoured. Similar recruitment patterns were observed for the slow and fast fibres within each muscle. By contrast, a commonly used modelling strategy that minimizes muscle excitations failed to predict force sharing and known physiological recruitment strategies, such as orderly motor unit recruitment. Our findings illustrate that this common strategy for recruiting motor units within muscles and coordination between muscles can explain the control of the plantarflexor muscles across a range of mechanical demands.

Keywords: human; motor unit recruitment; muscle; musculoskeletal modelling.

Figure 1.

(a) Musculoskeletal model of the leg used in the study. The model consisted of 10 segments, 19 d.f. and was actuated by six muscle–tendon units (MTUs) and one ideal hip actuator. (b) Schematic of the two-element Hill-type muscle model consisting of slower (CE_slow) and faster (CE_fast) contractile elements, a parallel elastic element (PEE) and a series elastic element (SEE). The Hill-type muscle model was used to compute tendon (F_t) and muscle force (F_m), accounting for pennation angle (β). (c) Excitation-to-activation dynamics of a square wave muscle excitation (i), active and passive force–length (F-L) curves (ii), the active force–velocity (F-V) curves (iii) and the mechanical efficiency (η) (iv) of the slower fibres (red) and faster fibres (blue) assigned in the muscle model. Mechanical efficiency was computed as mechanical work over metabolic work. Metabolic work was predicted using a metabolic model [20] adapted to account for the efficiency of different muscle fibre types [8]. F-L and F-V curves were normalized to optimal fibre length ($l_o^m$) and maximum contraction velocity (V_max), respectively, as well as maximum isometric force ($F_o^m$).

Figure 2.

Measured electromyography (EMG) intensities (a), model-predicted muscle activation (b) of the medial gastrocnemius (MG) and soleus (SO) and model-predicted muscle fibre activation (c) of the slow and fast muscle fibres in the MG and SO for all cyclists during four boundary pedalling conditions. EMG intensities were normalized to the maximum EMG signal obtained during either a maximum power output or a maximum cadence trial performed separately. Fibre activations were obtained from fibre-specific predicted excitations using a modelling strategy that aimed to minimize the metabolic cost of the muscle contractions (J_meta) and were normalized between 0 (no activation) and 1 (fully activated). Total muscle activation was the summed activations of the slower and faster fibres in each muscle divided by their respective fibre-type proportions and was also normalized between 0 (no activation) and 1 (fully activated). Note that the activations for the fast fibres in the SO are very low for every condition.

Figure 3.

Mean measured electromyography (EMG) intensities of the soleus (SO) and medial gastrocnemius (MG) for all cyclists across eight pedalling conditions. EMG intensities were normalized to the maximum EMG signal obtained during either a maximum power output or a maximum cadence trial performed separately. Mean EMG intensities were computed as the difference between the EMG intensities and the mean EMG intensity during 60 r.p.m. at 13 N m.

Figure 4.

Mean predicted total muscle force developed by the soleus (SO) and medial gastrocnemius (MG) during the crank cycle using two modelling strategies for all cyclists across eight pedalling conditions. J_meta was the strategy to minimize primarily the metabolic cost of the muscle contractions, and J_exc was a strategy to minimize the sum of muscle excitations. Muscle force was normalized to body weight (BW). The error bars represent the standard deviation across all cyclists for each pedalling condition.

Figure 5.

Force–force loops generated with time-varying soleus (SO) and medial gastrocnemius (MG) predicted total force plotted against each other for all cyclists with increases in crank torque (a) and pedalling cadence (b). The arrows denote the direction of the time-varying pattern through a crank cycle. BW, body weight.

Figure 6.

Mean predicted muscle fibre force developed by the slower and faster muscle fibres in the medial gastrocnemius (a) and soleus (b) during the crank cycle using two modelling strategies for all cyclists across eight pedalling conditions. J_meta was the strategy to minimize primarily the metabolic cost of the muscle contractions, and J_exc was a strategy to minimize the sum of muscle excitations. Muscle fibre force was normalized to body weight (BW).

Figure 7.

Force–force loops generated with time-varying predicted slower and faster muscle fibre force in the medial gastrocnemius plotted against each other for all cyclists with increases in crank torque (a) and pedalling cadence (b). Muscle fibre force was normalized to body weight (BW). The arrows denote the direction of the time-varying pattern through a crank cycle.