Nonlinearities make a difference: comparison of two common Hill-type models with real muscle

Compared to complex structural Huxley-type models, Hill-type models phenomenologically describe muscle contraction using only few state variables. The Hill-type models dominate in the ever expanding field of musculoskeletal simulations for simplicity and low computational cost. Reasonable parameters are required to gain insight into mechanics of movement. The two most common Hill-type muscle models used contain three components. The series elastic component is connected in series to the contractile component. A parallel elastic component is either connected in parallel to both the contractile and the series elastic component (model [CC+SEC]), or is connected in parallel only with the contractile component (model [CC]). As soon as at least one of the components exhibits substantial nonlinearities, as, e.g., the contractile component by the ability to turn on and off, the two models are mechanically different. We tested which model ([CC+SEC] or [CC]) represents the cat soleus better. Ramp experiments consisting of an isometric and an isokinetic part were performed with an in situ cat soleus preparation using supramaximal nerve stimulation. Hill-type models containing force–length and force–velocity relationship, excitation–contraction coupling and series and parallel elastic force–elongation relations were fitted to the data. To test which model might represent the muscle better, the obtained parameters were compared with experimentally determined parameters. Determined in situations with negligible passive force, the force–velocity relation and the series elastic component relation are independent of the chosen model. In contrast to model [CC+SEC], these relations predicted by model [CC] were in accordance with experimental relations. In conclusion model [CC] seemed to better represent the cat soleus contraction dynamics and should be preferred in the nonlinear regression of muscle parameters and in musculoskeletal modeling.     

A cross-bridge model that is able to explain mechanical and energetic properties of shortening muscle.

The responses of muscle to steady and stepwise shortening are simulated with a model in which actin-myosin cross-bridges cycle through two pathways distinct for the attachment-detachment kinetics and for the proportion of energy converted into work. Small step releases and steady shortening at low velocity (high load) favor the cycle implying approximately 5 nm sliding per cross-bridge interaction and approximately 100/s detachment-reattachment process; large step releases and steady shortening at high velocity (low load) favor the cycle implying approximately 10 nm sliding per cross-bridge interaction and approximately 20/s detachment-reattachment process. The model satisfactorily predicts specific mechanical properties of frog skeletal muscle, such as the rate of regeneration of the working stroke as measured by double-step release experiments and the transition to steady state during multiple step releases (staircase shortening). The rate of energy liberation under different mechanical conditions is correctly reproduced by the model. During steady shortening, the relation of energy liberation rate versus shortening speed attains a maximum (approximately 6 times the isometric rate) for shortening velocities lower than half the maximum velocity of shortening and declines for higher velocities. In addition, the model provides a clue for explaining how, in different muscle types, the higher the isometric maintenance heat, the higher the power output during steady shortening.     

From sarcomere to cell: An efficient algorithm for linking mathematical models of muscle contraction

Two classes of mathematical framework have previously been developed to model active tension generation in contracting muscle. Cross-bridge models of muscle are biophysically based but computationally expensive to solve, and thus unsuitable for embedding in spatially distributed continuum representations. Fading memory models are computationally efficient but provide limited biophysical insight. In this study a novel computational method is proposed for coupling these two frameworks such that biophysical events can be determined and computational tractability maintained. Within the cross-bridge model, the functional forms of the distribution of cross-bridges, as a function of strain in each state, are approximated using the distribution moment approach. Using the variables of area, mean and standard deviation of each distribution, analytic expressions are developed to calculate the temporal dynamics of stiffness, tension and energy. A root finding method is employed to adjust the variables such that the temporal dynamics of the cross-bridge model match those of an equivalent fading memory model. The method is demonstrated for sinusoidal perturbations in length at two frequencies, with an approximate 30-fold increase in computational efficiency over a conventional technique for finding a solution to the cross-bridge model.     

Fatigue and non-fatigue mathematical muscle models during functional electrical stimulation of paralyzed muscle

Electrical muscle stimulation demonstrates potential for preventing muscle atrophy and restoring functional movement after spinal cord injury (SCI). Control systems used to optimize delivery of electrical stimulation protocols depend upon the algorithms generated using computational models of paralyzed muscle force output. The Hill–Huxley-type model, while being highly accurate, is also very complex, making it difficult for real-time implementation. In this paper, we propose a Wiener–Hammerstein system to model the paralyzed skeletal muscle under electrical stimulus conditions. The proposed model has substantial advantages in identification algorithm analysis and implementation including computational complexity and convergence, which enable it to be used in real-time model implementation. Experimental data sets from the soleus muscles of 14 subjects with SCI were collected and tested. The simulation results show that the proposed model outperforms the Hill–Huxley-type model not only in peak force prediction, but also in fitting performance for force output of each individual stimulation train.     

Models in which many cross-bridges attach simultaneously can explain the filament movement per ATP split during muscle contraction

Contractile filaments in skeletal muscle are moved by less than 2 nm for each ATP used. If just one cross-bridge is attached to each thin filament at any instant then this distance represents the fundamental myosin cross-bridge step size (i.e. the distance one cross-bridge moves a thin filament in one ATP-splitting cycle). However, most contraction models assume many cross-bridges are attached at any instant along each thin filament. The purpose of this study was to establish whether the net filament sliding per ATP used could be explained quantitatively in terms of a cross-bridge model in which multiple cross-bridges are attached along each thin filament. It was found that the relationship between net filament sliding per ATP split and the load against which the muscle shortens is compatible with such a model and furthermore predicts that the cross-bridge step size is between 7.5 and 12.5 nm over most of the range of loads. These values were similar for different muscle fibre types.     

Optimal posture control of a musculo-skeletal arm model

 In this paper maximal performance posture control of the human arm is investigated by means of model simulations. Recent experiments (F.C.T. van der Helm, submitted, 2000) have shown that the reflexive feedback during postural control varies with the bandwidth of the applied force disturbances. This paper focusses on the influence of the frequency content of force disturbances on the reflexive feedback gains by means of optimization. The arm is modelled by a non-linear musculo-skeletal model with two degrees of freedom and six muscles. To facilitate the optimization of the model parameters, the arm model is linearized. A performance criterion is minimized for stochastic force disturbances in a two-step procedure: (1) optimization of static muscle activations using an additional energy criterion to obtain a unique and energy-efficient solution; and (2) optimization of reflex gains using an additional control effort criterion to obtain a unique solution. The optimization reveals that for the given task and posture, the shoulder muscles have the largest contribution, whereas the bi-articular muscles have a relatively small contribution to the behaviour. The dynamics at the endpoint level are estimated so that a comparison can be made with the experiments. Compared to the experiments, the intrinsic damping of the model is relatively large (about 150%), whereas the intrinsic stiffness is relatively small (about 60%). These differences can be attributed to unmodelled mechanical effects of cross-bridges in Hill-type muscle models. The optimized reflex gains show remarkable similarities with the values found in the experiments, implying that humans can adjust their reflexive feedback gains in an optimal way, weighting the performance and energy. The approach in this paper could be useful in the study of various posture tasks, for example in the prediction of the relation between the control parameters of various musculo-skeletal models and different experimental variables. Received: 24 January 2000 / Accepted in revised form: 7 July 2000     

Phenomenological models of the dynamics of muscle during isotonic shortening

We investigated the effectiveness of simple, Hill-type, phenomenological models of the force-length-velocity relationship for simulating measured length trajectories during muscle shortening, and, if so, what forms of the model are most useful. Using isotonic shortening data from mouse soleus and toad depressor mandibulae muscles, we showed that Hill-type models can indeed simulate the shortening trajectories with sufficiently good accuracy. However, we found that the standard form of the Hill-type muscle model, called the force-scaling model, is not a satisfactory choice. Instead, the results support the use of less frequently used models, the f-max scaling model and force-scaling with parallel spring, to simulate the shortening dynamics of muscle.     

Compliant realignment of binding sites in muscle: transient behavior and mechanical tuning.

The presence of compliance in the lattice of filaments in muscle raises a number of concerns about how one accounts for force generation in the context of the cross-bridge cycle--binding site motions and coupling between cross-bridges confound more traditional analyses. To explore these issues, we developed a spatially explicit, mechanochemical model of skeletal muscle contraction. With a simple three-state model of the cross-bridge cycle, we used a Monte Carlo simulation to compute the instantaneous balance of forces throughout the filament lattice, accounting for both thin and thick filament distortions in response to cross-bridge forces. This approach is compared to more traditional mass action kinetic models (in the form of coupled partial differential equations) that assume filament inextensibility. We also monitored instantaneous force generation, ATP utilization, and the dynamics of the cross-bridge cycle in simulations of step changes in length and variations in shortening velocity. Three critical results emerge from our analyses: 1) there is a significant realignment of actin-binding sites in response to cross-bridge forces, 2) this realignment recruits additional cross-bridge binding, and 3) we predict mechanical behaviors that are consistent with experimental results for velocity and length transients. Binding site realignment depends on the relative compliance of the filament lattice and cross-bridges, and within the measured range of these parameters, gives rise to a sharply tuned peak for force generation. Such mechanical tuning at the molecular level is the result of mechanical coupling between individual cross-bridges, mediated by thick filament deformations, and the resultant realignment of binding sites on the thin filament.     

Hill muscle model errors during movement are greatest within the physiologically relevant range of motor unit firing rates

This study evaluated the accuracy of Hill-type muscle models during movement. Hill-type models are ubiquitous in biomechanical simulations. They are attractive because of their computational simplicity and close relation to commonly measured experimental variables, but there have been surprisingly few experimental validations of these models during functionally relevant conditions. Our hypothesis was that model errors during movement are largest at the low motor unit firing rates most relevant to normal movement conditions. This hypothesis was evaluated in the cat soleus muscle activated either by electrical stimulation at physiological rates or via the crossed-extension reflex (CXR) thereby obtaining normal patterns of motor unit recruitment and rate modulation. These activation paradigms were applied during continuous movements approximately matched to locomotor length changes. The resulting muscle force was modeled using a common Hill model incorporating independent activation, tetanic length-tension and tetanic force-velocity properties. Errors for this model were greatest for stimulation rates between approximately 10-20Hz. Errors were especially large for muscles activated via the CXR, where most motor units appear to fire within this range. For large muscle excursions, such as those seen during normal locomotion, the errors for naturally activated muscle typically exceeded 50%, supporting our hypothesis and indicating that the Hill model is not appropriate for these conditions. Subsequent analysis suggested that model errors were due to the common Hill model's inability to account for the coupling between muscle activation and force-velocity properties that is most prevalent at the low motor unit firing rates relevant to normal activation.     

Force, work and power output of lower limb muscles during human maximal-effort countermovement jumping.

The purpose of this study was to simulate human maximal-effort countermovement jumping with a three-dimensional neuromusculoskeletal model. The specific aim was to investigate muscle force, work and power output of major lower limb muscles during the motion. A neuromusculoskeletal model that has nine rigid body segments, 20 degrees of freedom, 32 Hill-type lower limb muscles was developed. The neural activation input signal was represented by a series of step functions with step duration of 0.05 s. The excitation-contraction dynamics of the contractile element, the tissues around the joints to limit the joint range of motion, as well as the foot-ground interaction were implemented. A simulation was started from a standing posture. Optimal pattern of the activation input signal was searched through numerical optimization with a goal of maximizing the height reached by the mass center of body after jumping up. As a result, feasible kinematics, ground reaction force profile and muscle excitation profile were generated. It was found that monoarticular muscles had major contributions of mechanical work and power output, whereas biarticular muscles had minor contributions. Hip adductors, abductors and external rotator muscles were vigorously activated, although their mechanical work and power output was minor because of their limited length change during the motion. Joint flexor muscles such as m. iliopsoas, m. biceps femoris short head and m. tibialis anterior were activated in the beginning of the motion with an effect of facilitating the generation of a countermovement.     

Some self-consistent two-state sliding filament models of muscle contraction.

The general formalism required to treat two-state sliding filament models of muscle contraction, including free energy considerations, is first reviewed and amplified. This formalism is then used to examine, and modify as needed, three models studied previously by Podolsky and Nolan, in which cross-bridge attachment-detachment and ATP turnover are not tightly coupled. No attempt is made here to establish an optimal, self-consistent model of this type because our interest is primarily in methadology rather than in fitting experimental results. But it appears from this preliminary study that such a model, with satisfactory mechanical and thermodynamic properties, could be found. An extremely simple but unrealistic two-state model is also studied which is of interest because it demonstrates the fact that it is possible, in principle at least, for sliding filament models to work with very high thermodynamic efficiencies (50-100 percent). An appendix is included that is concerned with the form of the dependence of certain first-order rate constants on the concentrations of ATP, ADP, and P.     

Sarcomere lattice geometry influences cooperative myosin binding in muscle

In muscle, force emerges from myosin binding with actin (forming a cross-bridge). This actomyosin binding depends upon myofilament geometry, kinetics of thin-filament Ca²⁺ activation, and kinetics of cross-bridge cycling. Binding occurs within a compliant network of protein filaments where there is mechanical coupling between myosins along the thick-filament backbone and between actin monomers along the thin filament. Such mechanical coupling precludes using ordinary differential equation models when examining the effects of lattice geometry, kinetics, or compliance on force production. This study uses two stochastically driven, spatially explicit models to predict levels of cross-bridge binding, force, thin-filament Ca²⁺ activation, and ATP utilization. One model incorporates the 2-to-1 ratio of thin to thick filaments of vertebrate striated muscle (multi-filament model), while the other comprises only one thick and one thin filament (two-filament model). Simulations comparing these models show that the multi-filament predictions of force, fractional cross-bridge binding, and cross-bridge turnover are more consistent with published experimental values. Furthermore, the values predicted by the multi-filament model are greater than those values predicted by the two-filament model. These increases are larger than the relative increase of potential inter-filament interactions in the multi-filament model versus the two-filament model. This amplification of coordinated cross-bridge binding and cycling indicates a mechanism of cooperativity that depends on sarcomere lattice geometry, specifically the ratio and arrangement of myofilaments.     

Cross-bridge duty cycle in isometric contraction of skeletal myofibrils

During interaction of actin with myosin, cross-bridges impart mechanical impulses to thin filaments resulting in rotations of actin monomers. Impulses are delivered on the average every tc seconds. A cross-bridge spends a fraction of this time (ts) strongly attached to actin, during which it generates force. The "duty cycle" (DC), defined as the fraction of the total cross-bridge cycle that myosin spends attached to actin in a force generating state (ts/ tc), is small for cross-bridges acting against zero load, like freely shortening muscle, and increases as the load rises. Here we report, for the first time, an attempt to measure DC of a single cross-bridge in muscle. A single actin molecule in a half-sarcomere was labeled with fluorescent phalloidin. Its orientation was measured by monitoring intensity of the polarized TIRF images. Actin changed orientation when a cross-bridge bound to it. During isometric contraction, but not during rigor, actin orientation oscillated between two values, corresponding to the actin-bound and actin-free state of the cross-bridge. The average ts and tc were 3.4 and 6 s, respectively. These results suggest that, in isometrically working muscle, cross-bridges spend about half of the cycle time attached to actin. The fact that 1/ tc was much smaller than the ATPase rate suggests that the bulk of the energy of ATP hydrolysis is used for purposes other than performance of mechanical work.     

Characterization of cross-bridge elasticity and kinetics of cross-bridge cycling during force development in single smooth muscle cells

Force development in smooth muscle, as in skeletal muscle, is believed to reflect recruitment of force-generating myosin cross-bridges. However, little is known about the events underlying cross-bridge recruitment as the muscle cell approaches peak isometric force and then enters a period of tension maintenance. In the present studies on single smooth muscle cells isolated from the toad (Bufo marinus) stomach muscularis, active muscle stiffness, calculated from the force response to small sinusoidal length changes (0.5% cell length, 250 Hz), was utilized to estimate the relative number of attached cross-bridges. By comparing stiffness during initial force development to stiffness during force redevelopment immediately after a quick release imposed at peak force, we propose that the instantaneous active stiffness of the cell reflects both a linearly elastic cross-bridge element having 1.5 times the compliance of the cross-bridge in frog skeletal muscle and a series elastic component having an exponential length-force relationship. At the onset of force development, the ratio of stiffness to force was 2.5 times greater than at peak isometric force. These data suggest that, upon activation, cross-bridges attach in at least two states (i.e., low-force-producing and high-force-producing) and redistribute to a steady state distribution at peak isometric force. The possibility that the cross-bridge cycling rate was modulated with time was also investigated by analyzing the time course of tension recovery to small, rapid step length changes (0.5% cell length in 2.5 ms) imposed during initial force development, at peak force, and after 15 s of tension maintenance. The rate of tension recovery slowed continuously throughout force development following activation and slowed further as force was maintained. Our results suggest that the kinetics of force production in smooth muscle may involve a redistribution of cross-bridge populations between two attached states and that the average cycling rate of these cross-bridges becomes slower with time during contraction.     

Are the maximum shortening velocity and the shape parameter in a Hill-type model of whole muscle related to activation?

Mathematical models of the inter-relationship of muscle force, velocity, and activation are useful in forward dynamic simulations of human movement tasks. The objective of this work was to determine whether the parameters (maximum shortening velocity V(max) and shape parameter k) of a Hill-type muscle model, interrelating muscle force, velocity, and activation, are themselves dependent on the activation. To fulfill this objective, surface EMG signals from four muscles, as well as the kinematics and kinetics of the arm, were recorded from 14 subjects who performed rapid-release elbow extension tasks at 25%, 50%, 75%, and 100% activation (MVC). The experimental elbow flexion angle was tracked by a forward dynamic simulation of the task in which V(max) and k of the triceps brachii were varied at each activation level to minimize the difference between the simulated and experimental elbow flexion angle. Because a preliminary analysis demonstrated no dependency of k on activation, additional simulations were performed with constant k values of 0.15, 0.20, and 0.25. The optimized values of V(max) normalized to the average value within a subject were then regressed onto the activation. Normalized V(max) depended significantly on the activation (p<0.001) for all values of k. Furthermore, the estimated V(max) values were not sensitive to the selected k value. The results support the use of Hill-type models in which V(max) depends on activation in forward dynamic simulations modeling muscles with mixed fiber-type composition recruited in the range of 25-100% activation. The use of more accurate models will lend greater confidence to the results of forward dynamic simulations.     

Kinetic mechano-chemical model of a muscle with a six-stage cross-bridge cycle]

A kinetic scheme of the mechano-chemical cycle of the cross-bridges and a mathematical model based on this scheme are proposed. The main assumptions accepted in the scheme are: the step of the inorganic phosphate release precedes the force-generating step of a cross-bridge; the rate-limiting step of the ATP hydrolysis is isomerization of the actomyosin-ADP complex. It is shown that the model well describes the mechanical and biochemical transients initiated by the temperature jump and flash photolysis of the caged compounds in skinned muscle fibres.     

Muscle Compliance and the Longitudinal Transmission of Mechanical Impulses

The time required for a mechanical impulse to propagate from one end to the other was measured directly in frog sartorius muscles and in fiber bundles from the semitendinosus muscle. When the fibers were fully activated, the transmission velocity was 170 mm/ms. In resting fibers the transmission time was three to four times greater than in activated fibers. Control experiments indicated that the transmission time across the tendons was negligible. A muscle compliance of 55–80 Å per half sarcomere was estimated from these data. The "measurement time" of the method was calculated to be about 15 µs. This relatively short measurement time makes the method potentially useful for detecting changes in cross-bridge compliance.     

The smooth muscle cross-bridge cycle studied using sinusoidal length perturbations

The mechanical characteristics of smooth muscle can be broadly defined as either phasic, or fast contracting, and tonic, or slow contracting (, Pharmacol. Rev. 20:197-272). To determine if differences in the cross-bridge cycle and/or distribution of the cross-bridge states could contribute to differences in the mechanical properties of smooth muscle, we determined force and stiffness as a function of frequency in Triton-permeabilized strips of rabbit portal vein (phasic) and aorta (tonic). Permeabilized muscle strips were mounted between a piezoelectric length driver and a piezoresistive force transducer. Muscle length was oscillated from 1 to 100 Hz, and the stiffness was determined as a function of frequency from the resulting force response. During calcium activation (pCa 4, 5 mM MgATP), force and stiffness increased to steady-state levels consistent with the attachment of actively cycling cross-bridges. In smooth muscle, because the cross-bridge states involved in force production have yet to be elucidated, the effects of elevation of inorganic phosphate (P(i)) and MgADP on steady-state force and stiffness were examined. When portal vein strips were transferred from activating solution (pCa 4, 5 mM MgATP) to activating solution with 12 mM P(i), force and stiffness decreased proportionally, suggesting that cross-bridge attachment is associated with P(i) release. For the aorta, elevating P(i) decreased force more than stiffness, suggesting the existence of an attached, low-force actin-myosin-ADP- P(i) state. When portal vein strips were transferred from activating solution (pCa 4, 5 mM MgATP) to activating solution with 5 mM MgADP, force remained relatively constant, while stiffness decreased approximately 50%. For the aorta, elevating MgADP decreased force and stiffness proportionally, suggesting for tonic smooth muscle that a significant portion of force production is associated with ADP release. These data suggest that in the portal vein, force is produced either concurrently with or after P(i) release but before MgADP release, whereas in aorta, MgADP release is associated with a portion of the cross-bridge powerstroke. These differences in cross-bridge properties could contribute to the mechanical differences in properties of phasic and tonic smooth muscle.     

Cardiac Myosin Binding Protein C Phosphorylation Affects Cross-Bridge Cycle's Elementary Steps in a Site-Specific Manner

Based on our recent finding that cardiac myosin binding protein C (cMyBP-C) phosphorylation affects muscle contractility in a site-specific manner, we further studied the force per cross-bridge and the kinetic constants of the elementary steps in the six-state cross-bridge model in cMyBP-C mutated transgenic mice for better understanding of the influence of cMyBP-C phosphorylation on contractile functions. Papillary muscle fibres were dissected from cMyBP-C mutated mice of ADA (Ala273-Asp282-Ala302), DAD (Asp273-Ala282-Asp302), SAS (Ser273-Ala282-Ser302), and t/t (cMyBP-C null) genotypes, and the results were compared to transgenic mice expressing wide-type (WT) cMyBP-C. Sinusoidal analyses were performed with serial concentrations of ATP, phosphate (Pi), and ADP. Both t/t and DAD mutants significantly reduced active tension, force per cross-bridge, apparent rate constant (2πc), and the rate constant of cross-bridge detachment. In contrast to the weakened ATP binding and enhanced Pi and ADP release steps in t/t mice, DAD mice showed a decreased ADP release without affecting the ATP binding and the Pi release. ADA showed decreased ADP release, and slightly increased ATP binding and cross-bridge detachment steps, whereas SAS diminished the ATP binding step and accelerated the ADP release step. t/t has the broadest effects with changes in most elementary steps of the cross-bridge cycle, DAD mimics t/t to a large extent, and ADA and SAS predominantly affect the nucleotide binding steps. We conclude that the reduced tension production in DAD and t/t is the result of reduced force per cross-bridge, instead of the less number of strongly attached cross-bridges. We further conclude that cMyBP-C is an allosteric activator of myosin to increase cross-bridge force, and its phosphorylation status modulates the force, which is regulated by variety of protein kinases.