User:Asantorelli/Neuromuscular control of the ankle

Neuromuscular control of the ankle refers to both the body’s motor response to sensory information and the conscious decisions made to alter ankle joint mechanics. The ability to control the muscles surrounding the ankle is pivotal in movement and posture control. Furthermore, this control is necessary to prevent injury due to abnormal joint motion.

Conscious decisions about movement and muscle contraction are carried to muscles via the peripheral nervous system. This system allows for sensory feedback to exist such that the neurons controlling muscle contraction can respond to external stimuli. The correct response to these stimuli prevents the body from incurring injuries. This sensory information is relayed to the muscle through receptors located within the muscle. These receptors respond to deformation to the muscle and provide feedback about the type of deformation.

The mechanical properties of the ankle joint can be defined by the dynamic joint stiffness. The dynamic joint stiffness relates the changes in the joint position to the changes of the torque acting about it. In order to prevent external forces from causing abnormal joint movement it is necessary to increase the joint stiffness. The joint stiffness can be separated into two components: the intrinsic stiffness and the reflex stiffness.

Ankle injuries are the most common orthopedic injury. Changes within the neuromuscular system can explain the nature of some of these injuries, whereas, in some cases, injuries may cause the altered behaviour of this system. Studying changes to ankle joint stiffness resulting from injuries, or even muscle fatigue, will possibly help improve the treatment and prevention of these injuries.

Physiology
In order to understand the mechanisms through which the ankle joint can be controlled, it is first necessary to comprehend the complex interactions governing this behaviour. This section outlines the process in which signals are generated within the brain, how they propagate to the muscles of the ankle, and how this signal is converted to the generation of a force.

Central nervous system
The central nervous system (CNS) consists of the brain and the spinal cord. This region gathers information from all regions of the body, and, with this information, is responsible for coordinating all physical activities. It is the CNS which is concerned with initializing movements, stabilizing the position, and selecting the muscles to contract in order to control the ankle joint.

Peripheral nervous system
The peripheral nervous system (PNS) is responsible for transferring information from the CNS to the other structures of the body. It is necessary so that information may both leave and enter the CNS. Efferent or motor neurons are responsible for carrying information away from the CNS,  while afferent neurons are responsible for returning sensory information to the CNS. Unlike the central nervous system, the peripheral nervous system is not protected by bone, such as the skull and vertebrae, making it susceptible to external damages.

Made up of both the somatic nervous system and the autonomic nervous system, the peripheral nervous system is vital to ensure the proper function of both consciously controlled actions (somatic nervous system) and the unconsciously controlled visceral functions (autonomic nervous system). The somatic nervous system consists of all the neurons which are directly connected to skeletal muscle. It is this component of the nervous system which communicates with the muscles of the ankle in order to establish movement.

Motor neuron
Motor neurons connect the CNS to skeletal muscle. The body of these neurons reside within the CNS and it is the axons which are connected to the skeletal muscle which they innervate. There are two primary types of motor neurons. Large motor neurons which innervate muscle fibre are referred to as α-motor neurons. γ-motor neurons are smaller neurons which excite the sensory muscle fibres.

Motor unit
A motor unit is defined by a motor neuron and the muscle fibres it innervates. Each motor neuron can innervate numerous muscle fibres, however, each fibre can only be innervated by one neuron. The number of muscle fibres in a motor unit is dependent on the function of the muscle. Muscles controlling small and precise movement require few muscle fibres per neuron, whereas large muscles, required to generate large forces, require many muscle fibres.

Force generation
There are two mechanisms that have been identified for the control of motion through muscle force generation: rate coding and muscle recruitment.

The mean tension in the muscle increases in a non-linear fashion as the stimulus rate increases. A higher input rate leads to a more powerful force being generated at the muscle. Furthermore, it has been shown that muscle exhibits low-pass filter properties in response to variations in the stimulus rate.

Muscle recruitment has been shown to follow a specific pattern: increasingly larger motor units are recruited as the required force is increased. Hence, small motor units are recruited first while the largest are recruited last. Larger motor units have more muscle fibres connected to each neuron and are thus more difficult to stimulate, hence they will only be recruited when there is a requirement for a large force. This phenomenon is referred to as the Henneman size principle.

The importance of each mechanism in locomotion control remains debatable; certain movements favour one mechanism over the other. When attempting to maintain a steady force within a muscle, muscle recruitment offers a favourable mechanism to balance this force. Attempting to fire a small number of motor units at a high rate over a prolonged period of time would lead to considerable fatigue, especially when compared to firing many motor neurons at a low rate. However, when attempting to generate powerful forces over short periods of time, rate coding offers a beneficial means of increase the amount of force to be generated. If all the muscle fibres have been recruited the only means to generate a stronger contraction would be to increase the stimulus rate of each motor unit.

Neuromuscular junction
When a motor unit fires it causes all the muscles fibres it innervates to contract. The electrical signal generated at the body of the neuron travels to the terminal of the axons. At each axon terminal of a motor neuron there is a junction with the muscle fibres. These junctions are referred to as neuromuscular junctions, a type of synapse. At this synapse, neurotransmitters allow the electrical signal to reach the muscle cells. Action potentials traveling along the axons of a motor neuron cause the neurotransmitters to be released.

Control of the ankle
The ability to control the muscles of the ankle, whether it be for locomotion or posture control, is dependent on not only on the consciously decided signals generated within the CNS, but relies heavily on sensory information provided by the peripheral system. This feedback information is necessary to allow muscles to react to external stimuli, preventing both the destabilization of the body and the potential injury to joints and muscles due to excessive strain.

Muscle receptors
The two primary muscle receptors for providing sensory information are: muscle spindles and golgi tendon organs. These receptors can be classified as mechanoreceptors. They are able to transduce information about mechanical deformation, change in muscle length, to an electrical signal.

Muscle spindle
Muscle spindles are located within the muscle fibre and, as such, will undergo similar size deformations. They are innervated by both afferent and efferent nerves and therefore are able to relay information to the CNS and also respond to information from the CNS. Unlike the major muscle in which muscle spindles are embedded, muscle spindles are innervated by γ-neurons as opposed to α-neurons. The stimulation of muscle spindles by the γ-motor neurons will cause the muscle to contract.

When the muscle is stretched action potentials are created, sending a signal along the afferent nerves. The muscle spindles are able to detect both changes in muscle length as well as the velocity of this change. The afferent nerves of the muscle have been shown to play a prominent role in the stretch reflex. A stretch in the muscle will excite afferent nerves which will, in turn, excite the efferent nerves of that very same muscle. More precisely the Type Ia fibres will directly excite the α-motor neurons, causing the muscle to contract. This closed loop response to an external stimulus by muscle spindles is a reflex arc known as the stretch reflex.

The indirect excitation of efferent neurons by afferent neurons is possible due to interneurons connecting afferent sensory nerves and efferent motor nerves.

Golgi tendon organ
Golgi tendon organs (GTOs) are located within tendons, close to the musculotendon joint. They are able to detect changes in the tension within a muscle. Golgi tendon organs are stimulated by both the contraction and stretching of a muscle. This stimulus has the opposite effect of the muscle spindle; it produces a relaxation in the muscle by inhibiting motor neurons. This inhibition allows for a means of tension control within the muscle by the negative feedback provided by the golgi tendon organ.

Interneurons
The ability to excite, or inhibit, α-motor neurons is, primarily, dependent on the ability of afferent nerves from sensory receptors to communicate with the efferent nerves which innervate the motor neurons. Interneurons connect the afferent pathways, in this case the afferent neurons from GTOs and muscle spindles, with motor neurons. These connections allow feedback to take a prominent role in determining the function of α-motor neurons.

System Models
There has been several different control system models presented in literature describing muscle control. This section highlights three specific models, states the disadvantages of each, and how each subsequent model is an improvement to the previous model.

Direct control by α-motor neurons
In this system the muscle fibres are under direct control from the α-motor neurons. The α-motor neurons receive input signals from both the CNS and the sensory feedback provided by muscles spindles. This feedback from the muscle spindles is necessary to allow the system to resist external perturbations. Unfortunately, in this system model, the feedback system will also attempt to resist any changes that are brought about by the commands from the CNS. The central commands will appear to be external perturbations causing muscle length deformation, thus causing the muscle spindles to oppose these changes. Furthermore, the introduction of positive feedback into a system allows for the possibility of instability within the system.

Indirect control by γ-motor neurons
In this model the muscle fibres are now indirectly controlled by the firing of γ-motor neurons. The CNS communicates with the efferent nerve endings in the muscle spindles and causes their excitation. The α-motor neurons are no longer controlled by the command signal from the CNS, but are solely controlled by the afferent nerves of the muscle spindles. In this new system model there will be no signal within the feedback loop in a state of equilibrium (same stretch in the muscle spindle fibres as the skeletal muscle fibres), hence the muscle fibres can match the desired length exactly. Furthermore, this system is less susceptible to external perturbation.

The difficulties with this system model are primarily due to the indirect nature of the system. Due to this indirect control there is an increase in the pathway length from the CNS to the muscle fibre. Adding to the delay is the fact that γ-fibres are relatively slow conducting, hence indirect control by γ-motor neurons make it impossible to have sudden movements. This longer pathway from the CNS to the α-motor neurons also introduces noise into the system stemming from an increased number of neuromuscular junctions.

α-γ cross-link control
The previous models had control commands directly controlling the α-neurons or communicating directly with the γ-neurons. A simultaneous activation of both the α and γ-motor neurons can help to eliminate the issues outlined above. The direct control of the α-motor neurons allows for immediate control of movement, while commands to the γ-motor neurons allows for the direct tracking of movement. By including the negative feedback that is introduced by the golgi tendon organs, it is possible to have a comprehensive system model. The α-motor neurons are now controlled by commands from the CNS, afferent nerves from muscle spindles (which are in turn controlled both by commands from the CNS and by sensory feedback), and the negative feedback from afferent nerves of the golgi tendon organ.

System Identification
System identification studies use experimental data to create a mathematical model of the system under investigation. In order to create this model it is necessary to determine the relationship between the inputs and outputs of the given system.

Dynamic Joint Stiffness
Dynamic joint stiffness is defined by the relationship between changes in the angular position of a joint and the torque acting upon the joint at that moment. Joint stiffness plays a pivotal role in motor control system; it defines the mechanical behaviour of a joint by determining the amount of force required to move a limb to a desired location and the resulting motion from external perturbations. The joint stiffness can be modeled as the transfer function of the Torque-Position system, defined in the following equation as $$H(s)$$:
 * $$ H(s) = \frac{Tq(s)}{\theta(s)} $$

where $$H(s)$$ is the transfer function describing the joint stiffness, $$Tq(s)$$ is the function describing the torque applied to the ankle joint, and $$\theta(s)$$ is the position function. The inverse of joint stiffness is joint compliance, the relationship of the torque acting on a joint and the angular changes in the joints position.

There have been two primary methods in attempting to determine and evaluate a model for total joint stiffness of the ankle. First, is to present the system with predetermined input and measure the resulting change in position. This methodology is used to determine joint compliance. The second method is to manipulate the position of the ankle and record the torque, hence determining the dynamic stiffness of the joint. In order to record experimental data with live human subjects the second method is employed. Data can be collected in this experimental setup by having subject positioned in a fixed and rigid fashion, such as lying supine with their posture maintained by straps or seated in a chair. At this point one foot should be encased in an object that will allow for its position tracking. Electromyograms (EMG) are employed to measure the signals within the skeletal muscle, the torque can be measured by the use of a torque sensor. As perturbations are presented to the ankle the corresponding signals can be measured.

There are no major differences between the results obtained by either of these methods, assuring the legitimacy of using experimental data. Ankle stiffness can be modeled by a second order transfer function defined by elastic, inertial, and viscuous components, as denoted in the following equation:
 * $$ H(s) = \frac{Tq(s)}{\theta(s)}={I{s^2} + B{s} + K}$$,

where $$I$$, $$B$$, and $$K$$ represent the inertial, viscosity, and elastic parameters respectively. The inertial component remains relatively invariant as a function of torque, where as the viscosity and elasticity parameters vary with changing torque. It is this elastic component which is of interest, as it is the major parameter in defining the stiffness of the ankle joint. As the torque applied to the ankle is increased so is the overall joint stiffness (as well as the elastic parameter, K). Furthermore, this proportional relationship has been shown to be valid at varying levels of muscle activation up to maximal contraction. This straight line relationship has enabled a measure of the resting joint stiffness of the ankle within a range of 50-100Nm/rad.

Intrinsic and Reflex Stiffness
Dynamic joint stiffness is made up of two components; intrinsic and reflex stiffness.

The intrinsic stiffness refers to the contribution from the mechanical properties of the joint, the passive tissue surrounding the joint, and the active muscles around the joint. Intrinsic stiffness can be modeled by a linear second-order system, defined by inertial, viscous, and elastic parameters, akin to the second-order model describing the total stiffness. The elastic parameter, K, is referred to as the intrinsic stiffness gain. Much like the total dynamic joint stiffness, the intrinsic stiffness increases monotonically with an increase in torque.

Reflex stiffness is the contribution to the total joint stiffness that is caused by the stretch reflex due the positive feedback loop from muscle spindles. This activation of muscle fibres will increase the total stiffness. The reflex stiffness lags the intrinsic stiffness by approximately 50ms, corresponding to the amount of time required for a signal to travel along the afferent nerves, reach the motor neurons, and instruct the muscle to contract. The stretch reflex has been found to contribute significantly to the overall stiffness, allowing the stiffness to increase past the intrinsic stiffness, thus, the reflex stiffness contributes to the mechanical behaviour of the ankle.

As the dynamic stiffness changes both the intrinsic and reflex stiffness will be altered together. The relative contribution of each form of stiffness will vary due to the ankle position, the amount of contraction of the surrounding muscles , and the nature of external perturbations. By determining the contribution in these various scenarios from both the intrinsic and reflex stiffness, it is possible to discover the role the stretch reflex plays in the control of posture and movement.

Separating Intrinsic and Reflex Contributions
There exist several methods to separate the two contributions. The primary methods are to use electric stimulation, in particular stimulation of the deep peroneal nerve, to prevent the reflex pathway from activating, and to use system identification methods to analyze and separate each contribution.

When using electrical stimulation it becomes necessary to have two sets of experiments. One set is used to determine the intrinsic stiffness. It is at this point when the peroneal nerve is stimulated. The corresponding measurements correspond to the contribution from the intrinsic stiffness only. The measurements are repeated, this time without any electrical stimulation. The difference between these two sets of data can be attributed to the addition of the reflex stiffness to the torque-position system.

By modeling the system as the sum of two separate pathways, one for each the intrinsic and reflex stiffness, it is possible to separate the signals in a post-processing fashion. In this case only one set of experiments are required, increasing the time efficiency of the experiments. Knowing that the intrinsic stiffness is modeled by a linear second-order system, it is possible to convolve an estimate of this model with the input position. The result is the torque contribution from the intrinsic stiffness. This component is made to be no longer than 40ms in duration to avoid correlation with the reflex stiffness. By subtracting the intrinsic component from the total torque it is possible to obtain an estimate on the contribution from reflex stiffness. The reflex stiffness can be modeled as a 2nd or 3rd order system consisting of a delay, differentiator, and a linear and non-linear subsystem. Subsequent analysis using Hammerstein identification procedures are used to finalize the modelling of the intrinsic and reflex stiffness. In particular, it is necessary to define the non-linear elements of the reflex pathway in order to obtain the reflex stiffness. This is an iterative procedure, which continues until the estimate of the total torque, the sum sum of the intrinsic and reflex components, is sufficiently close to the total measured torque.

Reflex Stiffness Contributions
The reflex stiffness increases the total joint stiffness, altering the mechanical properties of the joint. The contribution from the reflex stiffness, defined as the ratio of reflex to intrinsic stiffness, is found to peak at about 100-300ms after the initial stretch of the muscle. This contribution is dependent on muscle activation, joint position, and the nature of external stimuli, and, in some cases, makes up half of the total joint stiffness.

The role of muscle contraction on the contribution of the reflex stiffness is controversial. While it has been reported that there is no contribution during both a relaxed state and at maximum contraction and a peak contribution in 30-50% maximum voluntary contraction (MVC) range, it has also been shown that the maximum contribution occurs at very low muscle activation. Furthermore, the maximum contribution occurs close to the neutral position. Hence, it is believed, that the maximum contribution from the stretch reflex occurs at a low activation level with the ankle in a neutral position.

The reflex stiffness is also affected by the velocity and the amount of stretch caused by an external perturbation. There is an inverse relationship between the velocity of perturbation and the contribution to the total stiffness from the reflex stiffness. At very high velocities, above 0.5rad/s, the contribution from the stretch reflex is minimal, often lower than 5%. However, when a contracted muscle is stretched, by rotating the ankle by a certain amount of degrees, the contribution is found to be relatively stable. For a varying stretch from 2-7° the reflex stiffness contribution is approximately 50%.

Stretch Reflex and Walking
The stretch reflex plays a prominent role in muscle activation during walking. In order to establish the exact contribution from the stretch reflex, subjects are made to walk on a treadmill whilst known perturbations are presented to the ankle throughout the entire walking cycle. The stretch reflex is most prominent during the stance phase (from heel contact to toe off the ground). During the swing cycle the stretch reflex is far less pronounced. As the foot leaves the ground, the phase between stance and swing, there is a complete suppression of the stretch reflex.

There is a proportional correlation between the stretch reflex amplitude and walking speed, hence, at high speeds, the afferent input to the ankle becomes more important. Furthermore, the high stiffness observed during the early stance phase is believed to be a means of protecting the ankle joint from irregularities on the ground by increasing total joint stiffness.

Voluntary Control of Reflex Stiffness
In order to identify whether the CNS can control the stretch reflex independently from muscle contraction it is necessary to know whether the individual is modifying the intrinsic or reflex stiffness. A recent experimental setup allows for the real-time measurement of both the intrinsic and reflex stiffness of a subject. This information is given as feedback to the subject. With this biofeedback it has been shown that the CNS can control the reflex stiffness independently from the intrinsic stiffness, thus proving that the CNS can adjust the mechanical properties (dynamic stiffness) of the ankle joint in order to meet the varying functional demands placed upon the joint.

Intrinsic stiffness while Standing
While standing upright the body experiences many small and short perturbations due to the minor sway of the body. Due to the short duration and magnitude of these forces the reflex stiffness provides no contribution to the joint stiffness. It has been shown that only 64-91% of the total ankle joint stiffness required to stay upright is due to the intrinsic stiffness of the ankle. The lower end of this scale corresponds to more pronounced sway, whereas smaller sway leads to a higher contribution from the intrinsic stiffness. This evidence suggests that there is a neural anticipation which increases the ankle stiffness to maintain stability. It is believed that as age increases the effectiveness of this anticipation is diminished leaving the elderly with difficulties to stand upright. Developing strategies to increase the joint stiffness, by voluntary muscle contraction for instance, become of interest to address the problem.

Injury
Ankle injuries cause damage to the surrounding afferent nerves and leads to joint inflammation. This nerve damage alters the sensory feedback system, which in turn modifies joint stiffness. Furthermore, joint inflammation leads to a painful sensation which overloads the sensory pathways and causes a diminished detection of joint motion and muscle stretch. These changes in the peripheral system leads to altered neuromuscular control of the ankle.

In subjects with chronic ankle instability(CAI) there is evidence of a deficit in both static and dynamic posture control. This deficit is classified as an alteration in neuromuscular control of the ankle. Furthermore, there is evidence of altered joint stiffness in subjects with CAI due to a delay in stretch reflex.

Rehabilitation techniques aim to re-establish the afferent-efferent nerve connections in order to help the patient relearn the muscular control pathways. Certain training techniques can be incorporated to increase joint stiffness in order to protect the joint from dangerous strain. Increasing muscle activation will increase the joint stiffness and decrease the latency time of the stretch reflex, helping to better protect the ankle joint from many common injuries.

Spastic Patients
The ability of determine the total joint stiffness, as well as the intrinsic and reflex contributions, at various ankle positions and in respect to various perturbations allows the mechanics governing the ankle joint to be well characterized. This characterization can be extended to determine mechanics governing ankle joint for spastic spinal cord injured (SCI) patients, shedding light on how these injuries impact the control of the ankle.

In SCI patients the total joint stiffness is found to be increased in comparison to a healthy control group. While both the intrinsic and reflex stiffness are increased, the increase in the reflex stiffness is far more substantial in both magnitude and in its relative contribution. This increase in reflex stiffness is found to be highly dependent on joint position. At full plantarflexion there was minimal difference between the control group and the spastic patients. At mid-range of dorsiflexion the reflex stiffness was found to be much higher in SCI patients, and at full dorsiflexion both the intrinsic and reflex stiffness in SCI patients was found to be significantly higher.

Fatigue
Most athletes report that ankle sprains or ankle inversions, the most common ankle injury, occur late in competitions. While there is no scientific proof for this anecdotal evidence, the role of muscle fatigue on ankle joint mechanics is of interest. The changes of dynamic ankle joint stiffness due to muscle fatigue seem to be inconclusive. Early investigation demonstrated that there is no change in the joint stiffness due to fatigue, citing no significant variation in the K, I, and B parameters which govern the joint stiffness model. However, later research concluded that muscle fatigue will cause a decrease in the force production due to a decreased firing rate. This decreased force is observed in the muscles which provide dynamic stabilization to the ankle joint; the joint stiffness has been decreased due to muscle fatigue thus leading to a predisposition to ankle inversions. In subjects with CAI, fatigue is found to play a more prominent role in decreasing the overall joint stiffness. It has been hypothesized that fatigue increases the muscle spindle firing threshold, disrupting the afferent feedback and, ultimately, decreasing overall joint stiffness.

Injury Prevention and Rehabilitation
Specific training programs can help to prevent ankle injuries and help speed up the recovery. Specifically joint stabilizing techniques, balance, and co-ordination exercises have been shown to reduce the risk to injury and reduce the rate of recurrence. Theses dynamic stabilization techniques encourage sensory muscle activity. Joint compression, via neoprene sleeves or ace bandages, stimulates cutaneous sensory receptors thus increasing overall joint stiffness.

Future
In order to develop an ankle prosthesis it is necessary to understand the mechanics of the joint. By understanding the roles played by the intrinsic and reflex stiffness with varying joint position, perturbation velocity, muscle activation, and the effects of injury, it becomes possible to have a very well characterized joint. This characterization may allow for a highly evolved prosthesis to be developed, one capable of appropriately reacting to external and internal stimuli.