A Linked Muscular Activation Model for Movement Generation and Control

A new model for movement control is presented which incorporates characteristics of impulse-variability and mass-spring models. Movements in the model were controlled with phasic torque impulses in agonist and antagonist muscles and a tonic agonist torque.

Characteristics of the phasic agonist and antagonist torque profiles were based on observed properties of movement-related EMGs and muscle isometric torques. Variability of the phasic impulses depended on impulse magnitude as in impulse-variability models. The model therefore predicted a speed-accuracy tradeoff for limb movement. The time of onset and magnitude of the antagonist torque depended on the magnitude of the preceding agonist torque as indicated in studies of movement-related EMGs. This led to the new concept of linkage between the agonist and antagonist muscle forces which was shown to be important for reducing variability of fast movements. Progressive development of linkage during practice could explain the previous findings of decreased movement variability with practice coupled with increased variability of movement-related EMGs.

It was concluded that an inherently variable motor system deals with the variability associated with generation of large muscle forces by linking the forces produced by opposing muscles. In this way, variability in net joint torques and in movements can be decreased without the need for the nervous system to closely regulate the individual torques.     

A dynamical systems approach to manual tracking performance

Manual tracking performance was investigated from the perspective of dynamical systems theory. The authors manipulated the type of visual display, the control system dynamics, and the frequency of the sinusoidal input signal to examine couplings with various phases between the visual signal and control movements. Analyses of the system output amplitude ratio and relative phase showed that participants (N = 24) performed poorly with 90 degrees relative phase coupling. All the couplings became less stable as the movement frequency increased. The authors developed an adaptive oscillator model with linear damping to describe the coupled system consisting of the human performer, the visual display, and the control system dynamics. A geometric account of the stability of performance at different relative phases is also presented.     

A linked muscular activation model for movement generation and control

A new model for movement control is presented which incorporates characteristics of impulse-variability and mass-spring models. Movements in the model were controlled with phasic torque impulses in agonist and antagonist muscles and a tonic agonist torque. Characteristics of the phasic agonist and antagonist torque profiles were based on observed properties of movement-related EMGs and muscle isometric torques. Variability of the phasic impulses depended on impulse magnitude as in impulse-variability models. The model therefore predicted a speed-accuracy tradeoff for limb movement. The time of onset and magnitude of the antagonist torque depended on the magnitude of the preceding agonist torque as indicated in studies of movement-related EMGs. This led to the new concept of linkage between the agonist and antagonist muscle forces which was shown to be important for reducing variability of fast movements. Progressive development of linkage during practice could explain the previous findings of decreased movement variability with practice coupled with increased variability of movement-related EMGs. It was concluded that an inherently variable motor system deals with the variability associated with generation of large muscle forces by linking the forces produced by opposing muscles. In this way, variability in net joint torques and in movements can be decreased without the need for the nervous system to closely regulate the individual torques.     

Simulation Studies of Descending and Reflex Control of Fast Movements

Several neurological control strategies for fast head movements are considered using computer simulations of a stretch reflex model. Each control strategy incorporates a different amount of proprioceptive feedback contributing to braking and/or clamping the movement. The model behavior for each control strategy is qualitatively compared to experimental data that includes the agonist and antagonist EMGs, and the head position, velocity, and acceleration. Significance of the study is discussed with respect to the characteristic tri-phasic EMG pattern for fast voluntary movements and the possible roles that the stretch reflex may have in contributing to this pattern of activation.     

A Dynamical Systems Approach to Manual Tracking Performance

Manual tracking performance was investigated from the perspective of dynamical systems theory. The authors manipulated the type of visual display, the control system dynamics, and the frequency of the sinusoidal input signal to examine couplings with various phases between the visual signal and control movements. Analyses of the system output amplitude ratio and relative phase showed that participants (N = 24) performed poorly with 90° relative phase coupling. All the couplings became less stable as the movement frequency increased. The authors developed an adaptive oscillator model with linear damping to describe the coupled system consisting of the human performer, the visual display, and the control system dynamics. A geometric account of the stability of performance at different relative phases is also presented.     

Joint control strategies and hand trajectories in multijoint pointing movements

The role of timing in the control of multijoint pointing movements was evaluated. Eight subjects performed rapid pointing movements to a variety of target locations. The subject's right arm was strapped to a 2 degrees of freedom manupilandum that permitted shoulder and elbow motion in the horizontal plane. Initial and final position of the hand and magnitude of displacement was varied to determine effects on timing characteristics. Kinematics and kinetics of the shoulder, elbow, and hand were analyzed. The hand paths and velocity profiles observed were consistent with prior reports. Multiple regression analysis of kinematic variables disclosed that timing of joint movement onset was independent of initial and final positions of the hand, but was linearly related to joint displacement: the joint that moved farther started moving first. Using computer simulations to create joint movement onset, times that were different from the observed ones always resulted in hand paths with increased curvatures and loss of the smooth velocity profiles. Secondly, a very stable, linear relationship was observed between peak velocity and displacement at both the elbow and shoulder joints. This relationship was not affected by variations in movement space. We suggest that space-time transformation based on difference in joint displacement is used to regulated timing of joint movement onset. The simulations indicate that this transformation is set to produce smooth velocity profiles. The relationships between timing of movement onset and displacement and between peak velocity and displacement complement each other: by maintaining a linear relationship between velocity and displacement, a linear space time transformation can be used to control timing. Furthermore, these relationships are probably used to simplify coordination between the moving joints.     

Movement time and velocity as determinants of movement timing accuracy

Three experiments investigated the effect of movement time (MT) and movement velocity on the accuracy and initiation of linear timing movements. MTs of 100, 200, 500, 600, and 1000 msec were examined over various distances; timing accuracy decreased with longer MTs and slower average velocities. The velocity effect was independent of MT and occurred when the velocities were above and below about 15 cm/sec. Self-paced initiation times to movement increased directly with MT and inversely as a function of movement velocity. The latency data complement the MT findings in suggesting that average velocity is a key parameter in the initiation and control of discrete timing movements and, that there is some lower velocity below which movement control breaks down.     

Modelling the parallel bars in men's artistic gymnastics

The modelling of the parallel bars-gymnast system is considered. A 2D frontal plane model for the parallel bars apparatus is developed, enabling technique and injury analysis to be undertaken when combined with an interacting gymnast body model. We also demonstrate how such a gymnast body model may be combined with the parallel bars model by use of a simplifying symmetry consideration about the gymnast's sagittal plane. This symmetry consideration implies that just half the gymnast body and one of the two bars, are needed in the total model. We found that midpoint vertical parallel bars dynamics may be modelled by three parameters, using a single damped spring-mass model with linear force-displacement characteristics. Horizontally, as opposed to the vertical direction, bar endpoints accounted for a substantial part (35%) of the midpoint movement, demanding two serially connected springs for this direction. One spring represented the absolute horizontal movement of the bar endpoints, while the other spring represented the superimposed horizontal movement of bar midpoint relative to the endpoints. Both horizontal springs had the same characteristics as the vertical spring, giving a total of nine parameters for the three-spring bar model. Bar parameters were estimated by fitting the modelled bar movements to corresponding measured movements caused by a 140 kg lateral pendulum below the bar midpoint. Validation was then undertaken by comparing model-predicted bar movements to corresponding measurements using lateral pendulums of 100 kg and 60 kg, respectively. Finally, a gymnast handstand position was modelled and used to compare model-predicted and measured bar oscillations following a somersault backwards to a handstand position. The model gave convincing predictions of bar movements both for the 100 kg (1 period, RMS error of 7.0 mm) and 60 kg (1 period, RMS error of 3.7 mm) pendulums, as well as for the somersault landing (2 periods, RMS error of 8.1 mm).     

Controlling velocity in rapid movements

It has often been reported that subjects prefer to use a strategy in which they vary movement velocity and peak amplitude in a linear fashion. In this study, control of velocity and amplitude in rapid reciprocating movements of the interphalangeal joint of the thumb was investigated by examining movement trajectories and patterns of activity in the extensor pollicis longus (EPL) and flexor pollicis longus (FPL) muscles. In controlling either amplitude or peak flexion velocity without constraint, subjects always used a strategy in which peak extension velocity and peak flexion velocity had strong linear correlations with movement amplitude. When they were required to keep either amplitude or peak flexion velocity fixed their movements were still biased toward a strategy in which peak velocity and movement amplitude covaried. It is suggested that the preferred strategy is related to a basic principle of scaling the magnitude and duration of a velocity profile in order to achieve different movement amplitudes.     

Constraints in the performance of bimanual tasks and their expression in unskilled and skilled subjects

Unskilled and skilled subjects were asked to perform a variety of bimanual tapping tasks. Three major effects were seen. First, right-handers performed dual tasks better when the preferred hand took the “figure” and when the nonpreferred hand took the “ground” of the dual movement. This effect was not seen in left-handers. Second, subjects performed a simple slow/fast dual task better when they commenced the task with the fast rather than with the slow hand. This effect was seen in right- and lefthanders. Third, both unskilled and skilled subjects showed marked interdependence of movements such that performance of one hand was a function of movements in the other hand. The results are in agreement with a model that postulates the presence of a superordinate control mechanism that initiates action in subordinate control mechanisms, which in turn set the movement trajectories in the two hands. The results also show that attention is an important factor in the interaction between these two levels of control.     

Defining the boundary in a positioning task

Previous experiments have shown that overshoot rate in a linear positioning task is determined by the distance of the target from the boundary of the task in the direction of movement. Present experiments have served to specify distance more precisely as being relative rather than absolute, and as proximal rather than distal, and to show that the position of the boundary depends on the movements demanded by the task and not the visual and proprioceptive limits of the display. The operational boundary may be regarded as a cognitive construct by reference to which subjects locate targets.     

Visual short-term memory is influenced by haptic perception

The present study investigated whether and how visual memory and haptic perception are related. Participants were required to compare a visual reference velocity with a visual test velocity separated by a 4-s interval. During the retention interval, a fast or slow hand movement was performed. Although the hand movement was not visible, effects of the speed of the distracting body movement occurred. Slow movements resulted in a lowering of the represented visual velocity, whereas fast movements heightened the represented velocity. Subsequent experiments extended the effect to body movements that differed from the visual motion and ruled out the possibility that the effect was due to changes in visual perception or interference from semantic, verbal, and acoustic memory codes. Perhaps haptic velocity information and visual velocity information stored in short-term memory are blended.     

Speed-accuracy characteristics of saccadic eye movements

Speed-accuracy trade-off characteristic of horizontal saccadic eye movements were examined in this study. Unlike limb movements, saccadic eye movements are preprogrammed, unidimensional, and do not involve target impact. Hence, they provide an optimal test of the impulse variability account of the speed-accuracy trade-off in rapid movements. Subjects were required to alternately look at two target lights as fast and as accurately as possible for a period of 10 s. Target lights subtended angles of 5, 10, 15, and 20 degrees. By restricting target distances to less than 20 degrees of arc, the speed-accuracy relation was examined for single horizontal saccadic movements of the eye. movement of the dominant eye was tracked with an infra-red eye monitoring device. Fifty saccadic movements of the eye were recorded for each target distance and used to compute the average amplitude, duration, and velocity of eye movements, as well as, movement endpoint variability. An increase in both average velocity and movement endpoint variability with increasing movement amplitude was found. This, together with the unique features of the eye movement system, support the impulse variability account of the speed-accuracy trade-off in rapid movements.     

Space-time behavior of single and bimanual rhythmical movements: data and limit cycle model

How do space and time relate in rhythmical tasks that require the limbs to move singly or together in various modes of coordination? And what kind of minimal theoretical model could account for the observed data? Earlier findings for human cyclical movements were consistent with a nonlinear, limit cycle oscillator model (Kelso, Holt, Rubin, & Kugler, 1981) although no detailed modeling was performed at that time. In the present study, kinematic data were sampled at 200 samples/second, and a detailed analysis of movement amplitude, frequency, peak velocity, and relative phase (for the bimanual modes, in phase and antiphase) was performed. As frequency was scaled from 1 to 6 Hz (in steps of 1 Hz) using a pacing metronome, amplitude dropped inversely and peak velocity increased. Within a frequency condition, the movement's amplitude scaled directly with its peak velocity. These diverse kinematic behaviors were modeled explicitly in terms of low-dimensional (nonlinear) dissipative dynamics, with linear stiffness as the only control parameter. Data and model are shown to compare favorably. The abstract, dynamical model offers a unified treatment of a number of fundamental aspects of movement coordination and control.     

Progression-regression effects in tracking repeated patterns

Subjects used a position control system to perform compensatory tracking of a repeated input pattern. Tracking error was roughly proportional to the velocity of the input signal. Error magnitude decreased with practice and increased with the addition of a concurrent memory task. These effects can be modeled as progressive and regressive changes in how well subjects used control movement velocity and displayed error velocity to anticipate the input pattern and thereby reduce their effective time delay. The weighting of velocity cues in this model progressed with practice and regressed with the secondary task, even though the secondary task required no concurrent visual scanning or simultaneous motor response. This regression effect appears to indicate cognitive interference with the anticipation process. Stationary linear models provide a good approximation to the movement patterns; however, these models do not account for episodes of rapid pulse-like movements that were revealed in the ensemble-averaged trajectories.     

An Impulse-Timing Theory for Reciprocal Control of Muscular Activity in Rapid,Discrete Movements

Much remains to be learned about how agonist and antagonist muscles are controlled during the production of rapid, voluntary movements. In an effort to summarize a wide body of existing knowledge and stimulate future research on this subject, an impulse-timing theory is presented which attempts to predict the activity of reciprocal muscles based on certain characteristics of a movement. The basic tenet of the theory is that variables of movement time, movement distance, and inertial load have fairly predictable effects on the underlying muscular activity of the agonist and antagonist muscles during the production of rapid and discrete, voluntary movements. The theory is derived from the kinematic work of Schmidt, Zelaznik, Hawkins, Frank, and Quinn (1979) and supporting evidence from studies which have used electromyographic (EMG) recordings of agonist and antagonist muscles during rapid movements. Issues related to synergistic muscle control, central and peripheral control of reciprocal muscle activity, muscle control, and neurological disorder and the relationship between impulse-timing and mass-spring control are discussed in the final section.     

An impulse-timing theory for reciprocal control of muscular activity in rapid, discrete movements

Much remains to be learned about how agonist and antagonist muscles are controlled during the production of rapid, voluntary movements. In an effort to summarize a wide body of existing knowledge and stimulate future research on this subject, an impulse-timing theory is presented which attempts to predict the activity of reciprocal muscles based on certain characteristics of a movement. The basic tenet of the theory is that variables of movement time, movement distance and inertial load have fairly predictable effects on the underlying muscular activity of the agonist and antagonist muscles during the production of rapid and discrete, voluntary movements. The theory is derived from the kinematic work of Schmidt, Zelaznik, Hawkins, Frank and Quinn (1979) and supporting evidence from studies which have used electromyographic (EMG) recordings of agonist and antagonist muscles during rapid movements. Issues related to synergistic muscle control, central and peripheral control of reciprocal muscle activity, muscle control, and neurological disorder and the relationship between impulse-timing and mass-spring control are discussed in the final section.     

Three- to eight-month-old infants' catching under monocular and binocular vision

We report a cross-sectional and a longitudinal experiment that examined developmental changes in the relative contribution of monocular and binocular variables in the guidance of interceptive arm movements. Three- to eight-month-old infants were observed while presented with differently sized balls that approached frontally with a constant velocity under both monocular and binocular viewing conditions. Movement onset indicated that with age infants increasingly came to rely on binocular variables in controlling the timing of the interceptive arm movements. That is, from 7 to 8 months of age movement onset was independent from object size under binocular but not under monocular viewing. In contrast, binocular viewing enhanced the spatial accuracy of the interceptive arm movements at all ages. We concluded that attunement to binocular information is a key process in infants' gaining adaptive control of goal-directed arm movements. However, interceptive arm movements entail the formation of multiple on-line couplings between optic and movement variables, each of which appears to develop at its own pace.