Modeling of control and learning in a stepping motion

In a previous study (Beuter et al. 1986) the authors modeled a stepping motion using a three-body linkage with four degrees of freedom. Stepping was simulated by using three task parameters (i.e., step height, length, and duration) and sinusoidal joint angular velocity profiles. The results supported the concept of a hierarchical control structure with open-loop control during normal operation. In this study we refine the dynamic model and improve the simulation technique by incorporating the dynamics of the leg after landing, adding a foot segment to the model, and preprogramming the complete step motion using cycloids. The equations of the forces and torques developed on the ground by the foot during the landing phase are derived using the Lagrangian method. Simulation results are compared to experimental data collected on a subject stepping four times over an obstacle using a Selspot motion analysis system. A hierarchical control model that incorporates a learning process is proposed. The model allows an efficient combination of open and closed loop control strategies and involves hardwired movement segments. We also test the hypothesis of cycloidal velocity profiles in the joint programs against experimental data using a novel curve-fitting procedure based on analytical rather than numerical differentiation. The results suggest multiob-jective optimization of the joint's motion. The control and learning model proposed here will help the understanding of the mechanisms responsible for assembling selected movement segments into goaldirected movement sequences in humans.     

Muscle activation patterns of selected lower extremity muscles during stepping and cutting tasks

Lower extremity muscle activations during crossover and side step cut tasks are hypothesized to play an important role in controlling knee motion, and therefore, impact the design of knee injury prevention and rehabilitation programs. However, the contribution of lower extremity muscles to frontal and transverse plane moments during cutting tasks is unclear. The purpose of this study was to compare the muscle activation patterns of selected lower extremity muscles (vastus lateralis, medial/lateral hamstrings and medial/lateral gastrocnemius) of subjects performing a stepping down and side step cut, a stepping down and crossover cut and an equivalent straight ahead task. Ground reaction force was used to determine the cut angle, stance time and compare the lower limb loading during each task. Electromyography data during all tasks were normalized to the average activation during the straight ahead tasks to determine relative changes in muscle activation between the straight ahead and different cut styles (crossover and side step). There were no differences in the pattern of muscle activation of the vastus lateralis, or lateral hamstring muscles when comparing the cutting tasks to the equivalent straight ahead task. However, the crossover cut task resulted in significantly higher muscle activation of the medial hamstrings and lateral gastrocnemius muscles relative to both the side step cut and straight ahead tasks. These results suggest the medial/lateral hamstrings and medial/lateral gastrocnemius play a role in transverse and frontal plane control during cut tasks.     

Altered control strategy between leading and trailing leg increases knee adduction moment in the elderly while descending stairs

The aim of the study was to examine the external knee adduction moments in a group of older and younger adults while descending stairs and thus the possibility of an increased risk of knee osteoarthritis due to altered knee joint loading in the elderly. Twenty-seven older and 16 younger adults descended a purpose-built staircase. A motion capture system and a force plate were used to determine the subjects' 3D kinematics and ground reaction forces (GRF) during locomotion. Calculation of the leg kinematics and kinetics was done by means of a rigid, three-segment, 3D leg model. In the initial portion of the support phase, older adults showed a more medio-posterior GRF vector relative to the ankle joint, leading to lower ankle joint moments (P<0.05). At the knee, the older adults demonstrated a more medio-posterior directed GRF vector, increasing in knee flexion and adduction in the second part of the single support phase (P<0.05). Further, GRF magnitude was lower in the initial and higher in the mid-portions of the support phase for the elderly (P<0.05). The results show that older adults descend stairs by using the trailing leg before the initiation of the double support phase more compared to the younger ones. The consequence of this altered control strategy while stepping down is a more medially directed GRF vector increasing the magnitude of external knee adduction moment in the elderly. The observed changes between leading and trailing leg in the elderly may cause a redistribution of the mechanical load at the tibiofemoral joint, affecting the initiation and progression of knee osteoarthritis in the elderly.     

Predicting dynamic postural instability using center of mass time-to-contact information

Our purpose was to determine whether spatiotemporal measures of center of mass motion relative to the base of support boundary could predict stepping strategies after upper-body postural perturbations in humans. We expected that inclusion of center of mass acceleration in such time-to-contact (TtC) calculations would give better predictions and more advanced warning of perturbation severity. TtC measures were compared with traditional postural variables, which do not consider support boundaries, and with an inverted pendulum model of dynamic stability developed by Hof et al. [2005. The condition for dynamic stability. Journal of Biomechanics 38, 1-8]. A pendulum was used to deliver sequentially increasing perturbations to 10 young adults, who were strapped to a wooden backboard that constrained motion to sagittal-plane rotation about the ankle joint. Subjects were instructed to resist the perturbations, stepping only if necessary to prevent a fall. Peak center of mass and center of pressure velocity and acceleration demonstrated linear increases with postural challenge. In contrast, boundary-relevant minimum TtC values decreased nonlinearly with postural challenge, enabling prediction of stepping responses using quadratic equations. When TtC calculations incorporated center of mass acceleration, the quadratic fits were better and gave more accurate predictions of the TtC values that would trigger stepping responses. In addition, TtC minima occurred earlier with acceleration inclusion, giving more advanced warning of perturbation severity. Our results were in agreement with TtC predictions based on Hof's model, and suggest that TtC may function as a control parameter, influencing the postural control system's decision to transition from a stationary base of support to a stepping strategy.     

Dynamic motion planning for the design of robotic gait rehabilitation

In this paper we examine a method to control the stepping motion of a paralyzed person suspended over a treadmill using a robot attached to the pelvis. A leg swing motion is created by moving the pelvis without contact with the legs. The problem is formulated as an optimal control problem for an underactuated articulated chain. The optimal control problem is converted into a discrete parameter optimization and an efficient gradient-based algorithm is used to solve it. Motion capture data from an unimpaired human subject is compared to the simulation results from the dynamic motion optimization. Our results suggest that it is feasible to drive repetitive stepping on a treadmill by a paralyzed person by assisting in torso movement alone. The optimized, pelvic motion strategies are comparable to "hip-hiking" gait strategies used by people with lower limb prostheses or hemiparesis. The resulting motions can be found at the web site http://ww.eng.uci.edu/-chwang/project/stepper/stepper.html.     

Linear approximations for swing leg motion during gait

The applicability of a linear systems analysis of two-dimensional swing leg motion was investigated. Two different linear systems were developed. A linear time-varying system was developed by linearizing the nonlinear equations describing swing leg motion about a set of nominal system and control trajectories. Linear time invariant systems were developed by linearizing about three different fixed limb positions. Simulations of swing leg motion were performed with each of these linear systems. These simulations were compared to previously performed nonlinear simulations of two-dimensional swing leg motion and the actual subject motion. Additionally, a linear system analysis was used to gain some insight into the interdependency of the state variables and controls. It was shown that the linear time varying approximation yielded an accurate representation of limb motion for the thigh and shank but with diminished accuracy for the foot. In contrast, all the linear time invariant systems, if used to simulate more than a quarter of the swing phase, yielded generally inaccurate results for thigh shank and foot motion.     

Voluntary stepping behavior under single- and dual-task conditions in chronic stroke survivors: A comparison between the involved and uninvolved legs

ObjectiveIf balance is lost, quick step execution can prevent falls. Research has shown that speed of voluntary stepping was able to predict future falls in old adults. The aim of the study was to investigate voluntary stepping behavior, as well as to compare timing and leg push-off force–time relation parameters of involved and uninvolved legs in stroke survivors during single- and dual-task conditions. We also aimed to compare timing and leg push-off force–time relation parameters between stroke survivors and healthy individuals in both task conditions.MethodsTen stroke survivors performed a voluntary step execution test with their involved and uninvolved legs under two conditions: while focusing only on the stepping task and while a separate attention-demanding task was performed simultaneously. Temporal parameters related to the step time were measured including the duration of the step initiation phase, the preparatory phase, the swing phase, and the total step time. In addition, force–time parameters representing the push-off power during stepping were calculated from ground reaction data and compared with 10 healthy controls.ResultsThe involved legs of stroke survivors had a significantly slower stepping time than uninvolved legs due to increased swing phase duration during both single- and dual-task conditions. For dual compared to single task, the stepping time increased significantly due to a significant increase in the duration of step initiation. In general, the force time parameters were significantly different in both legs of stroke survivors as compared to healthy controls, with no significant effect of dual compared with single-task conditions in both groups.ConclusionsThe inability of stroke survivors to swing the involved leg quickly may be the most significant factor contributing to the large number of falls to the paretic side. The results suggest that stroke survivors were unable to rapidly produce muscle force in fast actions. This may be the mechanism of delayed execution of a fast step when balance is lost, thus increasing the likelihood of falls in stroke survivors.     

Parameter optimization model of learning in stepping motion

In this study we combine the representation of motion by a finite number of hardwired functions with parameter optimization to model learning during a stepping motion. Representation of experimental kinematic data by a finite number of predetermined functions and undetermined coefficients was analyzed. Least squares approximation was used to represent experimental data of stepping motions over obstacles of different heights. Functional relationships between coefficients and obstacles heights were also obtained. Learning of stepping over an obstacle was then formulated as a finite dimensional optimization problem. The pattern of foot path, and joint angles trajectories obtained by this learning model, were then compared to the experimental data. The results of the data fitting analysis and of the optimization process as a model for motion learning, indicate that motion can be adequately represented by a set of hardwired functions, and a finite number of task dependent coefficients.     

Muscle pre- and coactivity during downward stepping are associated with leg stiffness in aging.

We have previously reported that elderly compared to young women executed downward stepping with substantially greater leg stiffness. Because antagonist muscle coactivity increases joint stiffness we hypothesized that increased leg stiffness in aging is associated with increased muscle coactivity. We also explored the possibility that the magnitude of the preparatory muscle activity preceding impact also differed between young and old subjects. Young (n=11, 20. 8 yr) and old (n=12, 69 yr) women performed downward stepping from a platform set at 20% body height. The leg was modeled as a simple mass-spring system. From video and ground reaction force data leg stiffness was computed as the ratio of force under the foot and the linear shortening of the limb. EMG activity of the vastus lateralis, biceps femoris, gastrocnemius lateralis, and tibialis anterior were recorded with a telemetric system. Elders compared to young subjects had 64% greater leg stiffness during downward stepping. Muscle activity over a 200-ms period preceding touch down was 136% greater in elderly than in young subjects. Biceps femoris and tibialis anterior coactivity during ground contact was 120% greater in the elders. Muscle pre- and coactivity, respectively, accounted for about 50% of the variance in leg stiffness. In conclusion, elderly people elevate muscle pre- and coactivity during downward stepping to stiffen the leg in compensation for impaired neuromotor functions.     

Three-dimensional measurement of rearfoot motion during running

Excessive ranges of motion during running have been speculated to be connected to injuries to the lower extremities. Movement of the foot and lower leg has commonly been studied with two-dimensional techniques. However, differences in the alignment of the longitudinal axis of the foot with the camera axis will produce measurement errors for projected angles of the lower extremities. A three-dimensional approach would not have this limitation. The purpose of this study is to present a three-dimensional model for calculation of angles between lower leg and foot, lower leg and ground, and foot and ground, and to compare results from treadmill running derived from this model with results derived from a two-dimensional model for different alignment angles between foot axis and camera axis. A two camera Selspot system was used to obtain three-dimensional information on motion of the studied segments. It was found that several two-dimensional variables measured from a posterior view are very sensitive to the alignment angle between the foot and the camera axis. Some variables change as much as 1 degrees for every 2 degrees of change of the alignment angle. The large influence of rotations other than the measured one in two-dimensional measurements makes advisable the use of a three-dimensional model when studying motion between foot and lower leg during running.     

Rotational locomotion by the cockroach Blattella germanica

Locomotion on complex substrata can be expressed in a plane by two geometric components of body movement: linear locomotion and rotational locomotion. This study examined pure rotation by analysing the geometry of leg movements and stepping patterns during the courtship turns of male Blattella germanica. Strict rotation or translation by an insect requires that each side of the body cover equal distance with respect to the substrate. There are three mechanisms by which the legs can maintain this equality: frequency of stepping, magnitude of the leg arcs relative to the body and the degree to which legs flex and extend during locomotion. During the courtship behaviour of Blattella germanica selected males executed turns involving body rotation along with leg movements in which the legs on the outside of the turn swung through greater average arcs than those on the inside of the turn. This difference should have resulted in a translation component. However, legs on the inside of the turn compensated by flexion and extension movements which were greater than those of opposing legs. The net effect was that both sides of the body covered equal average ground. These cockroaches used a wide variety of stepping combinations to effect rotation. The frequency of these combinations was compared to an expected frequency distribution of stepping combinations and further to an expected frequency of these stepping combinations used for straight walking. These comparisons demonstrated a similarity between interleg coordination during straight walking and that during turning in place.     

Lower extremity control during turns initiated with and without hip external rotation

The pirouette turn is often initiated in neutral and externally rotated hip positions by dancers. This provides an opportunity to investigate how dancers satisfy the same mechanical objectives at the whole-body level when using different leg kinematics. The purpose of this study was to compare lower extremity control strategies during the turn initiation phase of pirouettes performed with and without hip external rotation. Skilled dancers (n=5) performed pirouette turns with and without hip external rotation. Joint kinetics during turn initiation were determined for both legs using ground reaction forces (GRFs) and segment kinematics. Hip muscle activations were monitored using electromyography. Using probability-based statistical methods, variables were compared across turn conditions as a group and within-dancer. Despite differences in GRFs and impulse generation between turn conditions, at least 90% of each GRF was aligned with the respective leg plane. A majority of the net joint moments at the ankle, knee, and hip acted about an axis perpendicular to the leg plane. However, differences in shank alignment relative to the leg plane affected the distribution of the knee net joint moment when represented with respect to the shank versus the thigh. During the initiation of both turns, most participants used ankle plantar flexor moments, knee extensor moments, flexor and abductor moments at the push leg׳s hip, and extensor and abductor moments at the turn leg׳s hip. Representation of joint kinetics using multiple reference systems assisted in understanding control priorities.     

The investigation of control mechanisms of stepping rhythm in human in the air-stepping conditions during passive and voluntary leg movements

In unloading condition the degree of activation of the central stepping program was investigated during passive leg movements in healthy subjects, as well as the excitability of spinal motoneurons during passive and voluntary stepping movement. Passive stepping movements with characteristics maximally approximated to those during voluntary stepping were accomplished by experimenter. The comparison of the muscle activity bursts during voluntary and imposed movements was made. In addition to that the influence of artificially created loading onto the foot to the leg movement characteristics was analyzed. Spinal motoneuron excitability was estimated by means of evaluation of amplitude modulation of the soleus H-reflex. The changes of H-reflexes under the fixation of knee or hip joints were also studied. In majority of subjects the passive movements were accompanied by bursts of EMG activity of hip muscles (and sometimes of knee muscles), which timing during step cycle was coincided with burst timing of voluntary step cycle. In many cases the bursts of EMG activity during passive movements exceeded activity in homonymous muscles during voluntary stepping. The foot loading imitation exerted essential influence on distal parts of moving extremity during voluntary as well passive movements, that was expressed in the appearance of movements in the ankle joint and accompanied by emergence and increasing of phasic EMG activity of shank muscles. The excitability of motoneurons during passive movements was greater then during voluntary ones. The changes and modulation of H-reflex throughout the step cycle without restriction of joint mobility and during exclusion of hip joint mobility were similar. The knee joint fixation exerted the greater influence. It is supposed that imposed movements activate the same mechanisms of rhythm generation as a supraspinal commands during voluntary movements. In the conditions of passive movements the presynaptic inhibition depend on afferent influences from moving leg in the most degree then on central commands. It seems that afferent inputs from pressure receptors of foot in the condition of "air-stepping" actively interact with central program of stepping and, irrespective of type of the performing movements (voluntary or passive), form the final pattern activity.     

Head mounted displays for capturing head kinematics in postural tasks

Tracking head motion in a simple, portable and accurate manner during performance of postural tasks in a virtual reality environment could have important implications for investigating normal and pathological head kinematics. We investigated concurrent validity of head tracking of two Head Mounted Displays (HMDs), Oculus Rift and HTC Vive, vs. a gold-standard motion capture system (Qualisys). Head kinematics of N = 20 healthy young adults was quantified during static and dynamic postural tasks. While wearing the Oculus Rift or HTC Vive, participants observed moving stars (static tasks) or a flying ball (dynamic task). Head kinematics were recorded simultaneously by the Rift or Vive and Qualisys camera system. We calculated head directional path, acceleration in 6 directions and volume of translation movement. Intra-Class Correlations (ICC) and 95% Limits of agreement were calculated. Most ICC values were around 0.9 with several at 0.99 indicating excellent agreement between the HMDs and Qualisys. Weaker agreement was observed for vertical displacement during a static task and moderate agreement was observed pitch and yaw displacement during a dynamic task. A negative bias of a small magnitude (indicating more movement in VR) was observed for most variables in static tasks, while a positive bias was observed for most variables in the dynamic task (indicating less movement in VR). Our results generally support the concurrent validity of Oculus Rift and HTC Vive head tracking during static and dynamic standing tasks in healthy young adults. Specific task- and direction-dependent differences should be considered when planning measurement studies using these novel tools.     

Balance control in stepping down expected and unexpected level changes

Stepping down an elevation in ongoing gait is a common task that can cause falls when the level change is unexpected. The aim of this study was to compare expected and unexpected stepping down. We hypothesized that unexpected stepping would lead to loss of control over the movement and potentially falls due to buckling of the leading leg at landing. Ten male subjects repeatedly walked over a platform on which they stepped down an expected 10-cm height difference. In 5 out of 50 trials, the height difference was encountered unexpectedly early. Kinematics and ground reaction forces under both feet were measured during the stride in which the height difference was negotiated. Stepping down involved a substantial increase in forward horizontal and angular momenta (approximately 40 N s and 20 N ms). In expected stepping down, step length was significantly increased (17%), which allowed control of these forward horizontal and angular momenta immediately following landing. In unexpected stepping down, the time between expected ground contact and actual ground contact (110 ms) appeared too short to substantially adjust leg movement and increase step length. Although buckling of the leg did not occur, presumably due to its more vertical orientation at landing, momentum could not be sufficiently attenuated at landing, but a fall was prevented by a rapid step of the trailing limb. The lack of control of momentum might cause a fall, when the capacity to make such a rapid step falls short, as in the elderly, or when the height difference is larger.     

A Neuro-Mechanical Model Explaining the Physiological Role of Fast and Slow Muscle Fibres at Stop and Start of Stepping of an Insect Leg

Stop and start of stepping are two basic actions of the musculo-skeletal system of a leg. Although they are basic phenomena, they require the coordinated activities of the leg muscles. However, little is known of the details of how these activities are generated by the interactions between the local neuronal networks controlling the fast and slow muscle fibres at the individual leg joints. In the present work, we aim at uncovering some of those details using a suitable neuro-mechanical model. It is an extension of the model in the accompanying paper and now includes all three antagonistic muscle pairs of the main joints of an insect leg, together with their dedicated neuronal control, as well as common inhibitory motoneurons and the residual stiffness of the slow muscles. This model enabled us to study putative processes of intra-leg coordination during stop and start of stepping. We also made use of the effects of sensory signals encoding the position and velocity of the leg joints. Where experimental observations are available, the corresponding simulation results are in good agreement with them. Our model makes detailed predictions as to the coordination processes of the individual muscle systems both at stop and start of stepping. In particular, it reveals a possible role of the slow muscle fibres at stop in accelerating the convergence of the leg to its steady-state position. These findings lend our model physiological relevance and can therefore be used to elucidate details of the stop and start of stepping in insects, and perhaps in other animals, too.     

Automatic imitation in budgerigars

A fully automated procedure, involving computer-controlled stimulus presentation and computer-recorded response measurement, was used for the first time to study imitation in non-human animals. After preliminary training to peck and step on a manipulandum, budgerigars were given a discrimination task in which they were rewarded with food for pecking during observation of pecking and for stepping during observation of stepping (Compatible group), or for pecking while observing stepping and for stepping while observing pecking (Incompatible group). The Incompatible group, which had to counter-imitate for food reward, showed weaker discrimination performance than the Compatible group. This suggests that, like humans, budgerigars are subject to 'automatic imitation'; they cannot inhibit online the tendency to imitate pecking and/or stepping, even when imitation of these behaviours interferes with the performance of an ongoing task. The difference between the two groups persisted over 10 test sessions, but the Incompatible group eventually acquired the discrimination, making more counter-imitative than imitative responses in the final sessions. These results are consistent with the associative sequence learning model, which suggests that, across species, the development of imitation and the mirror system depends on sensorimotor experience and phylogenetically ancient mechanisms of associative learning.     

The investigation of locomotor activity control system in humans under the air-stepping conditions during passive and voluntary leg movements

The degree of activation of the central stepping program during passive leg movement was studied in healthy subjects under unloading conditions; the excitability of spinal motoneurons was studied during passive and voluntary stepping movements. Passive stepping movements with characteristics maximally close to those during voluntary stepping were accomplished by the experimenter. The bursts of muscular activity during voluntary and imposed stepping movements were compared. In addition, the influence on the leg movement of artificially created loading onto the foot was studied. The excitability of spinal motoneurons was estimated by the amplitude of modulation of the m. soleus H reflex. Changes in the H reflex (Hoffmann’s reflex) after fixation of the knee and hip joints were also studied. In most subjects, passive movements were accompanied by bursts of electromyographic (EMG) activity in the hip muscles (sometimes in shank muscles); the timing of the EMG burst during the step cycle coincided with the burst’s timing during voluntary stepping. In many cases, the bursts in EMG activity exceeded the activity of homonymous muscles during voluntary stepping. Simulation of foot loading influenced significantly the distal part of the moving extremity during both voluntary and passive movements, which was expressed in the appearance of movements in the ankle joint and an increase in the phasic EMG activity of the shank muscles. The excitability of motoneurons during passive movements was higher than during voluntary movements. Changes and modulation of the H reflex throughout the step cycle were similar without restriction of joint mobility and without hip joint mobility. Fixation of the knee joint was of great importance. It is supposed that imposed movements activate the same mechanisms of rhythm generation as supraspinal commands during voluntary movements. During passive movements, presynaptic inhibition depends mostly on the afferent influences from the moving leg rather than on the central commands. Under the conditions of “air-stepping,” the afferent influences from the foot pressure receptors are likely to interact actively with the central program of stepping and to determine the final activity pattern irrespective of the movement type (voluntary or passive).     

Balance Asymmetry in Parkinson’s Disease and Its Contribution to Freezing of Gait

Balance control (the ability to maintain an upright posture) is asymmetrically controlled in a proportion of patients with Parkinson’s disease. Gait asymmetries have been linked to the pathophysiology of freezing of gait. We speculate that asymmetries in balance could contribute to freezing by a) hampering the unloading of the stepping leg and/or b) leading to a preferred stance leg during gait, which then results in asymmetric gait. To investigate this, we examined the relationship between balance control and weight-bearing asymmetries and freezing. We included 20 human patients with Parkinson (tested OFF medication; nine freezers) and nine healthy controls. Balance was perturbed in the sagittal plane, using continuous multi-sine perturbations, applied by a motion platform and by a force at the sacrum. Applying closed-loop system identification techniques, relating the body sway angle to the joint torques of each leg separately, determined the relative contribution of each ankle and hip joint to the total amount of joint torque. We also calculated weight-bearing asymmetries. We determined the 99-percent confidence interval of weight-bearing and balance-control asymmetry using the responses of the healthy controls. Freezers did not have larger asymmetries in weight bearing (p = 0.85) nor more asymmetrical balance control compared to non-freezers (p = 0.25). The healthy linear one-to-one relationship between weight bearing and balance control was significantly different for freezers and non-freezers (p = 0.01). Specifically, non-freezers had a significant relationship between weight bearing and balance control (p = 0.02), whereas this relation was not significant for freezers (p = 0.15). Balance control is asymmetrical in most patients (about 75 percent) with Parkinson’s disease, but this asymmetry is not related to freezing. The relationship between weight bearing and balance control seems to be less pronounced in freezers, compared to healthy controls and non-freezers. However, this relationship should be investigated further in larger groups of patients.     

Three-dimensional modular control of human walking

Recent studies have suggested that complex muscle activity during walking may be controlled using a reduced neural control strategy organized around the co-excitation of multiple muscles, or modules. Previous computer simulation studies have shown that five modules satisfy the sagittal-plane biomechanical sub-tasks of 2D walking. The present study shows that a sixth module, which contributes primarily to mediolateral balance control and contralateral leg swing, is needed to satisfy the additional non-sagittal plane demands of 3D walking. Body support was provided by Module 1 (hip and knee extensors, hip abductors) in early stance and Module 2 (plantarflexors) in late stance. In early stance, forward propulsion was provided by Module 4 (hamstrings), but net braking occurred due to Modules 1 and 2. Forward propulsion was provided by Module 2 in late stance. Module 1 accelerated the body medially throughout stance, dominating the lateral acceleration in early stance provided by Modules 4 and 6 (adductor magnus) and in late stance by Module 2, except near toe-off. Modules 3 (ankle dorsiflexors, rectus femoris) and 5 (hip flexors and adductors except adductor magnus) accelerated the ipsilateral leg forward in early swing whereas Module 4 decelerated the ipsilateral leg prior to heel-strike. Finally, Modules 1, 4 and 6 accelerated the contralateral leg forward prior to and during contralateral swing. Since the modules were based on experimentally measured muscle activity, these results provide further evidence that a simple neural control strategy involving muscle activation modules organized around task-specific biomechanical functions may be used to control complex human movements.