Left-Right Locomotor Coordination in Human Neonates.

Terrestrial locomotion requires coordinated bilateral activation of limb muscles, with left-right alternation in walking or running, and synchronous activation in hopping or skipping. The neural mechanisms involved in interlimb coordination at birth are well known in different mammalian species, but less so in humans. Here, 46 neonates (of either sex) performed bilateral and unilateral stepping with one leg blocked in different positions. By recording EMG activities of lower-limb muscles, we observed episodes of left-right alternating or synchronous coordination. In most cases, the frequency of EMG oscillations during sequences of consecutive steps was approximately similar between the two sides, but in some cases it was considerably different, with episodes of 2:1 interlimb coordination and episodes of activity deletions on the blocked side. Hip position of the blocked limb significantly affected ipsilateral, but not contralateral, muscle activities. Thus, hip extension backward engaged hip flexor muscle, and hip flexion engaged hip extensors. Moreover, the sudden release of the blocked limb in the posterior position elicited the immediate initiation of the swing phase of the limb, with hip flexion and a burst of an ankle flexor muscle. Extensor muscles showed load responses at midstance. The variable interlimb coordination and its incomplete sensory modulation suggest that the neonatal locomotor networks do not operate in the same manner as in mature locomotion, also because of the limited cortical control at birth. These neonatal mechanisms share many properties with spinal mammalian preparations (i.e., independent pattern generators for each limb, and for flexor and extensor muscles, load, and hip position feedback).SIGNIFICANCE STATEMENT Bilateral coupling and reciprocal activation of flexor and extensor burst generators represent the fundamental mechanisms used by mammalian limbed locomotion. Considerable progress has been made in deciphering the early development of the spinal networks and left-right coordination in different mammals, but less is known about human newborns. We compared bilateral and unilateral stepping in human neonates, where cortical control is still underdeveloped. We found neonatal mechanisms that share many properties with spinal mammalian preparations (i.e., independent pattern generators for each limb, the independent generators for flexor and extensor muscles, load, and hip-position feedback. The variable interlimb coordination and its incomplete sensory modulation suggest that the human neonatal locomotor networks do not operate in the same manner as in mature locomotion.

Club-mosses (Diphasiastrum, Lycopodiaceae) from the Far East - Introgression and possible cryptic speciation.

Hybridization occurs often in the genus Diphasiastrum (Lycopodiaceae), which corroborates reports for the two other recognized lycophyte families, Isoëtaceae and Selaginellaceae. Here we investigate the case of D. alpinum and D. sitchense from the Russian Far East (Kamchatka). Their hybrid, D. × takedae, was morphologically recognizable in 16 out of 22 accessions showing molecular signatures of hybridization; the remaining accessions displayed the morphology of either D. alpinum (3) or D. sitchense (3). We sequenced markers for chloroplast microsatellites (cp, 175 accessions from Kamchatka) and for the two nuclear markers RPB and LFY (175 and 152 accessions). A selection of 42 accessions, including all hybrid accessions, was analysed via genotyping by sequencing (GBS). We found multiple, but apparently uniparental hybridization, clearly characterized by a deviating group of haplotypes for D. sitchense and all hybrids. All accessions showing molecular signatures of hybridization in nuclear markers revealed the parental haplotype of D. sitchense, however only the LFY marker differentiated between the parent species. GBS, including 69,819 quality-filtered single nucleotid polymorphisms, unambiguously identified the hybrids and revealed introgression to occur. Most of the hybrids were F1, but three turned out to be backcrosses with D. alpinum (one) and with D. sitchense (two). These observations are in contrast to prior findings on three European species and their intermediates where all three hybrids turned out to be independent F1 crosses without evidence of recent backcrossing. In this study, backcrossing was detected, which indicates a limited fertility of the hybrid taxon D. × takedae. A comparison of accessions of Kamchatkian D. alpinum with plants from Europe indicated possible cryptic speciation. Accessions from the Far East had (i) a lower DNA content (7.0 vs. 7.5 pg/2C), (ii) different prevailing cp haplotypes, and (iii) RPB genotypes, and (iv) a clearly different SNP pattern in GBS. Diphasiastrum sitchense and the similar D. nikoënse, for the latter additional accessions from Japan were investigated, appeared as forms of one diverse species, sharing genotypes in both nuclear markers, although chloroplast haplotypes and DNA content show slight variations.

Drawing ellipses in water: evidence for dynamic constraints in the relation between velocity and path curvature.

Several types of continuous human movements comply with the so-called Two-Thirds Power Law (2/3-PL) stating that velocity (V) is a power function of the radius of curvature (R) of the endpoint trajectory. The origin of the 2/3-PL has been the object of much debate. An experiment investigated further this issue by comparing two-dimensional drawing movements performed in air and water. In both conditions, participants traced continuously quasi-elliptic trajectories (period T = 1.5 s). Other experimental factors were the movement plane (horizontal/vertical), and whether the movement was performed free-hand, or by following the edge of a template. In all cases a power function provided a good approximation to the V-R relation. The main result was that the exponent of the power function in water was significantly smaller than in air. This appears incompatible with the idea that the power relationship depends only on kinematic constraints and suggests a significant contribution of dynamic factors. We argue that a satisfactory explanation of the observed behavior must take into account the interplay between the properties of the central motor commands and the visco-elastic nature of the mechanical plant that implements the commands.

Visual gravity cues in the interpretation of biological movements: neural correlates in humans.

Our visual system takes into account the effects of Earth gravity to interpret biological motion (BM), but the neural substrates of this process remain unclear. Here we measured functional magnetic resonance (fMRI) signals while participants viewed intact or scrambled stick-figure animations of walking, running, hopping, and skipping recorded at normal or reduced gravity. We found that regions sensitive to BM configuration in the occipito-temporal cortex (OTC) were more active for reduced than normal gravity but with intact stimuli only. Effective connectivity analysis suggests that predictive coding of gravity effects underlies BM interpretation. This process might be implemented by a family of snapshot neurons involved in action monitoring.

Coordination of intrinsic and extrinsic foot muscles during walking.

PURPOSE: The human foot undergoes complex deformations during walking due to passive tissues and active muscles. However, based on prior recordings it is unclear if muscles that contribute to flexion/extension of the metatarsophalangeal (MTP) joints are activated synchronously to modulate joint impedance, or sequentially to perform distinct biomechanical functions. We investigated the coordination of MTP flexors and extensors with respect to each other, and to other ankle-foot muscles. METHODS: We analyzed surface electromyographic (EMG) recordings of intrinsic and extrinsic foot muscles for healthy individuals during level treadmill walking, and also during sideways and tiptoe gaits. We computed stride-averaged EMG envelopes and used the timing of peak muscle activity to assess synchronous vs. sequential coordination. RESULTS: We found that peak MTP flexor activity occurred significantly before peak MTP extensor activity during walking (P < 0.001). The period around stance-to-swing transition could be roughly characterized by sequential peak muscle activity from the ankle plantarflexors, MTP flexors, MTP extensors, and then ankle dorsiflexors. We found that foot muscles that activated synchronously during forward walking tended to dissociate during other locomotor tasks. For instance, extensor hallucis brevis and extensor digitorum brevis muscle activation peaks decoupled during sideways gait. CONCLUSIONS: The sequential peak activity of MTP flexors followed by MTP extensors suggests that their biomechanical contributions may be largely separable from each other and from other extrinsic foot muscles during walking. Meanwhile, the task-specific coordination of the foot muscles during other modes of locomotion indicates a high-level of specificity in their function and control.

Changes in the spinal segmental motor output for stepping during development from infant to adult.

Human stepping movements emerge in utero and show several milestones during development to independent walking. Recently, imaging has become an essential tool for investigating the development and function of pattern generation networks in the spinal cord. Here we examine the development of the spinal segmental output by mapping the distribution of motoneuron activity in the lumbosacral spinal cord during stepping in newborns, toddlers, preschoolers, and adults. Newborn stepping is characterized by an alternating bilateral motor output with only two major components that are active at all lumbosacral levels of the spinal cord. This feature was similar across different cycle durations of neonate stepping. The alternating spinal motor output is consistent with a simpler organization of neuronal networks in neonates. Furthermore, a remarkable feature of newborn stepping is a higher overall activation of lumbar versus sacral segments, consistent with a rostrocaudal excitability gradient. In toddlers, the stance-related motor pool activity migrates to the sacral cord segments, while the lumbar motoneurons are separately activated at touchdown. In the adult, the lumbar and sacral patterns become more dissociated with shorter activation times. We conclude that the development of human locomotion from the neonate to the adult starts from a rostrocaudal excitability gradient and involves a gradual functional reorganization of the pattern generation circuitry.

Can modular strategies simplify neural control of multidirectional human locomotion?

Each human lower limb contains over 50 muscles that are coordinated during locomotion. It has been hypothesized that the nervous system simplifies muscle control through modularity, using neural patterns to activate muscles in groups called synergies. Here we investigate how simple modular controllers based on invariant neural primitives (synergies or patterns) might generate muscle activity observed during multidirectional locomotion. We extracted neural primitives from unilateral electromyographic recordings of 25 lower limb muscles during five locomotor tasks: walking forward, backward, leftward and rightward, and stepping in place. A subset of subjects also performed five variations of forward (unidirectional) walking: self-selected cadence, fast cadence, slow cadence, tiptoe, and uphill (20% incline). We assessed the results in the context of dimensionality reduction, defined here as the number of neural signals needing to be controlled. For an individual task, we found that modular architectures could theoretically reduce dimensionality compared with independent muscle control, but we also found that modular strategies relying on neural primitives shared across different tasks were limited in their ability to account for muscle activations during multi- and unidirectional locomotion. The utility of shared primitives may thus depend on whether they can be adapted for specific task demands, for instance, by means of sensory feedback or by being embedded within a more complex sensorimotor controller. Our findings indicate the need for more sophisticated formulations of modular control or alternative motor control hypotheses in order to understand muscle coordination during locomotion.

Function dictates the phase dependence of vision during human locomotion.

In human and animal locomotion, sensory input is thought to be processed in a phase-dependent manner. Here we use full-field transient visual scene motion toward or away from subjects walking on a treadmill. Perturbations were presented at three phases of walking to test 1) whether phase dependence is observed for visual input and 2) whether the nature of phase dependence differs across body segments. Results demonstrated that trunk responses to approaching perturbations were only weakly phase dependent and instead depended primarily on the delay from the perturbation. Recording of kinematic and muscle responses from both right and left lower limb allowed the analysis of six distinct phases of perturbation effects. In contrast to the trunk, leg responses were strongly phase dependent. Leg responses during the same gait cycle as the perturbation exhibited gating, occurring only when perturbations were applied in midstance. In contrast, during the postperturbation gait cycle, leg responses occurred at similar response phases of the gait cycle over a range of perturbation phases. These distinct responses reflect modulation of trunk orientation for upright equilibrium and modulation of leg segments for both hazard accommodation/avoidance and positional maintenance on the treadmill. Overall, these results support the idea that the phase dependence of responses to visual scene motion is determined by different functional tasks during walking.

Patterned control of human locomotion.

There is much experimental evidence for the existence of biomechanical constraints which simplify the problem of control of multi-segment movements. In addition, it has been hypothesized that movements are controlled using a small set of basic temporal components or activation patterns, shared by several different muscles and reflecting global kinematic and kinetic goals. Here we review recent studies on human locomotion showing that muscle activity is accounted for by a combination of few basic patterns, each one timed at a different phase of the gait cycle. Similar patterns are involved in walking and running at different speeds, walking forwards or backwards, and walking under different loading conditions. The corresponding weights of distribution to different muscles may change as a function of the condition, allowing highly flexible control. Biomechanical correlates of each activation pattern have been described, leading to the hypothesis that the co-ordination of limb and body segments arises from the coupling of neural oscillators between each other and with limb mechanical oscillators. Muscle activations need only intervene during limited time epochs to force intrinsic oscillations of the system when energy is lost.

Influence of pregnancy related anthropometric changes on plantar pressure distribution during gait-A follow-up study.

BACKGROUND: As foot constitutes the base of support for the whole body, the pregnancy-related anthropometric changes can result in adaptive plantar pressure alterations. The present study aimed to investigate how pregnancy affects foot loading pattern in gait, and if it is related to body adjustments to growing foetus that occur in the course of pregnancy. METHODS: A prospective longitudinal study included 30 women. Three experimental sessions in accordance with the same procedure were carried out in the first, second and third trimesters of pregnancy. First, the anthropometric measures of the body mass and waist circumference were taken. Then walking trials at a self-selected speed along a ~6-m walkway were registered with the FreeMED force platform (Sensor Medica, Italy). Vertical foot pressure was recorded by the force plate located in the middle of the walkway. FINDINGS: The correlation of individual foot loading parameters across different trimesters was relatively high. Nevertheless, our results revealed a longitudinal foot arch flattening with the strongest effect in late pregnancy (P = 0.01). The anthropometric characteristics also influenced the foot loading pattern depending on the phase of pregnancy. In particular, arch flattening correlated with the body mass in all trimesters (r≥0.44, P≤0.006) while the medial-lateral loading index correlated only in the first (r = 0.45, P = 0.005) and second (r = 0.36, P = 0.03) trimesters. Waist circumference changes significantly influenced dynamic arch flattening but only in the late pregnancy (r≥0.46, P≤0.004). In the third trimester, a small though significant increase in the right foot angle was observed (P = 0.01). INTERPRETATION: The findings provided the characteristics of the relative foot areas loading throughout pregnancy. Growing abdominal size increases the risk of medial arch flattening, which can result in less stable gait. The observed increase in foot angle in late pregnancy may constitute a strategy to enhance gait stability.

Smooth changes in the EMG patterns during gait transitions under body weight unloading.

During gradual speed changes, humans exhibit a sudden discontinuous switch from walking to running at a specific speed, and it has been suggested that different gaits may be associated with different functioning of neuronal networks. In this study we recorded the EMG activity of leg muscles at slow increments and decrements in treadmill belt speed and at different levels of body weight unloading. In contrast to normal walking at 1 g, at lower levels of simulated gravity (<0.4 g) the transition between walking and running was generally gradual, without systematic abrupt changes in either intensity or timing of EMG patterns. This phenomenon depended to a limited extent on the gravity simulation technique, although the exact level of the appearance of smooth transitions (0.4-0.6 g) tended to be lower for the vertical than for the tilted body weight support system. Furthermore, simulations performed with a half-center oscillator neuromechanical model showed that the abruptness of motor patterns at gait transitions at 1 g could be predicted from the distinct parameters anchored already in the normal range of walking and running speeds, whereas at low gravity levels the parameters of the model were similar for the two human gaits. A lack of discontinuous changes in the pattern of speed-dependent locomotor characteristics in a hypogravity environment is consistent with the idea of a continuous shift in the state of a given set of central pattern generators, rather than the activation of a separate set of central pattern generators for each distinct gait.

Impulses of activation but not motor modules are preserved in the locomotion of subacute stroke patients.

It has been hypothesized that the coordinated activation of muscles is controlled by the central nervous system by means of a small alphabet of control signals (also referred to as activation signals) and motor modules (synergies). We analyzed the locomotion of 10 patients recently affected by stroke (maximum of 20 wk) and compared it with that of healthy controls. The aim was to assess whether the walking of subacute stroke patients is based on the same motor modules and/or activation signals as healthy subjects. The activity of muscles of the lower and upper limb and the trunk was measured and used for extracting motor modules. Four modules were sufficient to explain the majority of variance in muscle activation in both controls and patients. Modules from the affected side of stroke patients were different from those of healthy controls and from the unaffected side of stroke patients. However, the activation signals were similar between groups and between the affected and unaffected side of stroke patients, and were characterized by impulses at specific time instants within the gait cycle, underlying an impulsive controller of gait. In conclusion, motor modules observed in healthy subjects during locomotion are different from those used by subacute stroke patients, despite similar impulsive activation signals. We suggest that this pattern is consistent with a neuronal network in which the timing of activity generated by central pattern generators is directed to the motoneurons via a premotor network that distributes the activity in a task-dependent manner determined by sensory and descending control information.

Optimal walking speed following changes in limb geometry.

The principle of dynamic similarity states that the optimal walking speeds of geometrically similar animals are independent of size when speed is normalized to the dimensionless Froude number (Fr). Furthermore, various studies have shown similar dimensionless optimal speed (Fr ∼0.25) for animals with quite different limb geometries. Here, we wondered whether the optimal walking speed of humans depends solely on total limb length or whether limb segment proportions play an essential role. If optimal walking speed solely depends on the limb length then, when subjects walk on stilts, they should consume less metabolic energy at a faster optimal speed than when they walk without stilts. To test this prediction, we compared kinematics, electromyographic activity and oxygen consumption in adults walking on a treadmill at different speeds with and without articulated stilts that artificially elongated the shank segment by 40 cm. Walking on stilts involved a non-linear reorganization of kinematic and electromyography patterns. In particular, we found a significant increase in the alternating activity of proximal flexors-extensors during the swing phase, despite significantly shorter normalized stride lengths. The minimal metabolic cost per unit distance walked with stilts occurred at roughly the same absolute speed, corresponding to a lower Fr number (Fr ∼0.17) than in normal walking (Fr ∼0.25). These findings are consistent with an important role of limb geometry optimization and kinematic coordination strategies in minimizing the energy expenditure of human walking.

Motor patterns during walking on a slippery walkway.

Friction and gravity represent two basic physical constraints of terrestrial locomotion that affect both motor patterns and the biomechanics of bipedal gait. To provide insights into the spatiotemporal organization of the motor output in connection with ground contact forces, we studied adaptation of human gait to steady low-friction conditions. Subjects walked along a slippery walkway (7 m long; friction coefficient approximately 0.06) or a normal, nonslippery floor at a natural speed. We recorded gait kinematics, ground reaction forces, and bilateral electromyographic (EMG) activity of 16 leg and trunk muscles and we mapped the recorded EMG patterns onto the spinal cord in approximate rostrocaudal locations of the motoneuron (MN) pools to characterize the spatiotemporal organization of the motor output. The results revealed several idiosyncratic features of walking on the slippery surface. The step length, cycle duration, and horizontal shear forces were significantly smaller, the head orientation tended to be stabilized in space, whereas arm movements, trunk rotations, and lateral trunk inclinations considerably increased and foot motion and gait kinematics resembled those of a nonplantigrade gait. Furthermore, walking on the slippery surface required stabilization of the hip and of the center-of-body mass in the frontal plane, which significantly improved with practice. Motor patterns were characterized by an enhanced (roughly twofold) level of MN activity, substantial decoupling of anatomical synergists, and the absence of systematic displacements of the center of MN activity in the lumbosacral enlargement. Overall, the results show that when subjects are confronted with unsteady surface conditions, like the slippery floor, they adopt a gait mode that tends to keep the COM centered over the supporting limbs and to increase limb stiffness. We suggest that this behavior may represent a distinct gait mode that is particularly suited to uncertain surface conditions in general.

Migration of motor pool activity in the spinal cord reflects body mechanics in human locomotion.

During the evolution of bipedal modes of locomotion, a sequential rostrocaudal activation of trunk muscles due to the undulatory body movements was replaced by more complex and discrete bursts of activity. Nevertheless, the capacity for segmental rhythmogenesis and the rostrocaudal propagation of spinal cord activity has been conserved. In humans, motoneurons of different muscles are arranged in columns, with a specific grouping of muscles at any given segmental level. The muscle patterns of locomotor activity and the biomechanics of the body center of mass have been studied extensively, but their interrelationship remains poorly understood. Here we mapped the electromyographic activity recorded from 30 bilateral leg muscles onto the spinal cord in approximate rostrocaudal locations of the motoneuron pools during walking and running in humans. We found that the rostrocaudal displacements of the center of bilateral motoneuron activity mirrored the changes in the energy due to the center-of-body mass motion. The results suggest that biomechanical mechanisms of locomotion, such as the inverted pendulum in walking and the pogo-stick bouncing in running, may be tightly correlated with specific modes of progression of motor pool activity rostrocaudally in the spinal cord.

Kinematic strategies in newly walking toddlers stepping over different support surfaces.

In adults, locomotor movements are accommodated to various support surface conditions by means of specific anticipatory locomotor adjustments and changes in the intersegmental coordination. Here we studied the kinematic strategies of toddlers at the onset of independent walking when negotiating various support surface conditions: stepping over an obstacle, walking on an inclined surface, and on a staircase. Generally, toddlers could perform these tasks only when supported by the arm. They exhibited strategies very different from those of the adults. Although adults maintained walking speed roughly constant, toddlers markedly accelerated when walking downhill or downstairs and decelerated when walking uphill or upstairs. Their coordination pattern of thigh-shank-foot elevation angles exhibited greater inter-trial variability than that in adults, but it did not undergo the systematic change as a function of task that was present in adults. Thus the intersegmental covariance plane rotated across tasks in adults, whereas its orientation remained roughly constant in toddlers. In contrast with the adults, the toddlers often tended to place the foot onto the obstacle or across the edges of the stairs. We interpret such foot placements as part of a haptic exploratory repertoire and we argue that the maintenance of a roughly constant planar covariance--irrespective of the surface inclination and height--may be functional to the exploratory behavior. The latter notion is consistent with the hypothesis proposed decades ago by Bernstein that, when humans start to learn a skill, they may restrict the number of degrees of freedom to reduce the size of the search space and simplify the coordination.

The many roles of vision during walking.

Vision can improve bipedal upright stability during standing and locomotion. However, during locomotion, vision supports additional behaviors such as gait cycle modulation, navigation, and obstacle avoidance. Here, we investigate how the multiple roles of vision are reflected in the dynamics of trunk control as the neural control problem changes from a fixed to a moving base of support. Subjects were presented with either low- or high-amplitude broadband visual stimuli during standing posture or while walking on a treadmill at 1 km/h and 5 km/h. Frequency response functions between visual scene motion (input) and trunk kinematics (output) revealed little or no change in the gain of trunk orientation in the standing posture and walking conditions. However, a dramatic increase in gain was observed in trunk (hip and shoulder) horizontal displacement from posture to locomotion. Such increases in gain may be interpreted as an increased coupling to visual scene motion. However, we believe that the increased gain reflects a decrease in stability due to a change of the control problem from standing to locomotion. Indeed, keeping the body upright with the use of vision during walking is complicated by the additional locomotor processes at work. Unlike during standing, vision plays many roles during locomotion, providing information for upright stability as well as body position relative to the external environment.

Modular control of limb movements during human locomotion.

The idea that the CNS may control complex interactions by modular decomposition has received considerable attention. We explored this idea for human locomotion by examining limb kinematics. The coordination of limb segments during human locomotion has been shown to follow a planar law for walking at different speeds, directions, and levels of body unloading. We compared the coordination for different gaits. Eight subjects were asked to walk and run on a treadmill at different speeds or to walk, run, and hop over ground at a preferred speed. To explore various constraints on limb movements, we also recorded stepping over an obstacle, walking with the knees flexed, and air-stepping with body weight support. We found little difference among covariance planes that depended on speed, but there were differences that depended on gait. In each case, we could fit the planar trajectories with a weighted sum of the limb length and orientation trajectories. This suggested that limb length and orientation might provide independent predictors of limb coordination. We tested this further by having the subjects step, run, and hop in place, thereby varying only limb length and maintaining limb orientation fixed, and also by marching with knees locked to maintain limb length constant while varying orientation. The results were consistent with a modular control of limb kinematics where limb movements result from a superposition of separate length- and orientation-related angular covariance. The hypothesis finds support in the animal findings that limb proprioception may also be encoded in terms of these global limb parameters.

Plasticity of spinal centers in spinal cord injury patients: new concepts for gait evaluation and training.

Recent data on spinal cord plasticity after spinal cord injury (SCI) were reviewed to analyze the influence of training on the neurophysiological organization of locomotor spinal circuits in SCI patients. In particular, the authors studied the relationship between central pattern generators (CPGs) and motor neuron pool activation during gait. An analysis of the relations between locomotor recovery and compensatory mechanisms focuses on the hierarchical organization of gait parameters and allows characterizing kinematic parameters that are highly stable during different gait conditions and in recovered gait after SCI. The importance of training characteristics and the use of robotic/automated devices in gait recovery is analyzed and discussed. The role of CPG in defining kinematic gait parameters is summarized, and spatio-temporal maps of EMG activity during gait are used to clarify the role of CPG plasticity in sustaining gait recovery.

Pendular energy transduction within the step during human walking on slopes at different speeds.

When ascending (descending) a slope, positive (negative) work must be performed to overcome changes in gravitational potential energy at the center of body mass (COM). This modifies the pendulum-like behavior of walking. The aim of this study is to analyze how energy exchange and mechanical work done vary within a step across slopes and speeds. Ten subjects walked on an instrumented treadmill at different slopes (from -9° to 9°), and speeds (between 0.56 and 2.22 m s-1). From the ground reaction forces, we evaluated energy of the COM, recovery (i.e. the potential-kinetic energy transduction) and pendular energy savings (i.e. the theoretical reduction in work due to this recovered energy) throughout the step. When walking uphill as compared to level, pendular energy savings increase during the first part of stance (when the COM is lifted) and decreases during the second part. Conversely in downhill walking, pendular energy savings decrease during the first part of stance and increase during the second part (when the COM is lowered). In uphill and downhill walking, the main phase of external work occurs around double support. Uphill, the positive work phase is extended during the beginning of single support to raise the body. Downhill, the negative work phase starts before double support, slowing the downward velocity of the body. Changes of the pendulum-like behavior as a function of slope can be illustrated by tilting the 'classical compass model' backwards (uphill) or forwards (downhill).