Viability leads to the emergence of gait transitions in learning agile quadrupedal locomotion on challenging terrains.

Quadruped animals are capable of seamless transitions between different gaits. While energy efficiency appears to be one of the reasons for changing gaits, other determinant factors likely play a role too, including terrain properties. In this article, we propose that viability, i.e., the avoidance of falls, represents an important criterion for gait transitions. We investigate the emergence of gait transitions through the interaction between supraspinal drive (brain), the central pattern generator in the spinal cord, the body, and exteroceptive sensing by leveraging deep reinforcement learning and robotics tools. Consistent with quadruped animal data, we show that the walk-trot gait transition for quadruped robots on flat terrain improves both viability and energy efficiency. Furthermore, we investigate the effects of discrete terrain (i.e., crossing successive gaps) on imposing gait transitions, and find the emergence of trot-pronk transitions to avoid non-viable states. Viability is the only improved factor after gait transitions on both flat and discrete gap terrains, suggesting that viability could be a primary and universal objective of gait transitions, while other criteria are secondary objectives and/or a consequence of viability. Moreover, our experiments demonstrate state-of-the-art quadruped robot agility in challenging scenarios.

Real-world humanoid locomotion with reinforcement learning.

Humanoid robots that can autonomously operate in diverse environments have the potential to help address labor shortages in factories, assist elderly at home, and colonize new planets. Although classical controllers for humanoid robots have shown impressive results in a number of settings, they are challenging to generalize and adapt to new environments. Here, we present a fully learning-based approach for real-world humanoid locomotion. Our controller is a causal transformer that takes the history of proprioceptive observations and actions as input and predicts the next action. We hypothesized that the observation-action history contains useful information about the world that a powerful transformer model can use to adapt its behavior in context, without updating its weights. We trained our model with large-scale model-free reinforcement learning on an ensemble of randomized environments in simulation and deployed it to the real-world zero-shot. Our controller could walk over various outdoor terrains, was robust to external disturbances, and could adapt in context.

Dimensionality of locomotor behaviors in developing C. elegans.

Adult animals display robust locomotion, yet the timeline and mechanisms of how juvenile animals acquire coordinated movements and how these movements evolve during development are not well understood. Recent advances in quantitative behavioral analyses have paved the way for investigating complex natural behaviors like locomotion. In this study, we tracked the swimming and crawling behaviors of the nematode Caenorhabditis elegans from postembryonic development through to adulthood. Our principal component analyses revealed that adult C. elegans swimming is low dimensional, suggesting that a small number of distinct postures, or eigenworms, account for most of the variance in the body shapes that constitute swimming behavior. Additionally, we found that crawling behavior in adult C. elegans is similarly low dimensional, corroborating previous studies. Further, our analysis revealed that swimming and crawling are distinguishable within the eigenworm space. Remarkably, young L1 larvae are capable of producing the postural shapes for swimming and crawling seen in adults, despite frequent instances of uncoordinated body movements. In contrast, late L1 larvae exhibit robust coordination of locomotion, while many neurons crucial for adult locomotion are still under development. In conclusion, this study establishes a comprehensive quantitative behavioral framework for understanding the neural basis of locomotor development, including distinct gaits such as swimming and crawling in C. elegans.

Simulation of Human Movement in Zero Gravity.

In the era of expanding manned space missions, understanding the biomechanical impacts of zero gravity on human movement is pivotal. This study introduces a novel and cost-effective framework that demonstrates the application of Microsoft's Azure Kinect body tracking technology as a motion input generator for subsequent OpenSim simulations in weightlessness. Testing rotations, locomotion, coordination, and martial arts movements, we validate the results' realism under the constraints of angular and linear momentum conservation. While complex, full-body coordination tasks face limitations in a zero gravity environment, our findings suggest possible approaches to device-free exercise routines for astronauts and reveal insights into the feasibility of hand-to-hand combat in space. However, some challenges remain in distinguishing zero gravity effects in the simulations from discrepancies in the captured motion input or forward dynamics calculations, making a comprehensive validation difficult. The paper concludes by highlighting the framework's practical potential for the future of space mission planning and related research endeavors, while also providing recommendations for further refinement.

Spatio-temporal distribution and genetic background of elastic proteins inside the chitin/chitosan matrix of insects including their functional significance for locomotion.

In insects, cuticle proteins interact with chitin and chitosan of the exoskeleton forming crystalline, amorphic or composite material structures. The biochemical and mechanical composition of the structure defines the cuticle's physical properties and thus how the insect cuticle behaves under mechanical stress. The tissue-specific ratio between chitin and chitosan and its pattern of deacetylation are recognized and interpreted by cuticle proteins depending on their local position in the body. Despite previous research, the assembly of the cuticle composites in time and space including its functional impact is widely unexplored. This review is devoted to the genetics underlying the temporal and spatial distribution of elastic proteins and the potential function of elastic proteins in insects with a focus on Resilin in the fruit fly Drosophila. The potential impact and function of localized patches of elastic proteins is discussed for movements in leg joints, locomotion and damage resistance of the cuticle. We conclude that an interdisciplinary research approach serves as an integral example for the molecular mechanisms of generation and interpretation of the chitin/chitosan matrix, not only in Drosophila but also in other arthropod species, and might help to synthesize artificial material composites.

The Exon Junction Complex Factor RBM8A in Glial Fibrillary Acid Protein-Expressing Astrocytes Modulates Locomotion Behaviors.

The role of RNA Binding Motif Protein 8a (RBM8A), an exon junction complex (EJC) component, in neurodevelopmental disorders has been increasingly studied for its crucial role in regulating multiple levels of gene expression. It regulates mRNA splicing, translation, and mRNA degradation and influences embryonic development. RBM8A protein is expressed in both neurons and astrocytes, but little is known about RBM8A's specific role in glial fibrillary acid protein (GFAP)-positive astrocytes. To address the role of RBM8A in astrocytes, we generated a conditional heterozygous knockout (KO) mouse line of Rbm8a in astrocytes using a GFAP-cre line. We confirmed a decreased expression of RBM8A in astrocytes of heterozygous conditional KO mice via RT-PCR and Sanger sequencing, as well as qRT-PCR, immunohistochemistry, and Western blot. Interestingly, these mice exhibit significantly increased movement and mobility, alongside sex-specific altered anxiety in the open field test (OFT) and elevated plus maze (OPM) tests. These tests, along with the rotarod test, suggest that these mice have normal motor coordination but hyperactive phenotypes. In addition, the haploinsufficiency of Rbm8a in astrocytes leads to a sex-specific change in astrocyte density in the dentate gyrus. This study further reveals the contribution of Rbm8a deletion to CNS pathology, generating more insights via the glial lens of an Rbm8a model of neurodevelopmental disorder.

S-shaped rolling gait designed using curve transformations of a snake robot for climbing on a bifurcated pipe.

In this work, we focus on overcoming the challenge of a snake robot climbing on the outside of a bifurcated pipe. Inspired by the climbing postures of biological snakes, we propose an S-shaped rolling gait designed using curve transformations. For this gait, the snake robot's body presenting an S-shaped curve is wrapped mainly around one side of the pipe, which leaves space for the fork of the pipe. To overcome the difficulty in constructing and clarifying the S-shaped curve, we present a method for establishing the transformation between a plane curve and a 3D curve on a cylindrical surface. Therefore, we can intuitively design the curve in 3D space, while analytically calculating the geometric properties of the curve in simple planar coordinate systems. The effectiveness of the proposed gait is verified by actual experiments. In successful configuration scenarios, the snake robot could stably climb on the pipe and efficiently cross or climb to the bifurcation while maintaining its target shape.

Locomotion Efficiency in Children With Cerebral Palsy Experiencing Limited Gross Motor Function: Walking Versus Cycling.

PURPOSE: This study compares cycling and walking efficiency, and energy expenditure in children with bilateral spastic cerebral palsy (CP). In children with CP, locomotion with body weight support aids such as a tricycle is a potential alternative for less exhausting movements. METHODS: Nine children with CP traveled at comfortable speed for 6 minutes by cycling and walking. The energy expenditure index (EEI) and the percentage of the reserve heart rate (%HRR) were calculated. RESULTS: The EEI was lower while cycling than walking, the traveled distance was higher while cycling than walking, and %HRR remained similar between cycling and walking. CONCLUSION: Cycling appears an efficient alternative to walking for children with CP for adapted school environments and in the community.

ANYmal parkour: Learning agile navigation for quadrupedal robots.

Performing agile navigation with four-legged robots is a challenging task because of the highly dynamic motions, contacts with various parts of the robot, and the limited field of view of the perception sensors. Here, we propose a fully learned approach to training such robots and conquer scenarios that are reminiscent of parkour challenges. The method involves training advanced locomotion skills for several types of obstacles, such as walking, jumping, climbing, and crouching, and then using a high-level policy to select and control those skills across the terrain. Thanks to our hierarchical formulation, the navigation policy is aware of the capabilities of each skill, and it will adapt its behavior depending on the scenario at hand. In addition, a perception module was trained to reconstruct obstacles from highly occluded and noisy sensory data and endows the pipeline with scene understanding. Compared with previous attempts, our method can plan a path for challenging scenarios without expert demonstration, offline computation, a priori knowledge of the environment, or taking contacts explicitly into account. Although these modules were trained from simulated data only, our real-world experiments demonstrate successful transfer on hardware, where the robot navigated and crossed consecutive challenging obstacles with speeds of up to 2 meters per second.

Structure formation induced by non-reciprocal cell-cell interactions in a multicellular system.

Collective cellular behavior plays a crucial role in various biological processes, ranging from developmental morphogenesis to pathological processes such as cancer metastasis. Our previous research has revealed that a mutant cell of Dictyostelium discoideum exhibits collective cell migration, including chain migration and traveling band formation, driven by a unique tail-following behavior at contact sites, which we term "contact following locomotion" (CFL). Here, we uncover an imbalance of forces between the front and rear cells within cell chains, leading to an additional propulsion force in the rear cells. Drawing inspiration from this observation, we introduce a theoretical model that incorporates non-reciprocal cell-cell interactions. Our findings highlight that the non-reciprocal interaction, in conjunction with self-alignment interactions, significantly contributes to the emergence of the observed collective cell migrations. Furthermore, we present a comprehensive phase diagram, showing distinct phases at both low and intermediate cell densities. This phase diagram elucidates a specific regime that corresponds to the experimental system.

Cerebellar state estimation enables resilient coupling across behavioural domains.

Cerebellar computations are necessary for fine behavioural control and may rely on internal models for estimation of behaviourally relevant states. Here, we propose that the central cerebellar function is to estimate how states interact with each other, and to use these estimates to coordinates extra-cerebellar neuronal dynamics underpinning a range of interconnected behaviours. To support this claim, we describe a cerebellar model for state estimation that includes state interactions, and link this model with the neuronal architecture and dynamics observed empirically. This is formalised using the free energy principle, which provides a dual perspective on a system in terms of both the dynamics of its physical-in this case neuronal-states, and the inferential process they entail. As a demonstration of this proposal, we simulate cerebellar-dependent synchronisation of whisking and respiration, which are known to be tightly coupled in rodents, as well as limb and tail coordination during locomotion. In summary, we propose that the ubiquitous involvement of the cerebellum in behaviour arises from its central role in precisely coupling behavioural domains.

Tapping into the human spinal locomotor centres with transspinal stimulation.

Human locomotion is controlled by spinal neuronal networks of similar properties, function, and organization to those described in animals. Transspinal stimulation affects the spinal locomotor networks and is used to improve standing and walking ability in paralyzed people. However, the function of locomotor centers during transspinal stimulation at different frequencies and intensities is not known. Here, we document the 3D joint kinematics and spatiotemporal gait characteristics during transspinal stimulation at 15, 30, and 50 Hz at sub-threshold and supra-threshold stimulation intensities. We document the temporal structure of gait patterns, dynamic stability of joint movements over stride-to-stride fluctuations, and limb coordination during walking at a self-selected speed in healthy subjects. We found that transspinal stimulation (1) affects the kinematics of the hip, knee, and ankle joints, (2) promotes a more stable coordination at the left ankle, (3) affects interlimb coordination of the thighs, and (4) intralimb coordination between thigh and foot, (5) promotes greater dynamic stability of the hips, (6) increases the persistence of fluctuations in step length variability, and lastly (7) affects mechanical walking stability. These results support that transspinal stimulation is an important neuromodulatory strategy that directly affects gait symmetry and dynamic stability. The conservation of main effects at different frequencies and intensities calls for systematic investigation of stimulation protocols for clinical applications.

Walking modulates visual detection performance according to stride cycle phase.

Walking is among our most frequent and natural of voluntary behaviours, yet the consequences of locomotion upon perceptual and cognitive function remain largely unknown. Recent work has highlighted that although walking feels smooth and continuous, critical phases exist within each step for the successful coordination of perceptual and motor function. Here, we test whether these phasic demands impact upon visual perception, by assessing performance in a visual detection task during natural unencumbered walking. We finely sample visual performance over the stride cycle as participants walk along a smooth linear path at a comfortable speed in a wireless virtual reality environment. At the group-level, accuracy, reaction times, and response likelihood show strong oscillations, modulating at approximately 2 cycles per stride (~2 Hz) with a marked phase of optimal performance aligned with the swing phase of each step. At the participant level, Bayesian inference of population prevalence reveals highly prevalent oscillations in visual detection performance that cluster in two idiosyncratic frequency ranges (2 or 4 cycles per stride), with a strong phase alignment across participants.

Multimodal sensory control of motor performance by glycinergic interneurons of the mouse spinal cord deep dorsal horn.

Sensory feedback is integral for contextually appropriate motor output, yet the neural circuits responsible remain elusive. Here, we pinpoint the medial deep dorsal horn of the mouse spinal cord as a convergence point for proprioceptive and cutaneous input. Within this region, we identify a population of tonically active glycinergic inhibitory neurons expressing parvalbumin. Using anatomy and electrophysiology, we demonstrate that deep dorsal horn parvalbumin-expressing interneuron (dPV) activity is shaped by convergent proprioceptive, cutaneous, and descending input. Selectively targeting spinal dPVs, we reveal their widespread ipsilateral inhibition onto pre-motor and motor networks and demonstrate their role in gating sensory-evoked muscle activity using electromyography (EMG) recordings. dPV ablation altered limb kinematics and step-cycle timing during treadmill locomotion and reduced the transitions between sub-movements during spontaneous behavior. These findings reveal a circuit basis by which sensory convergence onto dorsal horn inhibitory neurons modulates motor output to facilitate smooth movement and context-appropriate transitions.

Spatiotemporal modulation of a common set of muscle synergies during unpredictable and predictable gait perturbations in older adults.

Muscle synergies as functional low-dimensional building blocks of the neuromotor system regulate the activation patterns of muscle groups in a modular structure during locomotion. The purpose of the current study was to explore how older adults organize locomotor muscle synergies to counteract unpredictable and predictable gait perturbations during the perturbed steps and the recovery steps. Sixty-three healthy older adults (71.2±5.2 years) participated in the study. Mediolateral and anteroposterior unpredictable and predictable perturbations during walking were introduced using a treadmill. Muscle synergies were extracted from the electromyographic activity of 13 lower limb muscles using Gaussian non-negative matrix factorization. The four basic synergies responsible for unperturbed walking (weight acceptance, propulsion, early swing and late swing) were preserved in all applied gait perturbations, yet their temporal recruitment and muscle contribution in each synergy were modified (P<0.05). These modifications were observed for up to four recovery steps and were more pronounced (P<0.05) following unpredictable perturbations. The recruitment of the four basic walking synergies in the perturbed and recovery gait cycles indicates a robust neuromotor control of locomotion by using activation patterns of a few and well-known muscle synergies with specific adjustments within the synergies. The selection of pre-existing muscle synergies while adjusting the time of their recruitment during challenging locomotor conditions may improve the effectiveness to deal with perturbations and promote the transfer of adaptation between different kinds of perturbations.

Comparative kinetics of humans and non-human primates during vertical climbing.

Climbing represents a critical behavior in the context of primate evolution. However, anatomically modern human populations are considered ill-suited for climbing. This adaptation can be attributed to the evolution of striding bipedalism, redirecting anatomical traits away from efficient climbing. Although prior studies have speculated on the kinetic consequences of this anatomical reorganization, there is a lack of data on the force profiles of human climbers. This study utilized high-speed videography and force plate analysis to assess single limb forces during climbing from 44 human participants of varying climbing experience and compared these data with climbing data from eight species of non-human primates (anthropoids and strepsirrhines). Contrary to expectations, experience level had no significant effect on the magnitude of single limb forces in humans. Experienced climbers did, however, demonstrate a predictable relationship between center of mass position and peak normal forces, suggesting a better ability to modulate forces during climbing. Humans exhibited significantly higher peak propulsive forces in the hindlimb compared with the forelimb and greater hindlimb dominance overall compared with non-human primates. All species sampled demonstrated exclusively tensile forelimbs and predominantly compressive hindlimbs. Strepsirrhines exhibited a pull-push transition in normal forces, while anthropoid primates, including humans, did not. Climbing force profiles are remarkably stereotyped across humans, reflecting the universal mechanical demands of this form of locomotion. Extreme functional differentiation between forelimbs and hindlimbs in humans may help to explain the evolution of bipedalism in ancestrally climbing hominoids.

Multiple forces facilitate the aquatic acrobatics of grasshopper and bioinspired robot.

Aquatic locomotion is challenging for land-dwelling creatures because of the high degree of fluidity with which the water yields to loads. We surprisingly found that the Chinese rice grasshopper Oxya chinensis, known for its terrestrial acrobatics, could swiftly launch itself off the water's surface in around 25 ms and seamlessly transition into flight. Biological observations showed that jumping grasshoppers use their front and middle legs to tilt up bodies first and then lift off by propelling the water toward the lower back with hind legs at angular speeds of up to 18°/ms, whereas the swimming grasshoppers swing their front and middle legs in nearly horizontal planes and move hind legs less violently (~8°/ms). Force measurement and model analysis indicated that the weight support could be achieved by hydrostatics which are proportionate to the mass of the grasshoppers, while the propulsions for motion are derived from the controlled limb-water interactions (i.e., the hydrodynamics). After learning the structural and behavioral strategies of the grasshoppers, a robot was created and was capable of swimming and jumping on the water surface like the insects, further demonstrating the effectiveness of decoupling the challenges of aquatic locomotion by the combined use of the static and dynamic hydro forces. This work not only uncovered the combined mechanisms responsible for facilitating aquatic acrobatics in this species but also laid a foundation for developing bioinspired robots that can locomote across multiple media.

Robot motor learning shows emergence of frequency-modulated, robust swimming with an invariant Strouhal number.

Fish locomotion emerges from diverse interactions among deformable structures, surrounding fluids and neuromuscular activations, i.e. fluid-structure interactions (FSI) controlled by fish's motor systems. Previous studies suggested that such motor-controlled FSI may possess embodied traits. However, their implications in motor learning, neuromuscular control, gait generation, and swimming performance remain to be uncovered. Using robot models, we studied the embodied traits in fish-inspired swimming. We developed modular robots with various designs and used central pattern generators (CPGs) to control the torque acting on robot body. We used reinforcement learning to learn CPG parameters for maximizing the swimming speed. The results showed that motor frequency converged faster than other parameters, and the emergent swimming gaits were robust against disruptions applied to motor control. For all robots and frequencies tested, swimming speed was proportional to the mean undulation velocity of body and caudal-fin combined, yielding an invariant, undulation-based Strouhal number. The Strouhal number also revealed two fundamental classes of undulatory swimming in both biological and robotic fishes. The robot actuators were also demonstrated to function as motors, virtual springs and virtual masses. These results provide novel insights in understanding fish-inspired locomotion.

Functional plasticity of glutamatergic neurons of medullary reticular nuclei after spinal cord injury in mice.

Spinal cord injury disrupts the descending command from the brain and causes a range of motor deficits. Here, we use optogenetic tools to investigate the functional plasticity of the glutamatergic reticulospinal drive of the medullary reticular formation after a lateral thoracic hemisection in female mice. Sites evoking stronger excitatory descending drive in intact conditions are the most impaired after injury, whereas those associated with a weaker drive are potentiated. After lesion, pro- and anti-locomotor activities (that is, initiation/acceleration versus stop/deceleration) are overall preserved. Activating the descending reticulospinal drive improves stepping ability on a flat surface of chronically impaired injured mice, and its priming enhances recovery of skilled locomotion on a horizontal ladder. This study highlights the resilience and capacity for reorganization of the glutamatergic reticulospinal command after injury, along with its suitability as a therapeutical target to promote functional recovery.

Repeated microdoses of LSD do not alter anxiety or boldness in zebrafish.

The therapeutic use of lysergic acid diethylamide (LSD) has resurfaced in the last decade, prompting further scientific investigation into its effectiveness in many animal models. Zebrafish (Danio rerio) are a popular model organism in medical sciences and are used to examine the repeated administration of pharmacological compounds. Previous zebrafish research found acute LSD altered behaviour and cortisol levels at high (250 µg/L) but not low (5-100 µg/L) levels. In this study, we used a motion tracking system to record and analyze the movement patterns of zebrafish after acute and repeated 10-day LSD exposure (1.5 µg/L, 15 µg/L, 150 µg/L) and after seven days of withdrawal. The open-field and novel object approach tests were used to examine anxiety-like behaviour, boldness, and locomotion. In the acute experiments we observed a significant decrease in high mobility with 1.5 µg/L, 15 µg/L, and 150 µg/L of LSD compared to the control and a decrease in velocity with 1.5 and 15 µg/L. In repeated experiments, there were no significant differences in the levels of anxiety, boldness, or locomotion between all LSD groups and controls immediately after 10-day treatment or after withdrawal.