AbstractCentral pattern generators (CPGs) are neural networks that can produce rhythmic patterned outputs without rhythmic sensory or central input. CPGs underlie the production of most rhythmic motor patterns and have been extensively studied as models of neural network function.

Quadruped animal locomotion emerges from the interactions between the spinal central pattern generator (CPG), sensory feedback, and supraspinal drive signals from the brain. Computational models of CPGs have been widely used for investigating the spinal cord contribution to animal locomotion control in computational neuroscience and in bio-inspired robotics. However, the contribution of supraspinal drive to anticipatory behavior, i.e. motor behavior that involves planning ahead of time (e.g. of footstep placements), is not yet properly understood. In particular, it is not clear whether the brain modulates CPG activity and/or directly modulates muscle activity (hence bypassing the CPG) for accurate foot placements. In this paper, we investigate the interaction of supraspinal drive and a CPG in an anticipatory locomotion scenario that involves stepping over gaps. By employing deep reinforcement learning (DRL), we train a neural network policy that replicates the supraspinal drive behavior. This policy can either modulate the CPG dynamics, or directly change actuation signals to bypass the CPG dynamics. Our results indicate that the direct supraspinal contribution to the actuation signal is a key component for a high gap crossing success rate. However, the CPG dynamics in the spinal cord are beneficial for gait smoothness and energy efficiency. Moreover, our investigation shows that sensing the front feet distances to the gap is the most important and sufficient sensory information for learning gap crossing. Our results support the biological hypothesis that cats and horses mainly control the front legs for obstacle avoidance, and that hind limbs follow an internal memory based on the front limbs' information. Our method enables the quadruped robot to cross gaps of up to 20 cm (50% of body-length) without any explicit dynamics modeling or Model Predictive Control (MPC).

Locomotion requires the continuous integration of descending motor commands and sensory inputs from the legs by spinal central pattern generator circuits. Modulation of spinal neural circuits by transspinal stimulation is well documented, but how transspinal stimulation affects corticospinal excitability during walking in humans remains elusive. We measured the motor evoked potentials (MEPs) at multiple phases of the step cycle conditioned with transspinal stimulation delivered at sub- and supra-threshold intensities of the spinally-mediated transspinal evoked potential (TEP). Transspinal stimulation was delivered before or after TMS during which summation between MEP and TEP responses in the surface EMG was absent or present. Relationships between MEP amplitude and background EMG activity, silent period duration, and phase-dependent EMG amplitude modulation during and after stimulation were also determined. Ankle flexor and extensor MEPs were depressed by suprathreshold transspinal stimulation when descending volleys were timed to interact with transspinal stimulation-induced motoneuron depolarization at the spinal cord. MEP depression coincided with decreased MEP gain, unaltered MEP threshold, and unaltered silent period duration. Locomotor EMG activity of bilateral knee and ankle muscles was significantly depressed during the step at which transspinal stimulation was delivered, but fully recovered at the subsequent step. The results support a model in which MEP depression by transspinal stimulation occurs via subcortical or spinal mechanisms. Transspinal stimulation disrupts the locomotor output of flexor and extensor motoneurons initially, but the intact nervous system has the ability to rapidly overcome this pronounced locomotor adaptation. In conclusion, transspinal stimulation directly affects spinal locomotor centers in healthy humans.

The ability to efficiently move in complex environments is a fundamental property both for animals and for robots, and the problem of locomotion and movement control is an area in which neuroscience and robotics can fruitfully interact. Animal locomotion control is in a large part based on spinal cord circuits that combine reflex loops and central pattern generators (CPGs), i.e. neural networks capable of producing complex rhythmic or discrete patterns while being activated and modulated by relatively simple control signals. These networks located in the spinal cord for vertebrate animals are modulated by descending control signals and interact with the musculoskeletal system for generating rich motor behaviors. This paper presents how numerical models and robots can be used to explore the interplay of these four components (CPGs, reflexes, descending modulation, and musculoskeletal system). Going from lamprey to human locomotion, a series of models are presented that tend to show that the respective roles of these components have changed during evolution with a dominant role of CPGs in lamprey and salamander locomotion, and a more important role for sensory feedback and descending modulation in human locomotion. Interesting properties for robot locomotion control are also discussed.

Typical legged locomotion controllers are designed or trained offline. This is in contrast to many animals, which are able to locomote at birth, and rapidly improve their locomotion skills with few real-world interactions. Such motor control is possible through oscillatory neural networks located in the spinal cord of vertebrates, known as Central Pattern Generators (CPGs). Models of the CPG have been widely used to generate locomotion skills in robotics, but can require extensive hand-tuning or offline optimization of inter-connected parameters with genetic algorithms. In this paper, we present a framework for the online optimization of the CPG parameters through Bayesian Optimization. We show that our framework can rapidly optimize and adapt to varying velocity commands and changes in the terrain, for example to varying coefficients of friction, terrain slope angles, and added mass payloads placed on the robot. We study the effects of sensory feedback on the CPG, and find that both force feedback in the phase equations, as well as posture control (Virtual Model Control) are both beneficial for robot stability and energy efficiency. In hardware experiments on the Unitree Go1, we show rapid optimization (in under 3 minutes) and adaptation of energy-efficient gaits to varying target velocities in a variety of scenarios: varying coefficients of friction, added payloads up to 15 kg, and variable slopes up to 10 degrees.

The complexity of motor control is, to a great extent, overcome by hierarchical organization of the controlling system. Lower levels of this system contain a set of central pattern generators (CPGs). the neuronal networks capable of producing the basic spatio-temporal pattern underlying different “automatic” movements (rhythmic movements like locomotion, respiration, as well as non-rhythmic ones like swallowing and defense reactions) in the absence of peripheral sensory feedback. Instead of controlling individual muscles involved in generation of a definite motor pattern, higher centers (through command system) activate the corresponding CPG that generates this pattern. The most detailed analysis of CPGs has been performed for rhythmical movements. In these experiments, sensory feedback was abolished using in vitro (see Figures 1D; 2D; 3D), immobilized (see Figure 4B-D), or deafferented preparations.

Central pattern generators (CPGs) are self-organizing biological neural circuits that produce rhythmic outputs in the absence of rhythmic input. They Central pattern generators (CPGs) are self-organizing biological neural circuits that produce rhythmic outputs in the absence of rhythmic input. They are the source of the tightly-coupled patterns of neural activity that drive rhythmic and stereotyped motor behaviors like walking, swimming, breathing, or chewing. The ability to function without input from higher brain areas still requires modulato Fig. 1 provides a simplified schematic that summarizes these proposed mechanisms. A command specifying desired body velocity descends from higher centers to the MLR, which drives the spinal locomotor CPG. The CPG timer produces the appropriate cadence and phase durations and the pattern formation layer modulates the motoneuronal outputs. The activated muscles resist stretch through their own intrinsic biomechanical properties, providing a rapid form of length and velocity feedback control. Reflexes mediated by Golgi tendon organ and other afferents provide additional load compensation, but the main role of sensory input may be to adjust or override the CPG at stance-swing-stance transitions. As described in Neuromodulation, the human locomotive CPG is very adaptable and can respond to sensory input. It receives input from the brainstem as well as from the environment to keep the network regulated. Newer studies have not only confirmed the presence of the CPG for human locomotion, but also confirmed its robustness and adaptability. For example, Choi and Bastian showed that the networks responsible for human walking are adaptable on short and long timescales. They showed adaptation to different gait patterns and different walking contexts. Also, they showed that different motor patterns can adapt independently. Adults could even walk on treadmills going in a different direction for each leg. This study showed that independent networks control forward and backward walking and that networks controlling each leg can adapt independently and be trained to walk independently. Thus, humans also possess a central pattern generator for locomotion that is capable not only of rhythmic pattern…

consists of : e Segmental central pattern generators (CPGs) which control the motor neurons and ultimately … while others are under the control of central pattern generators (@PGs). 2.1 Coupling Between Afferents … consider the dynamic filtering applied by central pattern generators (Fig. 4, right). We have shown that during

1 nrz 1 Pattern generators Central pattern generators Premotor networks Pattern generators Pattern … inputs — ini- Invertebrate Central Pattern Generators 269 tially the motor pattern ceases but returns … pyloric central pattern generator, 256, 263 Homeostatic plasticity, central pattern generators, 268-269