How the brain can be trained to achieve an intermittent control strategy for stabilizing quiet stance by means of reinforcement learning.

The stabilization of human quiet stance is achieved by a combination of the intrinsic elastic properties of ankle muscles and an active closed-loop activation of the ankle muscles, driven by the delayed feedback of the ongoing sway angle and the corresponding angular velocity in a way of a delayed proportional (P) and derivative (D) feedback controller. It has been shown that the active component of the stabilization process is likely to operate in an intermittent manner rather than as a continuous controller: the switching policy is defined in the phase-plane, which is divided in dangerous and safe regions, separated by appropriate switching boundaries. When the state enters a dangerous region, the delayed PD control is activated, and it is switched off when it enters a safe region, leaving the system to evolve freely. In comparison with continuous feedback control, the intermittent mechanism is more robust and capable to better reproduce postural sway patterns in healthy people. However, the superior performance of the intermittent control paradigm as well as its biological plausibility, suggested by experimental evidence of the intermittent activation of the ankle muscles, leaves open the quest of a feasible learning process, by which the brain can identify the appropriate state-dependent switching policy and tune accordingly the P and D parameters. In this work, it is shown how such a goal can be achieved with a reinforcement motor learning paradigm, building upon the evidence that, in general, the basal ganglia are known to play a central role in reinforcement learning for action selection and, in particular, were found to be specifically involved in postural stabilization.

Artificial cognition vs. artificial intelligence for next-generation autonomous robotic agents.

The trend in industrial/service robotics is to develop robots that can cooperate with people, interacting with them in an autonomous, safe and purposive way. These are the fundamental elements characterizing the fourth and the fifth industrial revolutions (4IR, 5IR): the crucial innovation is the adoption of intelligent technologies that can allow the development of cyber-physical systems, similar if not superior to humans. The common wisdom is that intelligence might be provided by AI (Artificial Intelligence), a claim that is supported more by media coverage and commercial interests than by solid scientific evidence. AI is currently conceived in a quite broad sense, encompassing LLMs and a lot of other things, without any unifying principle, but self-motivating for the success in various areas. The current view of AI robotics mostly follows a purely disembodied approach that is consistent with the old-fashioned, Cartesian mind-body dualism, reflected in the software-hardware distinction inherent to the von Neumann computing architecture. The working hypothesis of this position paper is that the road to the next generation of autonomous robotic agents with cognitive capabilities requires a fully brain-inspired, embodied cognitive approach that avoids the trap of mind-body dualism and aims at the full integration of Bodyware and Cogniware. We name this approach Artificial Cognition (ACo) and ground it in Cognitive Neuroscience. It is specifically focused on proactive knowledge acquisition based on bidirectional human-robot interaction: the practical advantage is to enhance generalization and explainability. Moreover, we believe that a brain-inspired network of interactions is necessary for allowing humans to cooperate with artificial cognitive agents, building a growing level of personal trust and reciprocal accountability: this is clearly missing, although actively sought, in current AI. The ACo approach is a work in progress that can take advantage of a number of research threads, some of them antecedent the early attempts to define AI concepts and methods. In the rest of the paper we will consider some of the building blocks that need to be re-visited in a unitary framework: the principles of developmental robotics, the methods of action representation with prospection capabilities, and the crucial role of social interaction.

Discrete cortical control during quiet stance revealed by desynchronization and rebound of beta oscillations.

Postural sway during quiet stance often exhibits a repetition of micro-fall and the subsequent micro-recovery. The classical view -that the quiet bipedal stance is stabilized by the ankle joint stiffness- has been challenged by paradoxical non-spring-like behaviors of calf muscles: gastrocnemius muscles are shortened and then lengthened, respectively, during the micro-fall and the micro-recovery. Here, we examined EEG based brain activity during quiet stance, and identified desynchronization and synchronization of beta oscillations that were associated, respectively, with the micro-fall and the micro-recovery. Based on a widely accepted scenario for beta-band desynchronization during movement and post-movement rebound in the control of discrete voluntary movement, our results reveal that the beta rebound can be considered as a manifestation of stop command to punctuate the motor control for every fall-recovery cycle. Namely, cortical interventions to the automatic postural control are discrete, rather than continuous modulations. The finding is highly compatible with the intermittent control model, rather than the stiffness control model.

µ-band desynchronization in the contralateral central and central-parietal areas predicts proprioceptive acuity.

Introduction: Position sense, which belongs to the sensory stream called proprioception, is pivotal for proper movement execution. Its comprehensive understanding is needed to fill existing knowledge gaps in human physiology, motor control, neurorehabilitation, and prosthetics. Although numerous studies have focused on different aspects of proprioception in humans, what has not been fully investigated so far are the neural correlates of proprioceptive acuity at the joints. Methods: Here, we implemented a robot-based position sense test to elucidate the correlation between patterns of neural activity and the degree of accuracy and precision exhibited by the subjects. Eighteen healthy participants performed the test, and their electroencephalographic (EEG) activity was analyzed in its µ band (8-12 Hz), as the frequency band related to voluntary movement and somatosensory stimulation. Results: We observed a significant positive correlation between the matching error, representing proprioceptive acuity, and the strength of the activation in contralateral hand motor and sensorimotor areas (left central and central-parietal areas). In absence of visual feedback, these same regions of interest (ROIs) presented a higher activation level compared to the association and visual areas. Remarkably, central and central-parietal activation was still observed when visual feedback was added, although a consistent activation in association and visual areas came up. Conclusion: Summing up, this study supports the existence of a specific link between the magnitude of activation of motor and sensorimotor areas related to upper limb proprioceptive processing and the proprioceptive acuity at the joints.

The Quest for Cognition in Purposive Action: From Cybernetics to Quantum Computing.

Norbert Wiener and Nikolai Bernstein set the stage for a worldwide multidisciplinary attempt to understand how purposive action is integrated with cognition in a circular, bidirectional manner, both in life sciences and engineering. Such a 'workshop' is still open and far away from a satisfactory level of understanding, despite the current hype surrounding Artificial Intelligence (AI). The problem is that Cognition is frequently confused with Intelligence, overlooking a crucial distinction: the type of cognition that is required of a cognitive agent to meet the challenge of adaptive behavior in a changing environment is Embodied Cognition, which is antithetical to the disembodied and dualistic nature of the current wave of AI. This essay is the perspective formulation of a cybernetic framework for the representation of actions that, following Bernstein, is focused on what has long been considered the fundamental issue underlying action and motor control, namely the degrees of freedom problem. In particular, the paper reviews a solution to this problem based on a model of ideomotor/muscle-less synergy formation, namely the Passive Motion Paradigm (PMP). Moreover, it is shown how this modeling approach can be reformulated in a distributed manner based on a self-organizing neural paradigm consisting of multiple topology-representing networks with attractor dynamics. The computational implication of such an approach is also briefly analyzed looking at possible alternatives of the von Neuman paradigm, namely neuromorphic and quantum computing, aiming in perspective at a hybrid computational framework for integrating digital information, analog information, and quantum information. It is also suggested that such a framework is crucial not only for the neurobiological modeling of motor cognition but also for the design of the cognitive architecture of autonomous robots of industry 4.0 that are supposed to interact and communicate naturally with human partners.

Integrating ankle and hip strategies for the stabilization of upright standing: An intermittent control model.

Even in unperturbed upright standing of healthy young adults, body sway involves concurrent oscillations of ankle and hip joints, thus suggesting to using biomechanical models with at least two degrees of freedom, namely, a double inverted pendulum (DIP) framework. However, in a previous study, it was demonstrated that the observed coordinated ankle-hip patterns do not necessarily require the independent active control of the two joints but can be explained by a simpler hybrid control system, with a single active component (intermittent, delayed sensory feedback of the ongoing sway) applied to the ankle joint and a passive component (stiffness control) applied to the hip joint. In particular, the proposed active component was based on the internal representation of a virtual inverted pendulum (VIP) that links the ankle to the current position of the global center of mass (CoM). This hybrid control system, which can also be described as an ankle strategy, is consistent with the known kinematics of the DIP and, in particular, with the anti-phase correlation of the acceleration profiles of the two joints. The purpose of this study is to extend the hybrid control model in order to apply to both the ankle and hip strategy, clarifying as well the rationale of mixed strategies. The extension consists of applying the hybrid control scheme to both joints: a passive stiffness component and an active intermittent component, based on the same feedback signals derived from the common VIP but with independent parameter gains for the two joints. Thus, the hip gains are null in the pure ankle strategy, the ankle gains are null in the pure hip strategy, and both ankle and hip gains are specifically tuned in mixed strategies. The simulation of such an extended model shows that it can reproduce both strategies; moreover, the pure ankle strategy is more robust than the hip strategy, because the range of variation (RoV) of the intermittent control gains is larger in the former case than in the latter, and the pure ankle strategy is also more energy efficient. Generally, the simulations suggest that there is no advantage to employ mixed strategies, except in borderline situations in which the control gains are just outside the RoV that provides stable control for either pure strategy: in this case, a stable mixed strategy can emerge from the combination of two marginally unstable pure strategies.

A Dynamic Submaximal Fatigue Protocol Alters Wrist Biomechanical Properties and Proprioception.

Fatigue is a temporary condition that arises as a result of intense and/or prolonged use of muscles and can affect skilled human performance. Therefore, the quantitative analysis of these effects is a topic of crucial interest in both ergonomics and clinical settings. This study introduced a novel protocol, based on robotic techniques, to quantitatively assess the effects of fatigue on the human wrist joint. A wrist manipulandum was used for two concurrent purposes: (1) implementing the fatigue task and (2) assessing the functional changes both before and at four time points after the end of the fatigue task. Fourteen participants completed the experimental protocol, which included the fatigue task and assessment sessions over 2 days. Specifically, the assessments performed are related to the following indicators: (1) isometric forces, (2) biomechanical properties of the wrist, (3) position sense, and (4) stretch reflexes of the muscles involved. The proposed fatigue task was a short-term, submaximal and dynamic wrist flexion/extension task designed with a torque opposing wrist flexion. A novel task termination criterion was employed and based on a percentage decrease in the mean frequency of muscles measured using surface electromyography. The muscle fatigue analysis demonstrated a change in mean frequency for both the wrist flexors and extensors, however, only the isometric flexion force decreased 4 min after the end of the task. At the same time point, wrist position sense was significantly improved and stiffness was the lowest. Viscosity presented different behaviors depending on the direction evaluated. At the end of the experiment (about 12 min after the end of the fatigue task), wrist position sense recovered to pre-fatigue values, while biomechanical properties did not return to their pre-fatigue values. Due to the wide variety of fatigue tasks proposed in the literature, it has been difficult to define a complete framework that presents the dynamic of fatigue-related changes in different components associated with wrist function. This work enables us to discuss the possible causes and the mutual relationship of the changes detected after the same task.

Implementing a robust wrist dynamic fatigue task: repeatability and investigation of the features involved.

In this study, we implemented a protocol for the robotic assessment of the effects of forearm muscle fatigue on wrist dynamics. The potential of robotic devices lies in the possibility to control and measure a wide variety of kinematic and physiological variables, both in repeated sessions over time and during real-time assessments. The implemented fatigue task is tailored to the robotically assessed single-subject maximal force and based on a real-time evaluation of muscle activity. The protocol resulted to be repeatable across sessions evaluated on the same subject and a preliminary step toward a better understanding of which features should be monitored to design a robust and strongly controlled dynamic fatiguing task.

Evaluating Viscoelastic Properties of the Wrist Joint During External Perturbations: Influence of Velocity, Grip, and Handedness.

In this study, we designed a robot-based method to compute a mechanical impedance model that could extract the viscoelastic properties of the wrist joint. Thirteen subjects participated in the experiment, testing both dominant and nondominant hands. Specifically, the robotic device delivered position-controlled disturbances in the flexion-extension degree of freedom of the wrist. The external perturbations were characterized by small amplitudes and fast velocities, causing rotation at the wrist joint. The viscoelastic characteristics of the mechanical impedance of the joint were evaluated from the wrist kinematics and corresponding torques. Since the protocol used position inputs to determine changes in mean wrist torque, a detailed analysis of wrist joint dynamics could be made. The scientific question was whether and how these mechanical features changed with various grip demands and perturbation velocities. Nine experimental conditions were tested for each hand, given by the combination of three velocity perturbations (fast, medium, and slow) and three hand grip conditions [self-selected grip, medium and high grip force, as percentage of the maximum voluntary contraction (MVC)]. Throughout the experiments, electromyographic signals of the extensor carpi radialis (ECR) and the flexor carpi radialis (FCR) were recorded. The novelty of this work included a custom-made soft grip sensor, wrapped around the robotic handle, to accurately quantify the grip force exerted by the subjects during experimentation. Damping parameters were in the range of measurements from prior studies and consistent among the different experimental conditions. Stiffness was independent of both direction and velocity of perturbations and increased with increasing grip demand. Both damping and stiffness were not different between the dominant and nondominant hands. These results are crucial to improving our knowledge of the mechanical characteristics of the wrist, and how grip demands influence these properties. This study is the foundation for future work on how mechanical characteristics of the wrist are affected in pathological conditions.

Wrist Position Sense in Two Dimensions: Between-Hand Symmetry and Anisotropic Accuracy Across the Space.

A deep investigation of proprioceptive processes is necessary to understand the relationship between sensory afferent inputs and motor outcomes. In this work, we investigate whether and how perception of wrist position is influenced by the direction along which the movement occurs. Most previous studies have tested Joint Position Sense (JPS) through 1 degree of freedom (DoF) wrist movements, such as flexion/extension (FE) or radial/ulnar deviation (RUD). However, the wrist joint has 3-DoF and many activities of daily living produce combined movements, requiring at least 2-DoF wrist coordination. For this reason, in this study, target positions involved movement directions that combined wrist flexion or extension with radial or ulnar deviation. The chosen task was a robot-aided Joint Position Matching (JPM), in which blindfolded participants actively reproduced a previously passively assumed target joint configuration. The JPM performance of 20 healthy participants was quantified through measures of accuracy and precision, in terms of both perceived target direction and distance along each direction of movement. Twelve different directions of movement were selected and both hands tested. The left and right hand led to comparable results, both target extents and directions were differently perceived according to the target direction on the FE/RUD space. Moreover, during 2-DoF combined movements, subjects' perception of directions was impaired when compared to 1-DoF target movements. In summary, our results showed that human perception of wrist position on the FE/RUD space is symmetric between hands but not isotropic among movement directions.

Spontaneous movements in the newborns: a tool of quantitative video analysis of preterm babies.

BACKGROUND AND OBJECTIVES: The number of preterm babies is steadily growing world-wide and these neonates are at risk of neuro-motor-cognitive deficits. The observation of spontaneous movements in the first three months of age is known to predict such risk. However, the analysis by specifically trained physiotherapists is not suited for the clinical routine, motivating the development of simple computerized video analysis systems, integrated with a well-structured Biobank to make available for preterm babies a growing service with diagnostic, prognostic and epidemiological purposes. METHODS: MIMAS (Markerless Infant Movement Analysis System) is a simple, low-cost system of video analysis of spontaneous movements of newborns in their natural environment, based on a single standard RGB camera, without markers attached to the body. The original videos are transformed into binarized sequences highlighting the silhouette of the baby, in order to minimize the illumination effects and increase the robustness of the analysis; such sequences are then coded by a large set of parameters (39) related to the spatial and spectral changes of the silhouette. The parameter vectors of each baby were stored in the Biobank together with related clinical information. RESULTS: The preliminary test of the system was carried out at the Gaslini Pediatric Hospital in Genoa, where 46 preterm (PT) and 21 full-term (FT) babies (as controls) were recorded at birth (T0) and 8-12 weeks thereafter (T1). A simple statistical analysis of the data showed that the coded parameters are sensitive to the degree of maturation of the newborns (comparing T0 with T1, for both PT and FT babies), and to the conditions at birth (PT vs. FT at T0), whereas this difference tends to vanish at T1. Moreover, the coding method seems also able to detect the few 'abnormal' preterm babies in the PT populations that were analyzed as specific case studies. CONCLUSIONS: Preliminary results motivate the adoption of this tool in clinical practice allowing for a systematic accumulation of cases in the Biobank, thus for improving the accuracy of data analysis performed by MIMAS and ultimately allowing the adoption of data mining techniques.

A Vexing Question in Motor Control: The Degrees of Freedom Problem.

The human "marionette" is extremely complex and multi-articulated: anatomical redundancy (in terms of Degrees of Freedom: DoFs), kinematic redundancy (movements can have different trajectories, velocities, and accelerations and yet achieve the same goal, according to the principle of Motor Equivalence), and neurophysiological redundancy (many more muscles than DoFs and multiple motor units for each muscle). Although it is quite obvious that such abundance is not noxious at all because, in contrast, it is instrumental for motor learning, allowing the nervous system to "explore" the space of feasible actions before settling on an elegant and possibly optimal solution, the crucial question then boils down to figure out how the nervous system "chooses/selects/recruits/modulates" task-dependent subsets of countless assemblies of DoFs as functional motor synergies. Despite this daunting conceptual riddle, human purposive behavior in daily life activities is a proof of concept that solutions can be found easily and quickly by the embodied brain of the human cognitive agent. The point of view suggested in this essay is to frame the question above in the old-fashioned but still seminal observation by Marr and Poggio that cognitive agents should be regarded as Generalized Information Processing Systems (GIPS) and should be investigated according to three nearly independent but complementary levels of analysis: 1) the computational level, 2) the algorithmic level, and 3) the implementation level. In this framework, we attempt to discriminate as well as aggregate the different hypotheses and solutions proposed so far: the optimal control hypothesis, the muscle synergy hypothesis, the equilibrium point hypothesis, or the uncontrolled manifold hypothesis, to mention the most popular ones. The proposed GIPS follows the strategy of factoring out shaping and timing by adopting a force-field based approach (the Passive Motion Paradigm) that is inspired by the Equilibrium Point Hypothesis, extended in such a way to represent covert as well overt actions. In particular, it is shown how this approach can explain spatio-temporal invariances and, at the same time, solve the Degrees of Freedom Problem.

Postural instability via a loss of intermittent control in elderly and patients with Parkinson's disease: A model-based and data-driven approach.

Postural instability is one of the major symptoms of Parkinson's disease. Here, we assimilated a model of intermittent delay feedback control during quiet standing into postural sway data from healthy young and elderly individuals as well as patients with Parkinson's disease to elucidate the possible mechanisms of instability. Specifically, we estimated the joint probability distribution of a set of parameters in the model using the Bayesian parameter inference such that the model with the inferred parameters can best-fit sway data for each individual. It was expected that the parameter values for three populations would distribute differently in the parameter space depending on their balance capability. Because the intermittent control model is parameterized by a parameter associated with the degree of intermittency in the control, it can represent not only the intermittent model but also the traditional continuous control model with no intermittency. We showed that the inferred parameter values for the three groups of individuals are classified into two major groups in the parameter space: one represents the intermittent control mostly for healthy people and patients with mild postural symptoms and the other the continuous control mostly for some elderly and patients with severe postural symptoms. The results of this study may be interpreted by postulating that increased postural instability in most Parkinson's patients and some elderly persons might be characterized as a dynamical disease.

Centre of pressure versus centre of mass stabilization strategies: the tightrope balancing case.

This study proposes a generalization of the ankle and hip postural strategies to be applied to the large class of skills that share the same basic challenge of defeating the destabilizing effect of gravity on the basis of the same neuromotor control organization, adapted and specialized to a variable number of degrees of freedom, different body parts, different muscles and different sensory feedback channels. In all the cases, we can identify two crucial elements (the CoP, centre of pressure and the CoM, centre of mass) and the central point of the paper is that most balancing skills can be framed in the CoP-CoM interplay and can be modelled as a combination/alternation of two basic stabilization strategies: the standard well-investigated COPS (or CoP stabilization strategy, the default option), where the CoM is the controlled variable and the CoP is the control variable, and the less investigated COMS (or CoM stabilization strategy), where CoP and CoM must exchange their role because the range of motion of the CoP is strongly constrained by environmental conditions. The paper focuses on the tightrope balancing skill where sway control in the sagittal plane is modelled in terms of the COPS while the more challenging sway in the coronal plane is modelled in terms of the COMS, with the support of a suitable balance pole. Both stabilization strategies are implemented as state-space intermittent, delayed feedback controllers, independent of each other. Extensive simulations support the degree of plausibility, generality and robustness of the proposed approach.

State-space intermittent feedback stabilization of a dual balancing task.

Balancing the body in upright standing and balancing a stick on the fingertip are two examples of unstable tasks that, in spite of strong motor and sensory differences, appear to share a similar motor control paradigm, namely a state-space intermittent feedback stabilization mechanism. In this study subjects were required to perform the two tasks simultaneously, with the purpose of highlighting both the coordination between the two skills and the underlying interaction between the corresponding controllers. The experimental results reveal, in particular, that upright standing (the less critical task) is modified in an adaptive way, in order to facilitate the more critical task (stick balancing), but keeping the overall spatio-temporal signature well known in regular upright standing. We were then faced with the following question: to which extent the physical/biomechanical interaction between the two independent intermittent controllers is capable to explain the dual task coordination patterns, without the need to introduce an additional, supervisory layer/module? By comparing the experimental data with the output of a simulation study we support the former hypothesis, suggesting that it is made possible by the intrinsic robustness of both state-space intermittent feedback stabilization mechanisms.

Vibrotactile Feedback for Improving Standing Balance.

Maintaining balance standing upright is an active process that complements the stabilizing properties of muscle stiffness with feedback control driven by independent sensory channels: proprioceptive, visual, and vestibular. Considering that the contribution of these channels is additive, we investigated to what extent providing an additional channel, based on vibrotactile stimulation, may improve balance control. This study focused only on healthy young participants for evaluating the effects of different encoding methods and the importance of the informational content. We built a device that provides a vibrotactile feedback using two vibration motors placed on the anterior and posterior part of the body, at the L5 level. The vibration was synchronized with an accelerometric measurement encoding a combination of the position and acceleration of the body center of mass in the anterior-posterior direction. The goal was to investigate the efficacy of the information encoded by this feedback in modifying postural patterns, comparing, in particular, two different encoding methods: vibration always on and vibration with a dead zone, i.e., silent in a region around the natural stance posture. We also studied if after the exposure, the participants modified their normal oscillation patterns, i.e., if there were after effects. Finally, we investigated if these effects depended on the informational content of the feedback, introducing trials with vibration unrelated to the actual postural oscillations (sham feedback). Twenty-four participants were asked to stand still with their eyes closed, alternating trials with and without vibrotactile feedback: nine were tested with vibration always on and sham feedback, fifteen with dead zone feedback. The results show that synchronized vibrotactile feedback reduces significantly the sway amplitude while increasing the frequency in anterior-posterior and medial-lateral directions. The two encoding methods had no different effects of reducing the amount of postural sway during exposure to vibration, however only the dead-zone feedback led to short-term after effects. The presence of sham vibration, instead, increased the sway amplitude, highlighting the importance of the encoded information.

Tactile and proprioceptive dysfunction differentiates cervical dystonia with and without tremor.

OBJECTIVE: To determine whether different phenotypes of cervical dystonia (CD) express different types and levels of somatosensory impairment. METHODS: We assessed somatosensory function in patients with CD with and without tremor (n = 12 each) and in healthy age-matched controls (n = 22) by measuring tactile temporal discrimination thresholds of the nondystonic forearm and proprioceptive acuity in both the dystonic (head/neck) and nondystonic body segments (forearm/hand) using a joint position-matching task. The head or the wrist was passively displaced along different axes to distinct joint positions by the experimenter or through a robotic exoskeleton. Participants actively reproduced the experienced joint position, and the absolute joint position-matching error between the target and the reproduced positions served as a marker of proprioceptive acuity. RESULTS: Tactile temporal discrimination thresholds were significantly elevated in both CD subgroups compared to controls. Proprioceptive acuity of both the dystonic and nondystonic body segments was elevated in patients with CD and tremor with respect to both healthy controls and patients with CD without tremor. That is, tactile abnormalities were a shared dysfunction of both CD phenotypes, while proprioceptive dysfunction was observed in patients with CD with tremor. CONCLUSIONS: Our findings suggest that the pathophysiology in CD can be characterized by 2 abnormal neural processes: a dysfunctional somatosensory gating mechanism involving the basal ganglia that triggers involuntary muscle spasms and abnormal processing of proprioceptive information within a defective corticocerebellar loop, likely affecting the feedback and feedforward control of head positioning. This dysfunction is expressed mainly in CD with tremor.

Phase resetting and intermittent control at the edge of stability in a simple biped model generates 1/f-like gait cycle variability.

The 1/f-like gait cycle variability, characterized by temporal changes in stride-time intervals during steady-state human walking, is a well-documented gait characteristic. Such gait fractality is apparent in healthy young adults, but tends to disappear in the elderly and patients with neurological diseases. However, mechanisms that give rise to gait fractality have yet to be fully clarified. We aimed to provide novel insights into neuro-mechanical mechanisms of gait fractality, based on a numerical simulation model of biped walking. A previously developed heel-toe footed, seven-rigid-link biped model with human-like body parameters in the sagittal plane was implemented and expanded. It has been shown that the gait model, stabilized rigidly by means of impedance control with large values of proportional (P) and derivative (D) gains for a linear feedback controller, is destabilized only in a low-dimensional eigenspace, as P and D decrease below and even far below critical values. Such low-dimensional linear instability can be compensated by impulsive, phase-dependent actions of nonlinear controllers (phase resetting and intermittent controllers), leading to the flexible walking with joint impedance in the model being as small as that in humans. Here, we added white noise to the model to examine P-value-dependent stochastic dynamics of the model for small D-values. The simulation results demonstrated that introduction of the nonlinear controllers in the model determined the fractal features of gait for a wide range of the P-values, provided that the model operates near the edge of stability. In other words, neither the model stabilized only by pure impedance control even at the edge of linear stability, nor the model stabilized by specific nonlinear controllers, but with P-values far inside the stability region, could induce gait fractality. Although only limited types of controllers were examined, we suggest that the impulsive nonlinear controllers and criticality could be major mechanisms for the genesis of gait fractality.