Walking and falling: Using robot simulations to model the role of errors in infant walking.

What is the optimal penalty for errors in infant skill learning? Behavioral analyses indicate that errors are frequent but trivial as infants acquire foundational skills. In learning to walk, for example, falling is commonplace but appears to incur only a negligible penalty. Behavioral data, however, cannot reveal whether a low penalty for falling is beneficial for learning to walk. Here, we used a simulated bipedal robot as an embodied model to test the optimal penalty for errors in learning to walk. We trained the robot to walk using 12,500 independent simulations on walking paths produced by infants during free play and systematically varied the penalty for falling-a level of precision, control, and magnitude impossible with real infants. When trained with lower penalties for falling, the robot learned to walk farther and better on familiar, trained paths and better generalized its learning to novel, untrained paths. Indeed, zero penalty for errors led to the best performance for both learning and generalization. Moreover, the beneficial effects of a low penalty were stronger for generalization than for learning. Robot simulations corroborate prior behavioral data and suggest that a low penalty for errors helps infants learn foundational skills (e.g., walking, talking, and social interactions) that require immense flexibility, creativity, and adaptability. RESEARCH HIGHLIGHTS: During infant skill acquisition, errors are commonplace but appear to incur a low penalty; when learning to walk, for example, falls are frequent but trivial. To test the optimal penalty for errors, we trained a simulated robot to walk using real infant paths and systematically manipulated the penalty for falling. Lower penalties in training led to better performance on familiar, trained paths and on novel untrained paths, and zero penalty was most beneficial. Benefits of a low penalty were stronger for untrained than for trained paths, suggesting that discounting errors facilitates acquiring skills that require immense flexibility and generalization.

The power of multivariate approach in identifying EEG correlates of interlimb coupling.

Interlimb coupling refers to the interaction between movements of one limb and movements of other limbs. Understanding mechanisms underlying this effect is important to real life because it reflects the level of interdependence between the limbs that plays a role in daily activities including tool use, cooking, or playing musical instruments. Interlimb coupling involves multiple brain regions working together, including coordination of neural activity in sensory and motor regions across the two hemispheres. Traditional neuroscience research took a univariate approach to identify neural features that correspond to behavioural coupling measures. Yet, this approach reduces the complexity of the neural activity during interlimb tasks to one value. In this brief research report, we argue that identifying neural correlates of interlimb coupling would benefit from a multivariate approach in which full patterns from multiple sources are used to predict behavioural coupling. We demonstrate the feasibility of this approach in an exploratory EEG study where participants (n = 10) completed 240 trials of a well-established drawing paradigm that involves interlimb coupling. Using artificial neural network (ANN), we show that multivariate representation of the EEG signal significantly captures the interlimb coupling during bimanual drawing whereas univariate analyses failed to identify such correlates. Our findings demonstrate that analysing distributed patterns of multiple EEG channels is more sensitive than single-value techniques in uncovering subtle differences between multiple neural signals. Using such techniques can improve identification of neural correlates of complex motor behaviours.

The effects of typical ageing on cognitive control: recent advances and future directions.

Cognitive control is one of the most fundamental aspects of human life. Its ageing is an important contemporary research area due to the needs of the growing ageing population, such as prolonged independence and quality of life. Traditional ageing research argued for a global decline in cognitive control with age, typically characterised by slowing processing speed and driven by changes in the frontal cortex. However, recent advances questioned this perspective by demonstrating high heterogeneity in the ageing data, domain-specific declines, activity changes in resting state networks, and increased functional connectivity. Moreover, improvements in neuroimaging techniques have enabled researchers to develop compensatory models of neural reorganisation that helps negate the effects of neural losses and promote cognitive control. In this article on typical ageing, we review recent behavioural and neural findings related to the decline in cognitive control among older adults. We begin by reviewing traditional perspectives and continue with how recent work challenged those perspectives. In the discussion section, we propose key areas of focus for future research in the field.

Motor learning in hemi-Parkinson using VR-manipulated sensory feedback.

AIMS: Modalities for rehabilitation of the neurologically affected upper-limb (UL) are generally of limited benefit. The majority of patients seriously affected by UL paresis remain with severe motor disability, despite all rehabilitation efforts. Consequently, extensive clinical research is dedicated to develop novel strategies aimed to improve the functional outcome of the affected UL. We have developed a novel virtual-reality training tool that exploits the voluntary control of one hand and provides real-time movement-based manipulated sensory feedback as if the other hand is the one that moves. The aim of this study was to expand our previous results, obtained in healthy subjects, to examine the utility of this training setup in the context of neuro-rehabilitation. METHODS: We tested the training setup in patient LA, a young man with significant unilateral UL dysfunction stemming from hemi-parkinsonism. LA underwent daily intervention in which he intensively trained the non-affected upper limb, while receiving online sensory feedback that created an illusory perception of control over the affected limb. Neural changes were assessed using functional magnetic resonance imaging (fMRI) scans before and after training. RESULTS: Training-induced behavioral gains were accompanied by enhanced activation in the pre-frontal cortex and a widespread increase in resting-state functional connectivity. DISCUSSION: Our combination of cutting edge technologies, insights gained from basic motor neuroscience in healthy subjects and well-known clinical treatments, hold promise for the pursuit of finding novel and more efficient rehabilitation schemes for patients suffering from hemiplegia.Implications for rehabilitationAssistive devices used in hospitals to support patients with hemiparesis require expensive equipment and trained personnel - constraining the amount of training that a given patient can receive. The setup we describe is simple and can be easily used at home with the assistance of an untrained caregiver/family member. Once installed at the patient's home, the setup is lightweight, mobile, and can be used with minimal maintenance . Building on advances in machine learning, our software can be adapted to personal use at homes. Our findings can be translated into practice with relatively few adjustments, and our experimental design may be used as an important adjuvant to standard clinical care for upper limb hemiparesis.

Real-time processes in the development of action planning.

Across species and ages, planning multi-step actions is a hallmark of intelligence and critical for survival. Traditionally, researchers adopt a "top-down" approach to action planning by focusing on the ability to create an internal representation of the world that guides the next step in a multi-step action. However, a top-down approach does not inform on underlying mechanisms, so researchers can only speculate about how and why improvements in planning occur. The current study takes a "bottom-up" approach by testing developmental changes in the real-time, moment-to-moment interplay among perceptual, neural, and motor components of action planning using simultaneous video, motion-tracking, head-mounted eye tracking, and electroencephalography (EEG). Preschoolers (n = 32) and adults (n = 22) grasped a hammer with their dominant hand to pound a peg when the hammer handle pointed in different directions. When the handle pointed toward their non-dominant hand, younger children ("nonadaptive planners") used a habitual overhand grip that interfered with wielding the hammer, whereas adults and older children ("adaptive planners") used an adaptive underhand grip. Adaptive and nonadaptive children differed in when and where they directed their gaze to obtain visual information, neural activation of the motor system before reaching, and straightness of their reach trajectories. Nonadaptive children immediately used a habitual overhand grip before gathering visual information, leaving insufficient time to form a plan before acting. Our novel bottom-up approach transcends mere speculation by providing converging evidence that the development of action planning depends on a real-time "tug of war" between habits and information gathering and processing.

Lie to my face: An electromyography approach to the study of deceptive behavior.

BACKGROUND: Deception is present in all walks of life, from social interactions to matters of homeland security. Nevertheless, reliable indicators of deceptive behavior in real-life scenarios remain elusive. METHODS: By integrating electrophysiological and communicative approaches, we demonstrate a new and objective detection approach to identify participant-specific indicators of deceptive behavior in an interactive scenario of a two-person deception task. We recorded participants' facial muscle activity using novel dry screen-printed electrode arrays and applied machine-learning algorithms to identify lies based on brief facial responses. RESULTS: With an average accuracy of 73%, we identified two groups of participants: Those who revealed their lies by activating their cheek muscles and those who activated their eyebrows. We found that the participants lied more often with time, with some switching their telltale muscle groups. Moreover, while the automated classifier, reported here, outperformed untrained human detectors, their performance was correlated, suggesting reliance on shared features. CONCLUSIONS: Our findings demonstrate the feasibility of using wearable electrode arrays in detecting human lies in a social setting and set the stage for future research on individual differences in deception expression.

Children do not distinguish efficient from inefficient actions during observation.

Observation is a powerful way to learn efficient actions from others. However, the role of observers' motor skill in assessing efficiency of others is unknown. Preschoolers are notoriously poor at performing multi-step actions like grasping the handle of a tool. Preschoolers (N = 22) and adults (N = 22) watched video-recorded actors perform efficient and inefficient tool use. Eye tracking showed that preschoolers and adults looked equally long at the videos, but adults looked longer than children at how actors grasped the tool. Deep learning analyses of participants' eye gaze distinguished efficient from inefficient grasps for adults, but not for children. Moreover, only adults showed differential action-related pupil dilation and neural activity (suppressed oscillation power in the mu frequency) while observing efficient vs. inefficient grasps. Thus, children observe multi-step actions without "seeing" whether the initial step is efficient. Findings suggest that observer's own motor efficiency determines whether they can perceive action efficiency in others.

"Dancing" Together: Infant-Mother Locomotor Synchrony.

Pre-mobile infants and caregivers spontaneously engage in a sequence of contingent facial expressions and vocalizations that researchers have referred to as a social "dance." Does this dance continue when both partners are free to move across the floor? Locomotor synchrony was assessed in 13- to 19-month-old infant-mother dyads (N = 30) by tracking each partner's step-to-step location during free play. Although infants moved more than mothers, dyads spontaneously synchronized their locomotor activity. For 27 dyads, the spatiotemporal path of one partner uniquely identified the path of the other. Clustering analyses revealed two patterns of synchrony (mother-follow and yo-yo), and infants were more likely than mothers to lead the dance. Like face-to-face synchrony, locomotor synchrony scaffolds infants' interactions with the outside world.

Modeling Infant Free Play Using Hidden Markov Models.

Infants' free-play behavior is highly variable. However, in developmental science, traditional analysis tools for modeling and understanding variable behavior are limited. Here, we used Hidden Markov Models (HMMs) to capture behavioral states that govern infants' toy selection during 20 minutes of free play in a new environment. We demonstrate that applying HMMs to infant data can identify hidden behavioral states and thereby reveal the underlying structure of infant toy selection and how toy selection changes in real time during spontaneous free play. More broadly, we propose that hidden-state models provide a fruitful avenue for understanding individual differences in spontaneous infant behavior.

Single-cell activity in human STG during perception of phonemes is organized according to manner of articulation.

One of the central tasks of the human auditory system is to extract sound features from incoming acoustic signals that are most critical for speech perception. Specifically, phonological features and phonemes are the building blocks for more complex linguistic entities, such as syllables, words and sentences. Previous ECoG and EEG studies showed that various regions in the superior temporal gyrus (STG) exhibit selective responses to specific phonological features. However, electrical activity recorded by ECoG or EEG grids reflects average responses of large neuronal populations and is therefore limited in providing insights into activity patterns of single neurons. Here, we recorded spiking activity from 45 units in the STG from six neurosurgical patients who performed a listening task with phoneme stimuli. Fourteen units showed significant responsiveness to the stimuli. Using a Naïve-Bayes model, we find that single-cell responses to phonemes are governed by manner-of-articulation features and are organized according to sonority with two main clusters for sonorants and obstruents. We further find that 'neural similarity' (i.e. the similarity of evoked spiking activity between pairs of phonemes) is comparable to the 'perceptual similarity' (i.e. to what extent two phonemes are judged as sounding similar) based on perceptual confusion, assessed behaviorally in healthy subjects. Thus, phonemes that were perceptually similar also had similar neural responses. Taken together, our findings indicate that manner-of-articulation is the dominant organization dimension of phoneme representations at the single-cell level, suggesting a remarkable consistency across levels of analyses, from the single neuron level to that of large neuronal populations and behavior.

Real-Time Assembly of Coordination Patterns in Human Infants.

Flexibility and generativity are fundamental aspects of functional behavior that begin in infancy and improve with experience. How do infants learn to tailor their real-time solutions to variations in local conditions? On a nativist view, the developmental process begins with innate prescribed solutions, and experience elaborates on those solutions to suit variations in the body and the environment. On an emergentist view, infants begin by generating a variety of strategies indiscriminately, and experience teaches them to select solutions tailored to the current relations between their body and the environment. To disentangle these accounts, we observed coordination patterns in 11-month-old pre-walking infants with a range of cruising (moving sideways in an upright posture while holding onto a support) and crawling experience as they cruised over variable distances between two handrails they held for support. We identified infants' coordination patterns using a novel combination of computer-vision, machine-learning, and time-series analyses. As predicted by the emergentist view, the least experienced infants generated multiple coordination patterns inconsistently regardless of body size and handrail distance, whereas the most experienced infants tailored their coordination patterns to body-environment relations and switched solutions only when necessary. Moreover, the beneficial effects of experience were specific to cruising and not crawling, although both skills involve anti-phase coordination among the four limbs. Thus, findings support an emergentist view and suggest that everyday experience with the target skill may promote "learning to learn," where infants learn to assemble the appropriate solution for new problems on the fly.

Look before you fit: The real-time planning cascade in children and adults.

Goal-directed actions involve problem solving-how to coordinate perception and action to get the job done. Whereas previous work focused on the ages at which children succeed in problem solving, we focused on how children solve motor problems in real time. We used object fitting as a model system to understand how perception and action unfold from moment to moment. Preschoolers (N = 25) and adults (N = 24) inserted three-dimensional objects into their corresponding openings in a "shape-sorting" box. We applied a new combination of real-time methods to the problem of object fitting-head-mounted eye tracking to record looking behaviors, video microcoding to record adjustments in object orientation between reach and insertion, and real-time analysis techniques (recurrent quantification analysis and Granger causality) to test the timing relations between visual and manual actions. Children, like adults, solved the problem successfully. However, adults outperformed children in terms of their speed of fitting, and speed depended on when adjustments of object orientation occurred. Adults adjusted object orientation during transport, whereas children adjusted object orientation after arriving at the box. Children's delays in adjustment resulted from delays in looking at the target shape and its corresponding aperture. Findings show that planning is a real-time cascade of perception and action, and looking provides the basis for planning actions prospectively. We suggest that developmental improvements in problem solving are driven by real-time changes in the instigation of the planning cascade and the timing of its components.

The ties that bind: Cradling in Tajikistan.

A traditional childrearing practice-"gahvora" cradling-in Tajikistan and other parts of Central Asia purportedly restricts movement of infants' body and limbs. However, the practice has been documented only informally in anecdotal reports. Thus, this study had two research questions: (1) To what extent are infants' movements restricted in the gahvora? (2) How is time in the gahvora distributed over a 24-hour day in infants from 1-24 months of age? To answer these questions, we video-recorded 146 mothers cradling their infants and interviewed them using 24-hour time diaries to determine the distribution of time infants spent in the gahvora within a day and across age. Infants' movements were indeed severely restricted. Although mothers showed striking uniformity in how they restricted infants' movements, they showed large individual differences in amount and distribution of daily use. Machine learning algorithms yielded three patterns of use: day and nighttime cradling, mostly nighttime cradling, and mostly daytime cradling, suggesting multiple functions of the cradling practice. Across age, time in the gahvora decreased, yet 20% of 12- to 24-month-olds spent more than 15 hours bound in the gahvora. We discuss the challenges and benefits of cultural research, and how the discovery of new phenomena may defy Western assumptions about childrearing and development. Future work will determine whether the extent and timing of restriction impacts infants' physical and psychological development.

Variety Wins: Soccer-Playing Robots and Infant Walking.

Although both infancy and artificial intelligence (AI) researchers are interested in developing systems that produce adaptive, functional behavior, the two disciplines rarely capitalize on their complementary expertise. Here, we used soccer-playing robots to test a central question about the development of infant walking. During natural activity, infants' locomotor paths are immensely varied. They walk along curved, multi-directional paths with frequent starts and stops. Is the variability observed in spontaneous infant walking a "feature" or a "bug?" In other words, is variability beneficial for functional walking performance? To address this question, we trained soccer-playing robots on walking paths generated by infants during free play and tested them in simulated games of "RoboCup." In Tournament 1, we compared the functional performance of a simulated robot soccer team trained on infants' natural paths with teams trained on less varied, geometric paths-straight lines, circles, and squares. Across 1,000 head-to-head simulated soccer matches, the infant-trained team consistently beat all teams trained with less varied walking paths. In Tournament 2, we compared teams trained on different clusters of infant walking paths. The team trained with the most varied combination of path shape, step direction, number of steps, and number of starts and stops outperformed teams trained with less varied paths. This evidence indicates that variety is a crucial feature supporting functional walking performance. More generally, we propose that robotics provides a fruitful avenue for testing hypotheses about infant development; reciprocally, observations of infant behavior may inform research on artificial intelligence.

Suppression of EEG mu rhythm during action observation corresponds with subsequent changes in behavior.

Movement is intrinsically linked to perception such that observing an action induces in the observer behavioral changes during execution of similar actions. Electroencephalogram (EEG) studies have revealed that at the group level, action observation suppresses oscillatory power in mu (8-12 Hz) and beta (15-25 Hz) bands over the sensorimotor cortex - a phenomenon associated with increased excitability of cortical neurons. However, it is unclear whether differences in suppression level across individuals is linked with individual differences in subsequent behavioral changes. Here 32 subjects performed self-paced finger tapping with their right hand before and after observation of a video displaying finger-tapping at either 2 or 4 Hz. Behaviorally, subjects' rate of self-pace tapping increased following observation, with higher increases following 4 Hz observation. The level of EEG power suppression in the low frequency range (low mu; 8-10 Hz) during observation corresponded to subsequent behavioral changes in tapping rate across individuals. Our results demonstrate that observing actions implicitly shifts subsequent execution rates, and that individual differences in the level of this implicit shift can be explained by activity in the sensorimotor cortex during observation.

Perception as a Route for Motor Skill Learning: Perspectives from Neuroscience.

Learning a motor skill requires physical practice that engages neural networks involved in movement. These networks have also been found to be engaged during perception of sensory signals associated with actions. Nonetheless, despite extensive evidence for the existence of such sensory-evoked neural activity in motor pathways, much less is known about their contribution to learning and actual changes in behavior. Primate studies usually involve an overlearned task while studies in humans have largely focused on characterizing activity of the action observation network (AON) in the context of action understanding, theory of mind, and social interactions. Relatively few studies examined neural plasticity induced by perception and its role in transfer of motor knowledge. Here, we review this body of literature and point to future directions for the development of alternative, physiologically grounded ways in which sensory signals could be harnessed to improve motor skills.

Behavioral and neural effects of congruency of visual feedback during short-term motor learning.

Visual feedback can facilitate or interfere with movement execution. Here, we describe behavioral and neural mechanisms by which the congruency of visual feedback during physical practice of a motor skill modulates subsequent performance gains. 18 healthy subjects learned to execute rapid sequences of right hand finger movements during fMRI scans either with or without visual feedback. Feedback consisted of a real-time, movement-based display of virtual hands that was either congruent (right virtual hand movement), or incongruent (left virtual hand movement yoked to the executing right hand). At the group level, right hand performance gains following training with congruent visual feedback were significantly higher relative to training without visual feedback. Conversely, performance gains following training with incongruent visual feedback were significantly lower. Interestingly, across individual subjects these opposite effects correlated. Activation in the Supplementary Motor Area (SMA) during training corresponded to individual differences in subsequent performance gains. Furthermore, functional coupling of SMA with visual cortices predicted individual differences in behavior. Our results demonstrate that some individuals are more sensitive than others to congruency of visual feedback during short-term motor learning and that neural activation in SMA correlates with such inter-individual differences.

Using Virtual Reality to Transfer Motor Skill Knowledge from One Hand to Another.

As far as acquiring motor skills is concerned, training by voluntary physical movement is superior to all other forms of training (e.g. training by observation or passive movement of trainee's hands by a robotic device). This obviously presents a major challenge in the rehabilitation of a paretic limb since voluntary control of physical movement is limited. Here, we describe a novel training scheme we have developed that has the potential to circumvent this major challenge. We exploited the voluntary control of one hand and provided real-time movement-based manipulated sensory feedback as if the other hand is moving. Visual manipulation through virtual reality (VR) was combined with a device that yokes left-hand fingers to passively follow right-hand voluntary finger movements. In healthy subjects, we demonstrate enhanced within-session performance gains of a limb in the absence of voluntary physical training. Results in healthy subjects suggest that training with the unique VR setup might also be beneficial for patients with upper limb hemiparesis by exploiting the voluntary control of their healthy hand to improve rehabilitation of their affected hand.

Short Term Motor-Skill Acquisition Improves with Size of Self-Controlled Virtual Hands.

Visual feedback in general, and from the body in particular, is known to influence the performance of motor skills in humans. However, it is unclear how the acquisition of motor skills depends on specific visual feedback parameters such as the size of performing effector. Here, 21 healthy subjects physically trained to perform sequences of finger movements with their right hand. Through the use of 3D Virtual Reality devices, visual feedback during training consisted of virtual hands presented on the screen, tracking subject's hand movements in real time. Importantly, the setup allowed us to manipulate the size of the displayed virtual hands across experimental conditions. We found that performance gains increase with the size of virtual hands. In contrast, when subjects trained by mere observation (i.e., in the absence of physical movement), manipulating the size of the virtual hand did not significantly affect subsequent performance gains. These results demonstrate that when it comes to short-term motor skill learning, the size of visual feedback matters. Furthermore, these results suggest that highest performance gains in individual subjects are achieved when the size of the virtual hand matches their real hand size. These results may have implications for optimizing motor training schemes.

Neural Network Underlying Intermanual Skill Transfer in Humans.

Physical practice with one hand results in performance gains of the other (un-practiced) hand, yet the role of sensory feedback and underlying neurophysiology is unclear. Healthy subjects learned sequences of finger movements by physical training with their right hand while receiving real-time movement-based visual feedback via 3D virtual reality devices as if their immobile left hand was training. This manipulation resulted in significantly enhanced performance gain with the immobile hand, which was further increased when left-hand fingers were yoked to passively follow right-hand voluntary movements. Neuroimaging data show that, during training with manipulated visual feedback, activity in the left and right superior parietal lobule and their degree of coupling with motor and visual cortex, respectively, correlate with subsequent left-hand performance gain. These results point to a neural network subserving short-term motor skill learning and may have implications for developing new approaches for learning and rehabilitation in patients with unilateral motor deficits.