Competitive settings may be promising for neurorehabilitation as they may lead to stronger patient motivation. However, the mechanisms underlying the development of an interaction in competitive settings are not well understood. We use a dyadic haptic interface to investigate the dynamics of competitive interaction. Pairs of participants engaged in a ball game (a penalty kick challenge), with an attacker and a defender. To explore how the participants adapt to each other over game iterations, we manipulated the amount of information available to each one about the location of their opponent. We found that under the different experimental conditions, the participants adjusted their behavior in response to their opponents' strategy change. These observations are consistent with simulations with a computational model assuming that during interaction the participants use optimal (Bayesian) perception and action selection (based on game theory). Taken together, experiments and simulations suggest that the participants develop and update a flexible, adaptive predictor of their opponents next action, which helps them adjust their behavior to optimize performance. These findings provide a foundation for further research into the mechanisms that drive adaptive behavior in competitive scenarios, and provide new insights into how competitive environments can be used in rehabilitation.

Personalized technology-assisted neurorehabilitation should rely on understanding the mechanisms of spontaneous recovery. We describe a minimal model of neural control of arm movements and neuromotor recovery, consisting of an optimal feedback controller, responsible for generating the motor command, and a corticospinal model, responsible for transferring the motor command to the muscles. Impairment is described as a lack or reduced gain in specific corticospinal pathways. Recovery is modeled as the interplay between restitution or true recovery - through the reorganization of the spared cortico-spinal connectivity, mediated by a form of use-dependent plasticity - and compensation - determining the high-level commands that lead to movements that maximize the desired payoff by minimizing the effect of impairment.True recovery relies on the amount of activation of the impaired actuators. Compensatory strategies emerge as a consequence of adaptation of the internal model of the body to incorporate the impairment.Clinical relevance Modeling the mechanisms underlying recovery after brain injury may provide insights for developing patient digital twins, which can be embedded in robot controllers to allow for personalized rehabilitation.

In stroke rehabilitation, tailoring assistance to individual needs is crucial for more effective training. This study investigates the control architecture of an artificial partner (AP) inspired by game theory. The AP modulates assistance in a planar reaching task using adaptive control strategies. We compare AP performance in "lazy" and "generous" conditions. Results show that the AP adjusts its assistance effectively based on game-theoretic principles. This approach shows promise for enhancing robot-assisted rehabilitation through personalized therapy. Future research will explore the long-term effects of these policies and refine the AP's sensory system and state observer for improved precision.

Sensorimotor impairments following stroke frequently result in diminished voluntary control of the ankle, contributing to deficits in balance and gait. Robotic training paradigms targeting ankle motor control often use an assist-as-needed strategy, where compliant guidance is provided to assist movements towards a target trajectory. However, interaction with “perfect” reference trajectories may overly constrain movements during training and has been shown to limit learning in many upper-limb contexts; alternatives to robotic assistance have rarely been explored for post-stroke ankle training. Inspired by human-robot-human interaction studies, we investigated whether physical interaction with a therapist—termed human interaction—offers advantages over traditional trajectory guidance regarding short-term learning. In a within-subject design, nine individuals with chronic stroke (61.6 ± 14.3 years) performed a 1-DoF visuomotor tracking task while wearing ankle robots designed to train dorsiflexion and plantarflexion movements. Two robotic training methods were evaluated in separate visits: (1) compliant connection to a sinusoidal target trajectory (i.e., trajectory guidance) and (2) compliant connection to a physical therapist who tracked the same target trajectory (i.e., human interaction). In each visit, tracking performance (i.e., errors, movement smoothness) and muscle activation were evaluated during and immediately after training. Both training types improved tracking accuracy and movement smoothness during training, however random error was more significantly suppressed with trajectory guidance. Immediately after training, we found no significant difference in tracking accuracy or movement smoothness across training types. However, participants demonstrated significantly higher dorsiflexor activation after training with human interaction compared to trajectory guidance. Our results suggest that human interaction is a viable strategy for training ankle movements in chronic stroke participants, likely by providing assistance without over-constraining an individual’s movement smoothness or variability. Training while physically interacting with a partner could serve as an effective alternative to conventional robot-guided therapy for post-stroke ankle rehabilitation, though further studies with larger cohorts are needed to assess the generalization of this approach regarding long-term retention and functional improvement. Registry: clinicaltrials.gov, TRN: NCT04578665, Registration date: 8 October 2020.

Objective.In this study, we present a novel computational framework that combines the Hindmarsh-Rose (HR) neuronal model with evolutionary game theory on networks to simulate and interpret synaptic-level interactions within neuronal populations. Our approach preserves the features of the HR model-capable of generating both spiking and bursting dynamics-while integrating game-theoretic principles that govern the balance between emulative and non-emulative behaviors across neurons.Approach.Neurons were modeled as strategic agents whose interactions evolve according to game-theoretic principles, allowing us to capture emergent network dynamics beyond classical electrophysiological analyses. A key innovation of our work is the formulation of a parameter estimation method based on adaptive observers, which enables the recovery of game-theoretic parameters solely from partial state observations. The proposed framework is validated through numerical simulations, demonstrating its ability to recover hidden parameters and accurately predict system behavior under diverse conditions.Main results.By applying the devised approach to synthetic datasets mimicking real electrophysiological recordings, we highlight its applicability in distinguishing neuronal populations based on their strategic interactions. In this context, the model is shown to faithfully reproduce both spiking and bursting behaviors, capturing the diverse electrophysiological patterns observed inin vitroexperimental settings. Furthermore, we explore the potential of this model in experimental data analysis by suggesting that the estimated parameters may serve as discriminative markers for different neuronal types and structural characteristics.Significance.The integration of dynamical systems theory, game-theoretic modeling, and adaptive estimation provides a robust quantitative tool for investigating complex neuronal network dynamics. Our results quantitatively demonstrate the scalability and accuracy of the method in parameter estimation, reinforcing its value for systematic analysis of synaptic interactions and advancing our understanding of neuronal network dynamics.

Shoulder amputation is a complex condition that significantly challenges prosthetic design and control. Here we propose a master-slave control paradigm to operate a modular whole-arm prosthesis (HANNES Arm). Our solution uses Inertial Measurement Units (IMUs) to estimate the movements of the intact arm, which are then replicated in the prosthetic arm. The use of quaternion representations for the kinematics of both human and prosthetic arm kinematics ensures computationally efficient and singularity-free motion tracking and control. The algorithm was validated through simulations, integrating a virtual model of the prosthesis with a personalized human arm model. The human arm configuration was then used to compute the joint rotations of the prosthetic device, which served as reference signals for its control. Despite minor alignment errors, the results demonstrated satisfactory performance, aligning closely with qualitative assessments conducted using an optical motion tracking system (Vicon). Beyond technical innovation, this research contributes to the study of human arm movement synergies, offering new directions for future investigations. The algorithm’s adaptability to individual variations highlights its potential for personalized prosthetic applications, ultimately enhancing the quality of life for individuals with upper limb amputations.Clinical relevance This research not only contributes to technical advancements in the prosthesis control field, but also fosters a deeper understanding of the relationship between humans and prosthetic devices, paving the way for a future where prosthetics seamlessly integrate into daily life.

Despite advancements in myoelectric control of upper-limb prostheses, their use remains challenging and the rejection rates are still high. One step towards improving the human-prosthesis interfacing is to implement artificial sensory feedback as this can improve performance and user experience. While some late-generation prostheses incorporate supplementary feedback, there is still a lack of a principled method to investigate the impact of feedback and evaluate its effectiveness. In this study, we focused on how the subjects learn to operate an upper-limb prosthesis to control the grip force in a grasping task under visual control. We manipulated target grip force, and sensory and motor noise and investigated how these factors affected control performance. To interpret the empirical findings, we developed a computational model in which learning is described as the gradual acquisition of an internal representation of the prosthesis transfer function. The experimental data were collected in 20 non-disabled subjects and the results showed that the precision of control decreased with the increase in the target force and the level of motor noise, whereas the sensory noise did not have a significant impact. The model qualitatively reproduced the main experimental findings, in particular the dependence of performance on motor noise magnitude. The developed model is the first promising step towards capturing the behavior of prosthesis user and developing a more comprehensive understanding of the interaction between the different factors governing closed-loop prosthesis control.

This study investigates the dynamics of geometric hand synergies during manipulation tasks, contrasting them with grasping movements. The objective is to identify dominant synergies and their contribution to dexterous tasks. Using an 18-DoF Dataglove, we recorded the hand movements of six able-bodied participants during 18 daily activities. To analyze the data, principal component analysis (PCA) was used to determine the synergies for each subject. A k-means clustering approach was applied to compare synergies across subjects and identify similarities. Furthermore, the most relevant joints were identified based on their contributions to the principal components, with those exhibiting the highest loadings in the dominant synergy considered key for manipulation. Results revealed that manipulation tasks are described by 7 dominant synergies. Additionally, to perform the tasks a greater joint independence is required compared to grasping, with significant contributions from the proximal interphalangeal (PIP) joints of the middle, ring, and little fingers.Clinical relevance— Identifying the most involved hand joints during manipulation will help guide the design of the next generation of advanced bionic limbs.

During daily life activities we not only gather information regarding the environment, but also regarding other humans which are performing similar actions. Even if not specifically required by the task, people affect other motion plans through different sensory modalities and may align their plans in a subtle way even if not required. This study explores the role of visual and auditory feedback in tasks with a sensory connection between partners. We used a dual robotic interface to test various visual and auditory coupling modalities, translating the spatial dynamics of the partners into distinctive feedback. Results from visual experiment showed that the explicit representation of the partner position along with own position greatly improves coordination between them. The auditory experiment emphasized the effectiveness of binaurally presented spatio-temporally discrete auditory cues. The current study provides insights relevant to the design of novel enriched rehabilitative protocols which rely on the mechanisms underlying interaction.

Interpersonal coordination relies on the ability to predict a partner's future actions and use these predictions for decision-making. Social abilities are impaired in pathologies like dementia, Alzheimer's, and autism spectrum disorder. This study uses a robotic interface where humans play a strategic game (Stag Hunt) with artificial partners, whose behavior is based on a computational interaction model and can be adjusted to simulate different partner types. Humans adapt to the features of the artificial partners, making this setup a potential tool for assessing interaction abilities in both healthy individuals and those with cognitive impairments.

We refer to joint action as any social interaction where two or more people coordinate their actions in space and time to bring about a change in the environment. We are constantly involved in joint actions in our daily life lifting a piece of furniture, playing team sport, and during a rehabilitation session. Here, we aim at characterizing emergent coordination in pairs of participants mechanically coupled through a virtual elastic band while moving the hand through multiple via-points. We assessed the level of coordination in terms of temporal synchronization and trajectory similarity metrics. Overall, we found that participants changed their behaviour when mechanically coupled toward coordination, both in space and time, compared to their solo performances. However, coordination rapidly disappeared when the mechanical coupling was removed. These results provide insights into the understanding of the mechanisms underlying the development of joint coordination while haptically connected.

Coordinating with others is part of our everyday experience. Previous studies using sensorimotor coordination games suggest that human dyads develop coordination strategies that can be interpreted as Nash equilibria. However, if the players are uncertain about what their partner is doing, they develop coordination strategies which are robust to the actual partner’s actions. This has suggested that humans select their actions based on an explicit prediction of what the partner will be doing—a partner model—which is probabilistic by nature. However, the mechanisms underlying the development of a joint coordination over repeated trials remain unknown. Very much like sensorimotor adaptation of individuals to external perturbations (eg force fields or visual rotations), dynamical models may help to understand how joint coordination develops over repeated trials. Here we present a general computational model—based on game theory and Bayesian estimation—designed to understand the mechanisms underlying the development of a joint coordination over repeated trials. Joint tasks are modeled as quadratic games, where each participant’s task is expressed as a quadratic cost function. Each participant predicts their partner’s next move (partner model) by optimally combining predictions and sensory observations, and selects their actions through a stochastic optimization of its expected cost, given the partner model. The model parameters include perceptual uncertainty (sensory noise), partner representation (retention rate and internale noise), uncertainty in action selection and its rate of decay (which can be interpreted as the action’s learning rate). The model can be used in two ways: (i) to simulate interactive behaviors, thus helping to make specific predictions in the context of a given joint action scenario; and (ii) to analyze the action time series in actual experiments, thus providing quantitative metrics that describe individual behaviors during an actual joint action. We demonstrate the model in a variety of joint action scenarios. In a sensorimotor version of the Stag Hunt game, the model predicts that different representations of the partner lead to different Nash equilibria. In a joint two via-point (2-VP) reaching task, in which the actions consist of complex trajectories, the model captures well the observed temporal evolution of performance. For this task we also estimated the model parameters from experimental observations, which provided a comprehensive characterization of individual dyad participants. Computational models of joint action may help identifying the factors preventing or facilitating the development of coordination. They can be used in clinical settings, to interpret the observed behaviors in individuals with impaired interaction capabilities. They may also provide a theoretical basis to devise artificial agents that establish forms of coordination that facilitate neuromotor recovery.

Many of our everyday activities take place in social settings and are coordinated with other persons. Even seemingly simple interactions, like a pair of workers sawing timber with back-and-forth movements or a therapist providing physical therapy to a patient, require that two individual minds are somehow connected and their bodies coordinated. The characteristic feature of these interactions is that subjects influence each other's behavior through sensorimotor exchanges within continuous action spaces, continuously in time, and possibly over repeated trials. Previous joint action studies using sensorimotor games suggest that human dyads develop coordination strategies that can be interpreted as Nash equilibria. Uncertainty about the intended opponent's actions may shape the resulting coordination. However, the mechanisms underlying the development of joint coordination over repeated trials are poorly understood. This chapter reviews experimental studies on how two players that are mechanically connected develop joint coordination. We then present a general computational framework—based on game theory and Bayesian estimation—to understand the underlying mechanisms. Models can be used to implement artificial “partners” with an inherent ability to establish such coordination. Human-artificial dyads develop forms of joint coordinations that are similar to those observed in human-human dyads. We finally discuss the implications of these studies for the development of artificial robotic “therapists,” with an “optimal” capability to understand patient impairment and to facilitate their recovery.

As the global population ages, neurodegenerative diseases like Alzheimer (AD) and Parkinson (PD) are increasing, necessitating better remote monitoring methods. Here we explore quantitative analysis of handwriting skills as a tool for predicting cognitive impairment and thus, monitoring the degeneration of the disease. A novel probabilistic model using bump functions was developed to analyze handwriting's spatial organization, capturing their complexity and variability. The key findings are that AD subjects tend to organize the same dictated text into more rows than PD, and the rows' average tilt and height are predictive of cognitive functions. This approach offers a non-invasive, accessible, and effective solution for monitoring of neurodegenerative diseases.

Future healthcare is transitioning toward a decentralization of patient care, in which personal care is increasingly moved at the patient home and surrounding areas, while hospitals concentrate primarily on procedures that cannot be performed elsewhere, such as surgeries or outpatient examinations. The healthcare system in the Liguria region (Italy) is currently developing a new Center for Computational and Technological Medicine (CMCT), which is intended to facilitate and support this transition. As a component of the strategic planning and design process, this study examines the development and organization of telemedicine services across a range of chosen Italian and European institutions that share similarities with CMCT in terms of scope and scale. We specifically focus on telemedicine services - how they are governed, their main fields of application. The analysis confirmed the growing importance of telemedicine as part of the new vision of medicine, in which the patient is at the center.

During the development and assessment of an exoskeleton, many different analyzes need to be performed. The most frequently used evaluate the changes in muscle activations, metabolic consumption, kinematics, and kinetics. Since human-exoskeleton interactions are based on the exchange of forces and torques, the latter of these, kinetic analyzes, are essential and provide indispensable evaluation indices. Kinetic analyzes, however, require access to, and use of, complex experimental apparatus, involving many instruments and implicating lengthy data analysis processes. The proposed methodology in this paper, which is based on data collected via EMG and motion capture systems, considerably reduces this burden by calculating kinetic parameters, such as torque and power, without needing ground reaction force measurements. This considerably reduces the number of instruments used, allows the calculation of kinetic parameters even when the use of force sensors is problematic, does not need any dedicated software, and will be shown to have high statistical validity. The method, in fact, combines data found in the literature with those collected in the laboratory, allowing the analysis to be carried out over a much greater number of cycles than would normally be collected with force plates, thus enabling easy access to statistical analysis. This new approach evaluates the kinetic effects of the exoskeleton with respect to changes induced in the user's kinematics and muscular activation patterns and provides indices that quantify the assistance in terms of torque (AMI) and power (API). Following the User-Center Design approach, which requires driving the development process as feedback from the assessment process, this aspect is critical. Therefore, by enabling easy access to the assessment process, the development of exoskeletons could be positively affected.

In the last years, artificial partners have been proposed as tools to study joint action, as they would allow to address joint behaviors in more controlled experimental conditions. Here we present an artificial partner architecture which is capable of integrating all the available information about its human counterpart and to develop efficient and natural forms of coordination. The model uses an extended state observer which combines prior information, motor commands and sensory observations to infer the partner's ongoing actions (partner model). Over trials, these estimates are gradually incorporated into action selection. Using a joint planar task in which the partners are required to perform reaching movements while mechanically coupled, we demonstrate that the artificial partner develops an internal representation of its human counterpart, whose accuracy depends on the degree of mechanical coupling and on the reliability of the sensory information. We also show that human-artificial dyads develop coordination strategies which closely resemble those observed in human-human dyads and can be interpreted as Nash equilibria. The proposed approach may provide insights for the understanding of the mechanisms underlying human-human interaction. Further, it may inform the development of novel neuro-rehabilitative solutions and more efficient human-machine interfaces.

The sense of agency – the subjective feeling of being in control of our own actions – is one central aspect of the phenomenology of action. Computational models provided important contributions toward unveiling the mechanisms underlying the sense of agency in individual action. In particular, the sense of agency is believed to be related to the match between the actual and predicted consequences of our own actions (comparator model). In the study of joint action, models are even more necessary to understand the mechanisms underlying the development of coordination strategies and how the subjective experiences of control emerge during the interaction. In a joint action, we not only need to predict the consequences of our own actions; we also need to predict the actions and intentions of our partner, and to integrate these predictions to infer their joint consequences. Understanding our partner and developing mutually satisfactory coordination strategies are key components of joint action and in the development of the sense of joint agency. Here we discuss a computational architecture which addresses the sense of agency during intentional, real-time joint action. We first reformulate previous accounts of the sense of agency in probabilistic terms, as the combination of prior beliefs about the action goals and constraints, and the likelihood of the predicted movement outcomes. To look at the sense of joint agency, we extend classical computational motor control concepts - optimal estimation and optimal control. Regarding estimation, we argue that in joint action the players not only need to predict the consequences of their own actions, but also need to predict partner’s actions and intentions (a ‘partner model’) and to integrate these predictions to infer their joint consequences. As regards action selection, we use differential game theory – in which actions develop in continuous space and time - to formulate the problem of establishing a stable form of coordination and as a natural extension of optimal control to joint action. The resulting model posits two concurrent observer-controller loops, accounting for ‘joint’ and ‘self’ action control. The two observers quantify the likelihoods of being in control alone or jointly. Combined with prior beliefs, they provide weighing signals which are used to modulate the ‘joint’ and ‘self’ motor commands. We argue that these signals can be interpreted as the subjective sense of joint and self agency. We demonstrate the model predictions by simulating a sensorimotor interactive task where two players are mechanically coupled and are instructed to perform planar movements to reach a shared final target by crossing two differently located intermediate targets. In particular, we explore the relation between self and joint agency and the information available to each player about their partner. The proposed model provides a coherent picture of the inter-relation of prediction, control, and the sense of agency in a broader range of joint actions.

Abstract Assistive strategies for occupational back-support exoskeletons have focused, mostly, on lifting tasks. However, in occupational scenarios, it is important to account not only for lifting but also for other activities. This can be done exploiting human activity recognition algorithms that can identify which task the user is performing and trigger the appropriate assistive strategy. We refer to this ability as exoskeleton versatility. To evaluate versatility, we propose to focus both on the ability of the device to reduce muscle activation (efficacy) and on its interaction with the user (dynamic fit). To this end, we performed an experimental study involving $ 10 $ healthy subjects replicating the working activities of a manufacturing plant. To compare versatile and non-versatile exoskeletons, our device, XoTrunk, was controlled with two different strategies. Correspondingly, we collected muscle activity, kinematic variables and users’ subjective feedbacks. Also, we evaluated the task recognition performance of the device. The results show that XoTrunk is capable of reducing muscle activation by up to $ 40\% $ in lifting and $ 30\% $ in carrying. However, the non-versatile control strategy hindered the users’ natural gait (e.g., $ -24\% $ reduction of hip flexion), which could potentially lower the exoskeleton acceptance. Detecting carrying activities and adapting the control strategy, resulted in a more natural gait (e.g., $ +9\% $ increase of hip flexion). The classifier analyzed in this work, showed promising performance (online accuracy > 91%). Finally, we conducted 9 hours of field testing, involving four users. Initial subjective feedbacks on the exoskeleton versatility, are presented at the end of this work.