Intelligent Human-Computer Interaction: Combined Wrist and Forearm Myoelectric Signals for Handwriting Recognition.

Recent studies have highlighted the possibility of using surface electromyographic (EMG) signals to develop human-computer interfaces that are also able to recognize complex motor tasks involving the hand as the handwriting of digits. However, the automatic recognition of words from EMG information has not yet been studied. The aim of this study is to investigate the feasibility of using combined forearm and wrist EMG probes for solving the handwriting recognition problem of 30 words with consolidated machine-learning techniques and aggregating state-of-the-art features extracted in the time and frequency domains. Six healthy subjects, three females and three males aged between 25 and 40 years, were recruited for the study. Two tests in pattern recognition were conducted to assess the possibility of classifying fine hand movements through EMG signals. The first test was designed to assess the feasibility of using consolidated myoelectric control technology with shallow machine-learning methods in the field of handwriting detection. The second test was implemented to assess if specific feature extraction schemes can guarantee high performances with limited complexity of the processing pipeline. Among support vector machine, linear discriminant analysis, and K-nearest neighbours (KNN), the last one showed the best classification performances in the 30-word classification problem, with a mean accuracy of 95% and 85% when using all the features and a specific feature set known as TDAR, respectively. The obtained results confirmed the validity of using combined wrist and forearm EMG data for intelligent handwriting recognition through pattern recognition approaches in real scenarios.

Machine learning to detect, stage and classify diseases and their symptoms based on inertial sensor data: a mapping review.

This article presents a systematic review aimed at mapping the literature published in the last decade on the use of machine learning (ML) for clinical decision-making through wearable inertial sensors. The review aims to analyze the trends, perspectives, strengths, and limitations of current literature in integrating ML and inertial measurements for clinical applications. The review process involved defining four research questions and applying four relevance assessment indicators to filter the search results, providing insights into the pathologies studied, technologies and setups used, data processing schemes, ML techniques applied, and their clinical impact. When combined with ML techniques, inertial measurement units (IMUs) have primarily been utilized to detect and classify diseases and their associated motor symptoms. They have also been used to monitor changes in movement patterns associated with the presence, severity, and progression of pathology across a diverse range of clinical conditions. ML models trained with IMU data have shown potential in improving patient care by objectively classifying and predicting motor symptoms, often with a minimally encumbering setup. The findings contribute to understanding the current state of ML integration with wearable inertial sensors in clinical practice and identify future research directions. Despite the widespread adoption of these technologies and techniques in clinical applications, there is still a need to translate them into routine clinical practice. This underscores the importance of fostering a closer collaboration between technological experts and professionals in the medical field.

Modalities of sequential human robot collaboration trigger different modifications of trunk oscillations.

Introduction: Human robot collaboration is quickly gaining importance in the robotics and ergonomics fields due to its ability to reduce biomechanical risk on the human operator while increasing task efficiency. The performance of the collaboration is typically managed by the introduction of complex algorithms in the robot control schemes to ensure optimality of its behavior; however, a set of tools for characterizing the response of the human operator to the movement of the robot has yet to be developed. Methods: Trunk acceleration was measured and used to define descriptive metrics during various human robot collaboration strategies. Recurrence quantification analysis was used to build a compact description of trunk oscillations. Results and discussion: The results show that a thorough description can be easily developed using such methods; moreover, the obtained values highlight that, when designing strategies for human robot collaboration, ensuring that the subject maintains control of the rhythm of the task allows to maximize comfort in task execution, without affecting efficiency.

Walking ability of individuals fitted with transfemoral bone-anchored prostheses: A comparative study of gait parameters.

OBJECTIVE: This study presents the walking abilities of participants fitted with transfemoral bone-anchored prostheses using a total of 14 gait parameters. DESIGN: Two-centre retrospective cross-sectional comparative study. SETTING: Research facilities equipped with tridimensional motion capture systems. PARTICIPANTS: Two control arms included eight able-bodied participants arm (54 ± 9 years, 1.75 ± 0.07 m, 76 ± 7 kg) and nine participants fitted with transfemoral socket-suspended prostheses arm (59 ± 9 years, 1.73 ± 0.07 m, 80 ± 16 kg). The intervention arm included nine participants fitted with transfemoral bone-anchored prostheses arm (51 ± 13 years, 1.78 ± 0.09 m, 87.3 ± 16.1 kg). INTERVENTION: Fitting of transfemoral bone-anchored prostheses. MAIN MEASURES: Comparisons were performed for two spatio-temporal, three spatial and nine temporal gait parameters. RESULTS: The cadence and speed of walking were 107 ± 6 steps/min and 1.23 ± 0.19 m/s for the able-bodied participants arm, 88 ± 7 steps/min and 0.87 ± 0.17 m/s for the socket-suspended prosthesis arm, and 96 ± 6 steps/min and 1.03 ± 0.17 m/s for bone-anchored prosthesis arm, respectively. Able-bodied participants and bone-anchored prosthesis arms were comparable in age, height, and body mass index as well as cadence and speed of walking, but the able-bodied participant arm showed a swing phase 31% shorter. Bone-anchored and socket-suspended prostheses arms were comparable for age, height, mass, and body mass index as well as cadence and speed of walking, but the bone-anchored prosthesis arm showed a step width and duration of double support in seconds 65% and 41% shorter, respectively. CONCLUSIONS: Bone-anchored and socket-suspended prostheses restored equally well the gait parameters at a self-selected speed. This benchmark data provides new insights into the walking ability of individuals using transfemoral bionics bone-anchored prostheses.

Optimal Identification of Muscle Synergies From Typical Sit-to-Stand Clinical Tests.

Goal: The goal of this manuscript is to investigate the optimal methods for extracting muscle synergies from a sit-to-stand test; in particular, the performance in identifying the modular structures from signals of different length is characterized. Methods: Surface electromyography signals have been recorded from instrumented sit-to-stand trials. Muscle synergies have then been extracted from signals of different duration (i.e. 5 times sit to stand and 30 seconds sit to stand) from different portions of a complete sit-to-stand-to-sit cycle. Performance have then been characterized using cross-validation procedures. Moreover, an optimal method based on a modified Akaike Information Criterion measure is applied on the signal for selecting the correct number of synergies from each trial. Results: Results show that it is possible to identify correctly muscle synergies from relatively short signals in a sit-to-stand experiment. Moreover, the information about motor control structures is identified with a higher consistency when only the sit-to-stand phase of the complete cycle is considered. Conclusions: Defining a set of optimal methods for the extraction of muscle synergies from a clnical test such as the sit-to-stand is of key relevance to ensure the applicability of any synergy-related analysis in the clinical practice, without requiring knowledge of the technical signal processing methods and the underlying features of the signal.

Characterization of prosthetic knees through a low-dimensional description of gait kinematics.

The characterization of both limbs' behaviour in prosthetic gait is of key importance for improving the prosthetic components and increasing the biomechanical capability of trans-femoral amputees. When characterizing human gait, modular motor control theories have been proven to be powerful in providing a compact description of the gait patterns. In this paper, the planar covariation law of lower limb elevation angles is proposed as a compact, modular description of prosthetic gait; this model is exploited for a comparison between trans-femoral amputees walking with different prosthetic knees and control subjects walking at different speeds. Results show how the planar covariation law is maintained in prostheses users, with a similar spatial organization and few temporal differences. Most of the differences among the different prosthetic knees are found in the kinematic coordination patterns of the sound side. Moreover, different geometrical parameters have been calculated over the common projected plane, and their correlation with classical gait spatiotemporal and stability parameters has been investigated. The results from this latter analysis have highlighted a correlation with several parameters of gait, suggesting that this compact description of kinematics unravels a significant biomechanical meaning. These results can be exploited to guide the control mechanisms of prosthetic devices based purely on the measurement of relevant kinematic quantities.

Characterizing the Gait of People With Different Types of Amputation and Prosthetic Components Through Multimodal Measurements: A Methodological Perspective.

Prosthetic gait implies the use of compensatory motor strategies, including alterations in gait biomechanics and adaptations in the neural control mechanisms adopted by the central nervous system. Despite the constant technological advancements in prostheses design that led to a reduction in compensatory movements and an increased acceptance by the users, a deep comprehension of the numerous factors that influence prosthetic gait is still needed. The quantitative prosthetic gait analysis is an essential step in the development of new and ergonomic devices and to optimize the rehabilitation therapies. Nevertheless, the assessment of prosthetic gait is still carried out by a heterogeneous variety of methodologies, and this limits the comparison of results from different studies, complicating the definition of shared and well-accepted guidelines among clinicians, therapists, physicians, and engineers. This perspective article starts from the results of a project funded by the Italian Worker's Compensation Authority (INAIL) that led to the generation of an extended dataset of measurements involving kinematic, kinetic, and electrophysiological recordings in subjects with different types of amputation and prosthetic components. By encompassing different studies published along the project activities, we discuss the specific information that can be extracted by different kinds of measurements, and we here provide a methodological perspective related to multimodal prosthetic gait assessment, highlighting how, for designing improved prostheses and more effective therapies for patients, it is of critical importance to analyze movement neural control and its mechanical actuation as a whole, without limiting the focus to one specific aspect.

Classifying reaching height through muscle synergies in unconstrained scenarios.

Muscle synergy analysis has been widely adopted in the literature for the analysis of upper limb surface electromyographic signals during reaching tasks and for the prediction of movement direction for myoelectric control purposes. However, previous studies have characterized movements in constrained or semi-constrained scenarios, in which the subjects performing the movement were instructed to reach particular targets or were given some kind of feedback. In this work, the same synergy model has been applied to a completely unconstrained upper limb reaching experiment, with the aim of classifying the height of the target starting from the activity of the synergies. Results show that the synergistic model is able to extract compact features that can identify with good performance three different reaching heights. Moreover, this representation is able to isolate the signals that contain predictive information about the movement direction from the ones that are related to movement timing; this, together with the good performance of the synergy-based classifier supports the proposal of applying this model to the pre-processing of electromyographic signals when dealing with control systems that use signals from multiple muscles to predict movements.

The Influence of the sEMG Amplitude Estimation Technique on the EMG-Force Relationship.

The estimation of the sEMG-force relationship is an open problem in the scientific literature; current methods show different limitations and can achieve good performance only on limited scenarios, failing to identify a general solution to the optimization of this kind of analysis. In this work, this relationship has been estimated on two different datasets related to isometric force-tracking experiments by calculating the sEMG amplitude using different fixed-time constant moving-window filters, as well as an adaptive time-varying algorithm. Results show how the adaptive methods might be the most appropriate choice for the estimation of the correlation between the sEMG signal and the force time course. Moreover, the comparison between adaptive and standard filters highlights how the time constants exploited in the estimation strategy is not the only influence factor on this kind of analysis; a time-varying approach is able to constantly capture more information with respect to fixed stationary approaches with comparable window lengths.

An Objective, Information-Based Approach for Selecting the Number of Muscle Synergies to be Extracted via Non-Negative Matrix Factorization.

Muscle synergy analysis is a useful tool for the evaluation of the motor control strategies and for the quantification of motor performance. Among the parameters that can be extracted, most of the information is included in the rank of the modular control model (i.e. the number of muscle synergies that can be used to describe the overall muscle coordination). Even though different criteria have been proposed in literature, an objective criterion for the model order selection is needed to improve reliability and repeatability of MSA results. In this paper, we propose an Akaike Information Criterion (AIC)-based method for model order selection when extracting muscle synergies via the original Gaussian Non-Negative Matrix Factorization algorithm. The traditional AIC definition has been modified based on a correction of the likelihood term, which includes signal dependent noise on the neural commands, and a Discrete Wavelet decomposition method for the proper estimation of the number of degrees of freedom of the model, reduced on a synergy-by-synergy and event-by-event basis. We tested the performance of our method in comparison with the most widespread ones, proving that our criterion is able to yield good and stable performance in selecting the correct model order in simulated EMG data. We further evaluated the performance of our AIC-based technique on two distinct experimental datasets confirming the results obtained with the synthetic signals, with performances that are stable and independent from the nature of the analysed task, from the signal quality and from the subjective EMG pre-processing steps.

Smartphone-Based Answering to School Subject Questions Alters Gait in Young Digital Natives.

Smartphone texting while walking is a very common activity among people of different ages, with the so-called "digital natives" being the category most used to interacting with an electronic device during daily activities, mostly for texting purposes. Previous studies have shown how the concurrency of a smartphone-related task and walking can result in a worsening of stability and an increased risk of injuries for adults; an investigation of whether this effect can be identified also in people of a younger age can improve our understanding of the risks associated with this common activity. In this study, we recruited 29 young adolescents (12 ± 1 years) to test whether walking with a smartphone increases fall and injuries risk, and to quantify this effect. To do so, participants were asked to walk along a walkway, with and without the concurrent writing task on a smartphone; several different parameters linked to stability and risk of fall measures were then calculated from an inertial measurement unit and compared between conditions. Smartphone use determined a reduction of spatio-temporal parameters, including step length (from 0.64 ± 0.08 to 0.55 ± 0.06 m) and gait speed (1.23 ± 0.16 to 0.90 ± 0.16 m/s), and a general worsening of selected indicators of gait stability. This was found to be mostly independent from experience or frequency of use, suggesting that the presence of smartphone activities while walking may determine an increased risk of injury or falls also for a population that grew up being used to this concurrency.

Modular motor control of the sound limb in gait of people with trans-femoral amputation.

BACKGROUND: The above-knee amputation of a lower limb is a severe impairment that affects significantly the ability to walk; considering this, a complex adaptation strategy at the neuromuscular level is needed in order to be able to move safely with a prosthetic knee. In literature, it has been demonstrated that muscle activity during walking can be described via the activation of a small set of muscle synergies. The analysis of the composition and the time activation profiles of such synergies have been found to be a valid tool for the description of the motor control schemes in pathological subjects. METHODS: In this study, we used muscle synergy analysis techniques to characterize the differences in the modular motor control schemes between a population of 14 people with trans-femoral amputation and 12 healthy subjects walking at two different (slow and normal self-selected) speeds. Muscle synergies were extracted from a 12 lower-limb muscles sEMG recording via non-negative matrix factorization. Equivalence of the synergy vectors was quantified by a cross-validation procedure, while differences in terms of time activation coefficients were evaluated through the analysis of the activity in the different gait sub-phases. RESULTS: Four synergies were able to reconstruct the muscle activity in all subjects. The spatial component of the synergy vectors did not change in all the analysed populations, while differences were present in the activity during the sound limb's stance phase. Main features of people with trans-femoral amputation's muscle synergy recruitment are a prolonged activation of the module composed of calf muscles and an additional activity of the hamstrings' module before and after the prosthetic heel strike. CONCLUSIONS: Synergy-based results highlight how, although the complexity and the spatial organization of motor control schemes are the same found in healthy subjects, substantial differences are present in the synergies' recruitment of people with trans femoral amputation. In particular, the most critical task during the gait cycle is the weight transfer from the sound limb to the prosthetic one. Future studies will integrate these results with the dynamics of movement, aiming to a complete neuro-mechanical characterization of people with trans-femoral amputation's walking strategies that can be used to improve the rehabilitation therapies.

Wearable-based Temporal Parameters of Gait in Circuitous Routes under Dual-Task Conditions.

12 young adults were requested to walk along a circuitous path including turns, slaloms, stair ascending and descending, while wearing an inertial sensor placed on the back at the lumbar level. The path was completed under two conditions: with no additive cognitive task, and while performing a cognitive task and texting on a smartphone. Different temporal global parameters of gait were extracted from the inertial sensor data, to check for differences driven by the presence of the cognitive task. Regularity, durations, and temporal characteristics of gait resulted significantly affected from the presence of the additional task, and this effect was only in part due to a modification coming from the decrease in walking speed.

Comparison of Initialization Techniques for the Accurate Extraction of Muscle Synergies from Myoelectric Signals via Nonnegative Matrix Factorization.

The main goal of this work was to assess the performance of different initializations of matrix factorization algorithms for an accurate identification of muscle synergies. Currently, nonnegative matrix factorization (NNMF) is the most commonly used method to identify muscle synergies. However, it has been shown that NNMF performance might be affected by different kinds of initialization. The present study aims at optimizing the traditional NNMF initialization for data with partial or complete temporal dependencies. For this purpose, three different initializations are used: random, SVD-based, and sparse. NNMF was used to identify muscle synergies from simulated data as well as from experimental surface EMG signals. Simulated data were generated from synthetic independent and dependent synergy vectors (i.e., shared muscle components), whose activation coefficients were corrupted by simulating controlled degrees of correlation. Similarly, EMG data were artificially modified, making the extracted activation coefficients temporally dependent. By measuring the quality of identification of the original synergies underlying the data, it was possible to compare the performance of different initialization techniques. Simulation results demonstrate that sparse initialization performs significantly better than all other kinds of initialization in reconstructing muscle synergies, regardless of the correlation level in the data.

An automatic, adaptive, information-based algorithm for the extraction of the sEMG envelope.

Surface ElectroMyography (sEMG) is widely used as a non-invasive tool for the assessment of motor control strategies. However, the standardization of the methods used for the estimation of sEMG amplitude is a problem yet to be solved; in most cases, sEMG amplitude is estimated through the extraction of the envelope of the signal via different low-pass filtering procedures with fixed cut-off frequencies chosen by the experimenter. In this work, we have shown how it is not possible to find the optimal choice of the cut-off frequency without any a priori knowledge on the signal; considering this, we have proposed an updated version of an iterative adaptive algorithm already present in literature, aiming to completely automatize the sEMG amplitude estimation. We have compared our algorithm to most of the typical solutions (fixed window filters and the previous version of the adaptive algorithm) for the extraction of the sEMG envelope, showing how the proposed adaptive procedure significantly improves the quality of the estimation, with a lower fraction of variance unexplained by the extracted envelope for different simulated modulating waveforms (p < 0.005). The definition of an entropy-based convergence criterion has allowed for a complete automatization of the process. We infer that this algorithm can ensure repeatability of the estimation of the sEMG amplitude, due to its independence from the experimental choices, so allowing for a quantitative interpretation in a clinical environment.