The road forward for upper-extremity rehabilitation robotics

NSF DARE-Transforming modeling in neurorehabilitation: Four threads for catalyzing progress

We present an overview of the Conference on Transformative Opportunities for Modeling in Neurorehabilitation held in March 2023. It was supported by the Disability and Rehabilitation Engineering (DARE) program from the National Science Foundation's Engineering Biology and Health Cluster. The conference brought together experts and trainees from around the world to discuss critical questions, challenges, and opportunities at the intersection of computational modeling and neurorehabilitation to understand, optimize, and improve clinical translation of neurorehabilitation. We organized the conference around four key, relevant, and promising Focus Areas for modeling: Adaptation & Plasticity, Personalization, Human-Device Interactions, and Modeling 'In-the-Wild'. We identified four common threads across the Focus Areas that, if addressed, can catalyze progress in the short, medium, and long terms. These were: (i) the need to capture and curate appropriate and useful data necessary to develop, validate, and deploy useful computational models (ii) the need to create multi-scale models that span the personalization spectrum from individuals to populations, and from cellular to behavioral levels (iii) the need for algorithms that extract as much information from available data, while requiring as little data as possible from each client (iv) the insistence on leveraging readily available sensors and data systems to push model-driven treatments from the lab, and into the clinic, home, workplace, and community. The conference archive can be found at (dare2023.usc.edu). These topics are also extended by three perspective papers prepared by trainees and junior faculty, clinician researchers, and federal funding agency representatives who attended the conference.

Simulation of a control method for active kinesiotherapy with an upper extremity rehabilitation exoskeleton without force sensor

Exoskeleton-aided active rehabilitation is a process that requires sensing and acting upon the motion intentions of the user. Typically, force sensors are used for this. However, they increase the weight and cost of these wearable devices. This paper presents the methodology for detecting users' intentions only with encoders integrated with the drives. It is unique compared to other algorithms, as enables active kinesiotherapy while adding no sensory systems. The method is based on comparing the measured motion with the one computed with the idealised model of the multibody system. The investigation assesses the method's performance and its robustness to model and measurement inaccuracies, as well as patients' unintended motions. Moreover, the PID parameters are selected to provide the optimal regulation based on the dynamics requirements. The research proves the presented concept of the control approach. For all the tests with the final settings, the system reacts to a change in the user's intention below one second and minimises the changes in proportion between the system's acceleration and the generated user's joint torque. The results are comparable to those obtained by EMG-based systems and significantly better than low-cost force sensors.

Initial Testing of Robotic Exoskeleton Hand Device for Stroke Rehabilitation

The preliminary test results of a novel robotic hand rehabilitation device aimed at treatment for the loss of motor abilities in the fingers and thumb due to stroke are presented. This device has been developed in collaboration with physiotherapists who regularly treat individuals who have suffered from a stroke. The device was tested on healthy adults to ensure comfort, user accessibility, and repeatability for various hand sizes in preparation for obtaining permission from regulatory bodies and implementing the design in a full clinical trial. Trials were conducted with 52 healthy individuals ranging in age from 19 to 93 with an average age of 58. A comfort survey and force data ANOVA were performed to measure hand motions and ensure the repeatability and accessibility of the system. Readings from the force sensor (p < 0.05) showed no significant difference between repetitions for each participant. All subjects considered the device comfortable. The device scored a mean comfort value of 8.5/10 on all comfort surveys and received the approval of all physiotherapists involved. The device has satisfied all design specifications, and the positive results of the participants suggest that it can be considered safe and reliable. It can therefore be moved forward for clinical trials with post-stroke users.

Kinematic coordinations capture learning during human-exoskeleton interaction

Human-exoskeleton interactions have the potential to bring about changes in human behavior for physical rehabilitation or skill augmentation. Despite significant advances in the design and control of these robots, their application to human training remains limited. The key obstacles to the design of such training paradigms are the prediction of human-exoskeleton interaction effects and the selection of interaction control to affect human behavior. In this article, we present a method to elucidate behavioral changes in the human-exoskeleton system and identify expert behaviors correlated with a task goal. Specifically, we observe the joint coordinations of the robot, also referred to as kinematic coordination behaviors, that emerge from human-exoskeleton interaction during learning. We demonstrate the use of kinematic coordination behaviors with two task domains through a set of three human-subject studies. We find that participants (1) learn novel tasks within the exoskeleton environment, (2) demonstrate similarity of coordination during successful movements within participants, (3) learn to leverage these coordination behaviors to maximize success within participants, and (4) tend to converge to similar coordinations for a given task strategy across participants. At a high level, we identify task-specific joint coordinations that are used by different experts for a given task goal. These coordinations can be quantified by observing experts and the similarity to these coordinations can act as a measure of learning over the course of training for novices. The observed expert coordinations may further be used in the design of adaptive robot interactions aimed at teaching a participant the expert behaviors.

Technology Acceptance Model for Exoskeletons for Rehabilitation of the Upper Limbs from Therapists' Perspectives

Over the last few years, exoskeletons have been demonstrated to be useful tools for supporting the execution of neuromotor rehabilitation sessions. However, they are still not very present in hospitals. Therapists tend to be wary of this type of technology, thus reducing its acceptability and, therefore, its everyday use in clinical practice. The work presented in this paper investigates a novel point of view that is different from that of patients, which is normally what is considered for similar analyses. Through the realization of a technology acceptance model, we investigate the factors that influence the acceptability level of exoskeletons for rehabilitation of the upper limbs from therapists' perspectives. We analyzed the data collected from a pool of 55 physiotherapists and physiatrists through the distribution of a questionnaire. Pearson's correlation and multiple linear regression were used for the analysis. The relations between the variables of interest were also investigated depending on participants' age and experience with technology. The model built from these data demonstrated that the perceived usefulness of a robotic system, in terms of time and effort savings, was the first factor influencing therapists' willingness to use it. Physiotherapists' perception of the importance of interacting with an exoskeleton when carrying out an enhanced therapy session increased if survey participants already had experience with this type of rehabilitation technology, while their distrust and the consideration of others' opinions decreased. The conclusions drawn from our analyses show that we need to invest in making this technology better known to the public-in terms of education and training-if we aim to make exoskeletons genuinely accepted and usable by therapists. In addition, integrating exoskeletons with multi-sensor feedback systems would help provide comprehensive information about the patients' condition and progress. This can help overcome the gap that a robot creates between a therapist and the patient's human body, reducing the fear that specialists have of this technology, and this can demonstrate exoskeletons' utility, thus increasing their perceived level of usefulness.

Impact of Gravity Compensation on Upper Extremity Movements in Harmony Exoskeleton

Robots have been used to offset the limb weight through gravity compensation in upper body rehabilitation to delineate the effects of loss of strength and loss of dexterity, which are two common forms of post-stroke impairments. In this paper, we explored the impact of this anti-gravity support on the quality of movement during reaching and coordinated arm movements in a pilot study with two participants with chronic stroke. The subjects donned the Harmony exoskeleton which supported proper shoulder coordination in addition to providing gravity compensation. Participants had previously taken part in seven one-hour sessions with the Harmony exoskeleton, performing six sets of passive-stretching and active exercises. Pre- and post-training sessions included assessments of two separate tasks, planar reaching and a set of six coordinated arm movements, in two conditions, outside of and supported by the exoskeleton. The movements were recorded using an optical motion capture system and analyzed using spectral arc length (SPARC) and straight line deviation to quantify movement smoothness and quality. We observed that gravity compensation resulted in an increased smoothness for the subject with high level of impairment whereas compensation resulted in a reduction in smoothness for the subject with low level of impairment in the reaching task. Both participants demonstrated better coordination of the shoulder-arm joint with gravity compensation. This result motivates further studies into the role of gravity compensation during coordinated movement training and rehabilitation interventions.

Comparison of Machine Learning Techniques for Activities of Daily Living Classification with Electromyographic Data

Advances in data science and wearable robotic devices present an opportunity to improve rehabilitation outcomes. Some of these devices incorporate electromyography (EMG) electrodes that sense physiological patient activity, making it possible to develop rehabilitation systems able to assess the patient's progress when performing activities of daily living (ADLs). However, additional research is needed to improve the ability to interpret EMG signals. To address this issue, an off-line classification approach for the 26 upper-limb ADLs included in the KIN-MUS UJI dataset is presented in this paper. The ADLs were performed by 22 subjects, while seven EMG signals were recorded from their forearms. From variable-length EMG time windows, 18 features were computed, and 13 features more were extracted from frequency domain windows. The classification performance of five different machine learning techniques, including Support Vector Machines (SVM), Convolutional Neural Networks (CNN), Gated Recurrent Unit (GRU) network, XGBoost, and Random Forests, were compared. CNN performed best amongst individual models, with an accuracy above 80%, compared to SVM with 77%, GRU with 73.9%, and the tree-based models below 64%. Ensemble learning with four CNN models achieved an even higher accuracy of 86%. These results suggest that the CNN ensemble model is capable of classifying EMG signals for most ADLs, which could be used in off-line quantitative assessment of robotic rehabilitation outcomes.