Robot-assisted Rehabilitation Research Articles

Usability Assessment of a Cable-Driven Exoskeletal Robot for Hand Rehabilitation

Study design: Case series.Background: Robot-assisted rehabilitation mediated by exoskeletal devices is a popular topic of research. The biggest difficulty in the development of rehabilitation robots is the consideration of the clinical needs. This study investigated the usability of a novel cable-driven exoskeletal robot specifically designed for hand rehabilitation.Methods: The study consists of three steps, including prototype development, spasticity observation, and usability evaluation. First, we developed the prototype robot DexoHand to manipulate the patient's fingers based on the clinical needs and the cable-driven concept established in our previous work. Second, we applied DexoHand to patients with different levels of spasticity. Finally, we obtained the system usability scale (SUS) and assessed its usability.Results: Two healthy subjects were recruited in the pre-test, and 18 patients with stroke and four healthy subjects were recruited in the formal test for usability. The total SUS score obtained from the patients and healthy subjects was 94.77 ± 2.98 (n = 22), indicating an excellent level of usability. The satisfaction score was 4.74 ± 0.29 (n = 22), revealing high satisfaction with DexoHand. The tension profile measured by the cables showed the instantaneous force used to manipulate fingers among different muscle tone groups.Conclusions: DexoHand meets the clinical needs with excellent usability, satisfaction, and reliable tension force monitoring, yielding a feasible platform for robot-assisted hand rehabilitation.

Myoelectric control algorithm for robot-assisted therapy: a hardware-in-the-loop simulation study

BackgroundA direct blow to the knee is one way to injure the anterior cruciate ligament (ACL), e.g., during a football or traffic accident. Robot-assisted therapy (RAT) rehabilitation, simulating regular walking, improves walking and balance abilities, and extensor strength after ACL reconstruction. However, there is a need to perform RAT during other phases of ACL injury rehabilitation before attempting an advanced exercise such as walking. This paper aims to propose a myoelectric control (MEC) algorithm for a robot-assisted rehabilitation system, “Nukawa”, to assist knee movement during these types of exercises, i.e., such as in active-assisted extension exercises.MethodsSurface electromyography (sEMG) signal processing algorithm was developed to detect the motion intention of the knee joint. The sEMG signal processing algorithm and the movement control algorithm, reported by the authors in a previous publication, were joined together as a hardware-in-the-loop simulation to create and test the MEC algorithm, instead of using the actual robot.Experiments and resultsAn experimental protocol was conducted with 17 healthy subjects to acquire sEMG signals and their lower limb kinematics during 12 ACL rehabilitation exercises. The proposed motion intention algorithm detected the orientation of the intention 100% of the times for the extension and flexion exercises. Also, it detected in 94% and 59% of the cases the intensity of the movement intention in a comparable way to the maximum voluntary contraction (MVC) during extension exercises and flexion exercises, respectively. The maximum position mean absolute error was 0.1^{circ }, 6.3^{circ }, and 0.3^{circ } for the hip, knee, and ankle joints, respectively.ConclusionsThe MEC algorithm detected the intensity of the movement intention, approximately, in a comparable way to the MVC and the orientation. Moreover, it requires no prior training or additional torque sensors. Also, it controls the speed of the knee joint of Nukawa to assist the knee movement, i.e., such as in active-assisted extension exercises.

MyoTrack: Realtime Estimation of Subject Participation in Robotic Rehabilitation Using sEMG and IMU

In this paper, we present MyoTrack, a realtime classification procedure that identifies levels of subject participation during robot-assisted rehabilitation using a wearable surface electromyography (sEMG) Myo armband. We hypothesize, test, and prove that high sEMG correlates with high participation and vice versa. We then use Myo's inertial measurement unit to extract the subject's hand trajectory during the rehabilitation task. Comparing this hand trajectory with the ground-truth robot trajectory enables us to identify whether any high muscle activity corresponds to the active participation of the subject in robotic therapy or not (random gestures and motions). Since the robotic assistance implemented in this paper can autonomously complete the therapy task without any subject participation; it is crucial to identify the patient's participation level in realtime and develop a suitable intervention strategy. Using 15 healthy subjects, we demonstrate that the proposed methodology of combining sEMG activation and robot-hand trajectory matching is a reliable indicator of subject participation with a realtime accuracy of 91.45%. We also present a realtime application that uses MyoTrack in the back-end to identify the realtime participation of a subject and intervene accordingly.

RBFN-Based Adaptive Backstepping Sliding Mode Control of an Upper-Limb Exoskeleton With Dynamic Uncertainties

In recent decades, robot-assisted rehabilitation therapy has been widely researched and proven to be effective in the motor function recovery of disabled individuals. In this paper, an adaptive backstepping sliding mode control approach combined with neural uncertainty observer is developed for upper-limb exoskeleton, which can help the human operator perform repetitive rehabilitation training. Firstly, a comprehensive overview about the therapeutic exoskeleton hardware and real-time control system is introduced. Then, the neural adaptive backstepping sliding mode controller (NABSMC) is developed based on radial basis function network (RBFN) to improve the trajectory tracking accuracy with external disturbances and dynamics errors. Next, the closed-loop stability of the proposed controller is demonstrated according to the Lyapunov stability theory. Finally, further experimental investigation are conducted on three volunteers to compare the control performance of NABSMC strategy with an optimal backstepping sliding mode control (OBSMC) strategy. The comparison results show that the proposed NABSMC algorithm is capable of achieving higher trajectory tracking accuracy and better step response characteristic during repetitive passive rehabilitation training.

Adaptive Admittance Control of an Upper Extremity Rehabilitation Robot With Neural-Network-Based Disturbance Observer

Robot-assisted rehabilitation therapy has become an important technology applied to recover the motor functions of disabled individuals. In the present paper, an adaptive admittance control strategy combined with a neural-network-based disturbance observer (AACNDO) is developed for a therapeutic robot to provide upper extremity movement assistance. Firstly, a comprehensive overview of the robot hardware and real-time control system is introduced. Then, the dynamics-based adaptive admittance controller is designed to improve human-robot interaction compliance and induce the active participation of the patients during rehabilitation training. A disturbance observer with a radial basis function network is designed to guarantee the control performance with external uncertainties and dynamics error. Besides, an adaption law is integrated into the admittance model to adjust the interaction compliance in different working areas based on the motion intention and recovery phase of the patient. Further experimental investigations, including sinusoidal trajectory tracking experiments, circular trajectory tracking experiments with admittance adjustment, and intention-based resistive training experiments, are conducted by three volunteers. Finally, the experimental results validate the feasibility and effectiveness of the rehabilitation robot and AACNDO scheme in providing patient-passive and patient-cooperative rehabilitation training.

An Industrial Robot-Based Rehabilitation System for Bilateral Exercises

Robot-assisted rehabilitation devices can provide intensive and precise task-based training that differs from clinician-facilitated manual therapy. However, industrial robots are still rarely used in rehabilitation, especially in bilateral exercises. The main purpose of this research is to develop and evaluate the functionality of a bilateral upper-limb rehabilitation system based on two modern industrial robots. A `patient-cooperative' control strategy is developed based on an adaptive admittance controller, which can take into account patients' voluntary efforts. Three bilateral training protocols (passive, active, and self) are also proposed based on the system and the control strategy. Experimental results from 10 healthy subjects show that the proposed system can provide reliable bilateral exercises: the mean RMS values for the master error and the master-slave error are all less than 1.00 mm and 1.15 mm respectively, and the mean max absolute values for the master error and the master-slave error are no greater than 6.11 mm and 6.73 mm respectively. Meanwhile, the experimental results also confirm that the recalculated desired trajectory can present the voluntary efforts of subjects. These experimental findings suggest that industrial robots can be used in bilateral rehabilitation training, and also highlight the potential applications of the proposed system in further clinical practices.

Models for temporal-spatial parameters in walking with cadence ratio as the independent variable

Accurate models that describe temporal-spatial parameters are desirable in gait estimation and rehabilitation. This study aimed to explore simple but relatively accurate models to describe stride length (SL), speed (SP) and walk ratio (WR) at various cadences. Twenty-four able-bodied participants (16 in a test group and 8 in a validation group) walked at seven cadence ratios (CRs). The individual and group mean SL, SP and WR were studied. Suitable temporal-spatial model structures were proposed and used to approximate the individual SL, SP and WR at various CRs. After the temporal-spatial model structures were found to be feasible, the general temporal-spatial models were analysed using the test group mean SL, SP and WR. Accuracy was assessed using the validation group mean values. Individual approximation accuracies showed that the proposed model structure deduced from the linear SL model was suitable for WR approximation. The linear, deduced quadratic and power functions approximated the individual SL, SP and WR, respectively, with high accuracy. Based on the test group mean SL, SP and WR, the general temporal-spatial models were obtained and produced comparable approximation accuracies in the validation group. The general temporal-spatial models predicted well the individual gait parameters with similar individual errors for both groups. These temporal-spatial models clearly describe SL, SP and especially WR at various cadences. They provide accurate reference data for gait estimation and have potential to guide speed modulation in robot-assisted gait rehabilitation.Graphical abstractTwenty-four able-bodied participants (16 in test group and 8 in validation group) walked at seven cadence ratios (CRs), with the individual and group mean stride length (SL), speed (SP) and walk ratio (WR) studied. This work selected the cadence ratio as the independent variable and yielded general temporal-spatial models based on the test group data, which were a linear model for SL, a quadratic function for SP and a power function for WR. The general temporal-spatial model produced comparable approximation accuracies in the validation group. Clearly describing SL, SP and especially WR at various cadences, these temporal-spatial models provide accurate references for gait estimation and have the potential to guide speed modulation in robot-assisted gait rehabilitation. Approximation of the group mean temporal-spatial parameters at seven cadences. Solid lines in parts (a, b): the general linear SL model. Solid lines in (c, d): the general quadratic SP model. Solid lines in (e, f): the general WR model. Dots and stars in (a, c, e): the individual and group mean values for the test group. Dots and stars in (b, d, f): the individual and group mean values for the validation group.

A Multiparameter Approach to Evaluate Post-Stroke Patients: An Application on Robotic Rehabilitation

Multidomain instrumental evaluation of post-stroke chronic patients, coupled with standard clinical assessments, has rarely been exploited in the literature. Such an approach may be valuable to provide comprehensive insight regarding patients’ status, as well as orienting the rehabilitation therapies. Therefore, we propose a multidomain analysis including clinically compliant methods as electroencephalography (EEG), electromyography (EMG), kinematics, and clinical scales. The framework of upper-limb robot-assisted rehabilitation is selected as a challenging and promising scenario to test the multi-parameter evaluation, with the aim to assess whether and in which domains modifications may take place. Instrumental recordings and clinical scales were administered before and after a month of intensive robotic therapy of the impaired upper limb, on five post-stroke chronic hemiparetic patients. After therapy, all patients showed clinical improvement and presented pre/post modifications in one or several of the other domains as well. All patients performed the motor task in a smoother way; two of them appeared to change their muscle synergies activation strategies, and most subjects showed variations in their brain activity, both in the ipsi- and contralateral hemispheres. Changes highlighted by the new multiparametric instrumental approach suggest a recovery trend in agreement with clinical scales. In addition, by jointly demonstrating lateralization of brain activations, changes in muscle recruitment and the execution of smoother trajectories, the new approach may help distinguish between true functional recovery and the adoption of suboptimal compensatory strategies. In the light of these premises, the multi-domain approach may allow a finer patient characterization, providing a deeper insight into the mechanisms underlying the relearning procedure and the level (neuro/muscular) at which it occurred, at a relatively low expenditure. The role of this quantitative description in defining a personalized treatment strategy is of great interest and should be addressed in future studies.

Force feasible set prediction with artificial neural network and musculoskeletal model

Developing tools to predict the force capabilities of the human limbs through the Force Feasible Set (FFS) may be of great interest for robotic assisted rehabilitation and digital human modelling for ergonomics. Indeed, it could help to refine rehabilitation programs for active participation during exercise therapy and to prevent musculoskeletal disorders. In this framework, the purpose of this study is to use artificial neural networks (ANN) to predict the FFS of the upper-limb based on joint centre Cartesian positions and anthropometric data. Seventeen right upper-limb musculoskeletal models based on individual anthropometric data are created. For each musculoskeletal model, the FFS is computed for 8428 different postures. For any combination of force direction and joint positions, ANNs can predict the FFS with high values of coefficient of determination (R2 > 0.89) between the true and predicted data.

Robotic Assistance for Upper Limbs May Induce Slight Changes in Motor Modules Compared With Free Movements in Stroke Survivors: A Cluster-Based Muscle Synergy Analysis.

Background: The efficacy of robot-assisted rehabilitation as a technique for achieving motor recovery is still being debated. The effects of robotic assistance are generally measured using standard clinical assessments. Few studies have investigated the value of human-centered instrumental analysis, taking the modular organization of the human neuromotor system into account in assessing how stroke survivors interact with robotic set-ups. In this paper, muscle synergy analysis was coupled with clustering procedures to elucidate the effect of human-robot interaction on the spatial and temporal features, and directional tuning of motor modules during robot-assisted movements.Methods: Twenty-two stroke survivors completed a session comprising a series of hand-to-mouth movements with and without robotic assistance. Patients were assessed instrumentally, recording kinematic, and electromyographic data to extract spatial muscle synergies and their temporal components. Patients' spatial synergies were grouped by means of a cluster analysis, matched pairwise across conditions (free and robot-assisted movement), and compared in terms of their spatial and temporal features, and directional tuning, to examine how robotic assistance altered their motor modules.Results: Motor synergies were successfully extracted for all 22 patients in both conditions. Seven clusters (spatial synergies) could describe the original datasets, in both free and robot-assisted movements. Interacting with the robot slightly altered the spatial synergies' features (to a variable extent), as well as their temporal components and directional tuning.Conclusions: Slight differences were identified in the characteristics of spatial synergies, temporal components and directional tuning of the motor modules of stroke survivors engaging in free and robot-assisted movements. Such effects are worth investigating in the framework of a modular description of the neuromusculoskeletal system to shed more light on human-robot interaction, and the effects of robotic assistance and rehabilitation.

Wrist Robot-assisted Rehabilitation Treatment in Subacute and Chronic Stroke Patients: from Distal to Proximal Motor Recovery.

In this study, the recovery of proximal and distal segments in stroke patients who received distal training alone was investigated. Forty (20 subacute and 20 chronic) stroke patients were recruited to perform wrist robot-assisted rehabilitation training. The upper extremity, shoulder-elbow and wrist subsections of the Fugl-Meyer Assessment Scale were used to assess the motor recovery of distal and proximal segments. In addition, the Modified Ashworth Scale, the Motricity Index and the Box & Block test were used as clinical outcome measures together with kinematic parameters to evaluate the effects of the training. Significant increases in the wrist and shoulder-elbow subsections of the Fugl-Meyer Assessment Scale, Motricity Index and Box & Block test were found in both groups. Average changes in shoulder-elbow and upper extremity subsections of the Fugl-Meyer Assessment Scale in the subacute group (6.10 ± 6.60 and 15.65 ± 14.04) were significantly higher (p < 0.05) than those in the chronic group (2.30 ± 2.76 and 6.60 ± 4.64). In addition, significant increases in the movement velocity, movement smoothness and movement quality were observed in the subacute group. Our findings provide evidence that following a robot-assisted rehabilitation treatment there is a distal-to-proximal generalization in subacute stroke patients.

An Orthopaedic Robotic-Assisted Rehabilitation Method of the Forearm in Virtual Reality Physiotherapy.

The use of robotic rehabilitation in orthopaedics has been briefly explored. Despite its possible advantages, the use of computer-assisted physiotherapy of patients with musculoskeletal injuries has received little attention. In this paper, we detailed the development and evaluation of a robotic-assisted rehabilitation system as a new methodology of assisted physiotherapy in orthopaedics. The proposal consists of an enhanced end-effector haptic interface mounted in a passive mechanism for allowing patients to perform upper-limb exercising and integrates virtual reality games conceived explicitly for assisting the treatment of the forearm after injuries at the wrist or elbow joints. The present methodology represents a new approach to assisted physiotherapy for strength and motion recovery of wrist pronation/supination and elbow flexion-extension movements. We design specific game scenarios enriched by proprioceptive and haptic force feedback in three training modes: passive, active, and assisted exercising. The system allows the therapist to tailor the difficulty level on the observed motion capacity of the patients and the kinesiology measurements provided by the system itself. We evaluated the system through the analysis of the muscular activity of two healthy subjects, showing that the system can assign significant working loads during typical physiotherapy treatment profiles. Subsequently, a group of ten patients undergoing manual orthopaedic rehabilitation of the forearm tested the system, under similar conditions at variable intensities. Patients tolerated changes in difficulty through the tests, and they expressed a favourable opinion of the system through the administered questionnaires, which indicates that the system was well accepted and that the proposed methodology was feasible for the case study for subsequently controlled trials. Finally, a predictive model of the performance score in the form of a linear combination of kinesiology observations was implemented in function of difficult training parameters, as a way of systematically individualising the training during the therapy, for subsequent studies.

Exoskeleton-Robot Assisted Therapy in Stroke Patients: A Lesion Mapping Study.

Background: Technology-supported rehabilitation is emerging as a solution to support therapists in providing a high-intensity, repetitive and task-specific treatment, aimed at improving stroke recovery. End-effector robotic devices are known to positively affect the recovery of arm functions, however there is a lack of evidence regarding exoskeletons. This paper evaluates the impact of cerebral lesion load on the response to a validated robotic-assisted rehabilitation protocol.Methods: Fourteen hemiparetic patients were assessed in a within-subject design (age 66.9 ± 11.3 years; 10 men and 4 women). Patients, in post-acute phase, underwent 7 weeks of bilateral arm training assisted by an exoskeleton robot combined with a conventional treatment (consisting of simple physical activity together with occupational therapy). Clinical and neuroimaging evaluations were performed immediately before and after rehabilitation treatments. Fugl-Meyer (FM) and Motricity Index (MI) were selected to measure primary outcomes, i.e., motor function and strength. Functional independance measure (FIM) and Barthel Index were selected to measure secondary outcomes, i.e., daily living activities. Voxel-based lesion symptom mapping (VLSM) was used to determine the degree of cerebral lesions associated with motor recovery.Results: Robot-assisted rehabilitation was effective in improving upper limb motor function recovery, considering both primary and secondary outcomes. VLSM detected that lesion load in the superior region of the corona radiata, internal capsule and putamen were significantly associated with recovery of the upper limb as defined by the FM scores (p-level < 0.01).Conclusions: The probability of functional recovery from stroke by means of exoskeleton robotic rehabilitation relies on the integrity of specific subcortical regions involved in the primary motor pathway. This is consistent with previous evidence obtained with conventional neurorehabilitation approaches.

Effects of gait training with using HAL® on gait functions of chronic stroke patients

Recently, the robot-assisted rehabilitation has emerged as a promising new technology for patients with severe lower limb impairment The effect of rehabilitation including the concepts of repetitive, intensive, task-orientated training, which has been shown to be effective for chronic stroke patients, may be facilitated by using robotics in improving physical functions of patients The aim of the present study is to investigate whether gait training using a robot suit HAL ® (Hybrid Assistive Limb: an exoskeleton robot that enhances the limb motion of the human body by detecting the weak bioelectrical nerve signals) would improve gait functions in chronic stroke patients. Fourteen patients with chronic stroke in Kyoto University Hospital were participated in the study They received 8 gait training sessions (60 min/session) using HAL® of double leg type during 3 weeks Before and after intervention, gait speed (m/s), stride length (m), cadence (step/min), and 2-minute walk distance (m) were measured Paired t test was used to determine the differences in outcome measurements between before and after intervention Spearman's rank correlated coefficients were calculated between change percentage of gait speed and change percentage of gait parameter (stride length or cadence). Gait speed (0.56 ± 0.32 → 0.71 ± 0.41 m/s), stride length (0.72 ± 0.32 → 0.82 ± 0.35 m), cadence (90.0 ± 29.4 → 98.7 ± 33.2 step/min), and 2-minute walk distance (62.2 ± 34.7 → 77.9 ± 45.2 m) increased significantly after the intervention The change percentage of gait speed was significantly correlated with the change percentage of stride length ( r = 0.75), but not with cadence. The results of the present study demonstrated that gait training using HAL ® was an effective intervention, especially in improving spatiotemporal gait parameters and gait distance in chronic stroke patients.

Upper limb robot-assisted rehabilitation in sub-acute stroke: Slacking or engagement promotion? Evidence from patient-robot interactions analysis

After stroke, interventions based on active repeated movements stimulate neuroplasticity and enhances upper limb motor outcomes. Robotic devices using interactive and adaptive programs offer the ability to perform a large number of movements, including in patients with severe paresis. However, some authors argued that assist-as-needed algorithms would not ensure optimal patient's participation. This retrospective study analyzed robot-patient interactions by measuring motor performance and robotic assistance parameters during robotic sessions. Sixteen patients (age 51 ± 15) with upper limb paresis due to stroke carried out a 16-session robot-assisted training combined with usual care during the sub-acute phase (time from stroke, 43 ± 15 days). The robotic program (5 days a week) comprised at least 2 consecutive series (S1 and S2) of 320 assisted reaching movements performed in the horizontal plane. Time of series completion, active-assisted motor performance (range of motion, accuracy, smoothness) and robotic assistance parameters during movements’ execution were analysed between S1 and S2 every 4 sessions during the training program. Time of 320-movement series completion, robotic assistance and motor performance were not different between S1 and S2 over the training ( Table 1 ). The results indicated that a highly repetitive robot-assisted training for the upper limb can be associated with both sustained active participation and motor performance during robotic sessions. Robotic algorithms might optimize patient's active engagement and motor performance in spite of an unparalleled number of repetitions compared to usual care.

Translation of robot-assisted rehabilitation to clinical service: a comparison of the rehabilitation effectiveness of EMG-driven robot hand assisted upper limb training in practical clinical service and in clinical trial with laboratory configuration for chronic stroke

BackgroundRehabilitation robots can provide intensive physical training after stroke. However, variations of the rehabilitation effects in translation from well-controlled research studies to clinical services have not been well evaluated yet. This study aims to compare the rehabilitation effects of the upper limb training by an electromyography (EMG)-driven robotic hand achieved in a well-controlled research environment and in a practical clinical service.MethodsIt was a non-randomized controlled trial, and thirty-two participants with chronic stroke were recruited either in the clinical service (n = 16, clinic group), or in the research setting (n = 16, lab group). Each participant received 20-session EMG-driven robotic hand assisted upper limb training. The training frequency (4 sessions/week) and the pace in a session were fixed for the lab group, while they were flexible (1–3 sessions/week) and adaptive for the clinic group. The training effects were evaluated before and after the treatment with clinical scores of the Fugl-Meyer Assessment (FMA), Action Research Arm Test (ARAT), Functional Independence Measure (FIM), and Modified Ashworth Scale (MAS).ResultsSignificant improvements in the FMA full score, shoulder/elbow and wrist/hand (P < 0.001), ARAT (P < 0.001), and MAS elbow (P < 0.05) were observed after the training for both groups. Significant improvements in the FIM (P < 0.05), MAS wrist (P < 0.001) and MAS hand (P < 0.05) were only obtained after the training in the clinic group. Compared with the lab group, higher FIM improvement in the clinic group was observed (P < 0.05).ConclusionsThe functional improvements after the robotic hand training in the clinical service were comparable to the effectiveness achieved in the research setting, through flexible training schedules even with a lower training frequency every week. Higher independence in the daily living and a more effective release in muscle tones were achieved in the clinic group than the lab group.

Development of an RBFN-based neural-fuzzy adaptive control strategy for an upper limb rehabilitation exoskeleton

The patients of paralysis with motion impairment problems require extensive rehabilitation programs to regain motor functions. The great labor intensity and limited therapeutic effect of traditional human-based manual treatment have recently boosted the development of robot-assisted rehabilitation therapy. In the present work, a neural-fuzzy adaptive controller (NFAC) based on radial basis function network (RBFN) is developed for a rehabilitation exoskeleton to provide human arm movement assistance. A comprehensive overview is presented to describe the mechanical structure and electrical real-time control system of the therapeutic robot, which provides seven actuated degrees of freedom (DOFs) and achieves natural ranges of upper extremity movement. For the purpose of supporting the disable patients to perform repetitive passive rehabilitation training, the RBFN-based NFAC algorithm is proposed to guarantee trajectory tracking accuracy with parametric uncertainties and environmental disturbances. The stability of the proposed control scheme is demonstrated through Lyapunov stability theory. Further experimental investigation, involving the position tracking experiment and the frequency response experiment, are conducted to compare the control performance of the proposed method to those of cascaded proportional-integral-derivative controller (CPID) and fuzzy sliding mode controller (FSMC). The comparison results indicate that the proposed RBFN-based NFAC algorithm is capable of obtaining lower position tracking error and better frequency response characteristic.

Effectiveness of a single session of dual-transcranial direct current stimulation in combination with upper limb robotic-assisted rehabilitation in chronic stroke patients: a randomized, double-blind, cross-over study.

The impact of transcranial direct current stimulation (tDCS) is controversial in the neurorehabilitation literature. It has been suggested that tDCS should be combined with other therapy to improve their efficacy. To assess the effectiveness of a single session of upper limb robotic-assisted therapy (RAT) combined with real or sham-tDCS in chronic stroke patients. Twenty-one hemiparetic chronic stroke patients were included in a randomized, controlled, cross-over double-blind study. Each patient underwent two sessions 7 days apart in a randomized order: (a) 20 min of real dual-tDCS associated with RAT (REAL+RAT) and (b) 20 min of sham dual-tDCS associated with RAT (SHAM+RAT). Patient dexterity (Box and Block and Purdue Pegboard tests) and upper limb kinematics were evaluated before and just after each intervention. The assistance provided by the robot during the intervention was also recorded. Gross manual dexterity (1.8±0.7 blocks, P=0.008) and straightness of movement (0.01±0.03, P<0.05) improved slightly after REAL+RAT compared with before the intervention. There was no improvement after SHAM+RAT. The post-hoc analyses did not indicate any difference between interventions: REAL+RAT and SHAM+RAT (P>0.05). The assistance provided by the robot was similar during both interventions (P>0.05). The results showed a slight improvement in hand dexterity and arm movement after the REAL+RAT tDCS intervention. The observed effect after a single session was small and not clinically relevant. Repetitive sessions could increase the benefits of this combined approach.

Myoelectric Pattern Recognition for Controlling a Robotic Hand: A Feasibility Study in Stroke.

Myoelectric pattern recognition has been successfully applied as a human-machine interface to control robotic devices such as prostheses and exoskeletons, significantly improving the dexterity of myoelectric control. This study investigates the feasibility of applying myoelectric pattern recognition for controlling a robotic hand in stroke patients. Myoelectric pattern recognition of six hand motion patterns was performed using forearm electromyogram signals in paretic side of eight stroke subjects. Both the random cross validation (RCV) and the chronological handout validation (CHV) were applied to assess the offline myoelectric pattern recognition performance. Experiments on real-time myoelectric pattern recognition control of an exoskeleton robotic hand were also performed. An average classification accuracy of 84.1% (the mean value from two different classifiers) and individual subject differences were observed in the offline myoelectric pattern recognition analysis using the RCV, while the accuracy decreased to 65.7% when the CHV was used. The stroke subjects achieved an average accuracy of 61.3 ± 20.9% for controlling the robotic hand. However, our study did not reveal a clear correlation between the real-time control accuracy and the offline myoelectric pattern recognition performance, or any specific characteristics of the stroke subjects. The findings suggest that it is feasible to apply myoelectric pattern recognition to control the robotic hand in some but not all of the stroke patients. Each stroke subject should be individually online tested for the feasibility of applying myoelectric pattern recognition control for robot-assisted rehabilitation.

Clustering of Directions Improves Goodness of Fit in Kinematic Data Collected in the Transverse Plane During Robot-Assisted Rehabilitation of Stroke Patients.

The kinematic character of hand trajectory in reaching tasks varies by movement direction. Often, direction is not included as a factor in the analysis of data collected during multi-directional reach tasks; consequently, this directionally insensitive model (DI) may be prone to type-II error due to unexplained variance. On the other hand, directionally specific models (DS) that account separately for each movement direction, may reduce statistical power by increasing the amount of data groupings. We propose a clustered-by-similarity (CS) in which movement directions with similar kinematic features are grouped together, maximizing model fit by decreasing unexplained variance while also decreasing uninformative sub-groupings. We tested model quality in measuring change over time in 10 kinematic features extracted from 72 chronic stroke patients participating in the VA-ROBOTICS trial, performing a targeted reaching task over 16 movement directions (8 targets, back- and forth from center) in the horizontal plane. Across 49 participants surviving a quality control sieve, 4.3 ± 1.1 (min: 3; max: 7) clusters were found among the 16 movement directions; clusters varied between participants. Among 49 participants, and averaged across 10 features, the better-fitting model for predicting change in features was found to be CS assessed by the Akaike Information criterion (61.6 ± 7.3%), versus DS (31.0 ± 7.8%) and DI (7.1 ± 7.1%). Confirmatory analysis via Extra Sum of Squares F-test showed the DS and CS models out-performed the DI model in head-to-head (pairwise) comparison in >85% of all specimens. Thus, we find overwhelming evidence that it is necessary to adjust for direction in the models of multi-directional movements, and that clustering kinematic data by feature similarly may yield the optimal configuration for this co-variate.