Wearable Sensor-Based Real-Time Gait Detection: A Systematic Review.

Gait analysis has traditionally been carried out in a laboratory environment using expensive equipment, but, recently, reliable, affordable, and wearable sensors have enabled integration into clinical applications as well as use during activities of daily living. Real-time gait analysis is key to the development of gait rehabilitation techniques and assistive devices such as neuroprostheses. This article presents a systematic review of wearable sensors and techniques used in real-time gait analysis, and their application to pathological gait. From four major scientific databases, we identified 1262 articles of which 113 were analyzed in full-text. We found that heel strike and toe off are the most sought-after gait events. Inertial measurement units (IMU) are the most widely used wearable sensors and the shank and foot are the preferred placements. Insole pressure sensors are the most common sensors for ground-truth validation for IMU-based gait detection. Rule-based techniques relying on threshold or peak detection are the most widely used gait detection method. The heterogeneity of evaluation criteria prevented quantitative performance comparison of all methods. Although most studies predicted that the proposed methods would work on pathological gait, less than one third were validated on such data. Clinical applications of gait detection algorithms were considered, and we recommend a combination of IMU and rule-based methods as an optimal solution.

Increasing motivation in robot-aided arm rehabilitation with competitive and cooperative gameplay.

BACKGROUND: Several strategies have been proposed to improve patient motivation and exercise intensity during robot-aided stroke rehabilitation. One relatively unexplored possibility is two-player gameplay, allowing subjects to compete or cooperate with each other to achieve a common goal. In order to explore the potential of such games, we designed a two-player game played using two ARMin arm rehabilitation robots. METHODS: The game was an air-hockey task displayed on a computer monitor and controlled using shoulder movements in the ARMin robot. Three game modes were tested: single-player (competing against computer), competitive (competing against human), and cooperative (cooperating with human against computer). All modes were played by 30 unimpaired subjects and 8 impaired chronic stroke subjects. The subjects filled out the Intrinsic Motivation Inventory questionnaire after each game mode, as well as a final questionnaire about game preferences and their personality. RESULTS: Nearly all unimpaired subjects preferred playing the two-player game modes to the single-player one, as they enjoyed talking and interacting with another person. However, there were two distinct player groups: one liked the competitive mode but not the cooperative mode while the other liked the cooperative but not the competitive mode. Unimpaired subjects who liked the competitive mode also put significantly more effort into it than into the other modes. Results from impaired subjects were similar, with even impaired subjects over 60 years old enjoying competitive gameplay. The subjects' personalities roughly predicted which mode they would prefer, which was especially evident in a poorly-matched impaired pair that preferred the single-player mode. CONCLUSIONS: Results indicate great potential for two-player rehabilitation games, in the form of greater enjoyment as well as potentially more intensive exercise compared to single-player games. However, the right game type needs to be chosen for each subject depending on skill and personality, along with selecting an appropriate co-player. Further studies with patients that are currently enrolled in rehabilitation programs are recommended, and the subjective measures used in our study should be augmented with objective measures such as electromyography.

Comparing Walking-Related Everyday Life Tasks of Children with Gait Disorders in a Virtual Reality Setup With a Physical Setup: Cross-Sectional Noninferiority Study.

BACKGROUND: A frequent rehabilitation goal for children with gait disorders is to practice daily-life walking activities. Unfortunately, these are often difficult to practice in a conventional therapeutic setting. Virtual reality (VR) with head-mounted displays (HMDs) could be a promising approach in neurorehabilitation to train such activities in a safe environment. First, however, we must know whether obstacles in VR are indeed mastered as obstacles. OBJECTIVE: This study aimed to provide information on whether VR is feasible and motivating to induce and practice movements needed to master real obstacles in children and adolescents with gait disorders. Furthermore, this project aims to evaluate which kinds of everyday walking activities are appropriate to be practiced in VR. METHODS: In this cross-sectional study, participants stepped over a bar, crossed a gap, balanced over a beam, and circumvented stationary obstructions arranged in a course under real physical and virtual conditions wearing a VR HMD. We recorded the respective primary outcomes (step height, step length, step width, and minimal shoulder-obstacle distance) with motion capture. We then calculated the mean differences and 95% CI of the spatiotemporal parameters between the VR and physical setup and later compared them using noninferiority analysis with margins defined a priori by a clinical expert panel. Additionally, the participants responded to a standardized questionnaire while the therapists observed and evaluated their movement performance. RESULTS: We recruited 20 participants (mean age 12.0, range 6.6-17.8 years) with various diagnoses affecting their walking ability. At 3.77 (95% CI 1.28 to 6.26) cm, the mean difference in step height of the leading foot in the overstepping task did not exceed the predefined margin of -2 cm, thus signifying noninferiority of the VR condition compared to mastering the physical obstacles. The same was true for step length (-1.75, 95% CI -4.91 to 1.41 cm; margin -10 cm), step width (1.05, 95% CI 0.20 to -1.90 cm; margin 3 cm), and the minimal shoulder-obstacle distance (0.25, 95% CI -0.85 to 0.35 cm; margin -2 cm) in the other tasks. Only the trailing foot in the overstepping task yielded inconclusive results. CONCLUSIONS: Children with gait disorders perform everyday walking tasks like overstepping, crossing, balancing, or circumventing similarly in physical and VR environments, suggesting that VR could be a feasible therapeutic tool to practice everyday walking tasks.

Energy efficiency increase in a chemical production site.

Sustainability has become a key factor for the chemical industry. One element of sustainability is energy efficiency in manufacturing processes. This article illustrates the strategic energy initiatives of a leading global operating company and the implementation of its elements into practice. Some successful energy-saving projects are highlighted.

Virtual reality aided training of combined arm and leg movements of children with CP.

Cerebral palsy (CP) occurs in over 2 out of 1000 live births and can impair motor control and cognition. Our goal was to create a robotic rehabilitation environment that mimics real-life situations by allowing simultaneous exercise of upper and lower limbs. We chose to use the Lokomat as a gait robot and added a novel removable arm robot, called PASCAL (pediatric arm support robot for combined arm and leg training), that was integrated into the Lokomat environment. We also added a virtual reality (VR) environment that enables the subject to perform motivating game-like scenarios incorporating combined arm and leg movements. In this paper we summarize the design of PASCAL and present the novel virtual environment including first experimental results. The next step will be to test whether a combined application of the virtual environment and the two simultaneously working robots is feasible in healthy participants, and finally to clinically evaluate the entire system on children with CP.

Virtual Reality as a Therapy Tool for Walking Activities in Pediatric Neurorehabilitation: Usability and User Experience Evaluation.

BACKGROUND: Many essential walking activities in daily life, such as crossing a street, are challenging to practice in conventional therapeutic settings. Virtual environments (VEs) delivered through a virtual reality (VR) head-mounted display (HMD) would allow training such activities in a safe and attractive environment. Furthermore, the game-like character and high degree of immersion in these applications might help maintain or increase children's motivation and active participation during the rehabilitation process. OBJECTIVE: This study aimed to investigate the usability, user experience, and acceptability of an immersive VE experienced through a VR HMD to train everyday life walking activities in pediatric neurorehabilitation. METHODS: In a cross-sectional study, 21 youths (median age 12.1 years; range 6.8-17.7 years) with a neuromotor impairment undergoing inpatient or outpatient neurorehabilitation tested a VE experienced through the VR HMD Oculus Quest. The participants, accompanied by their physiotherapists, moved freely around a 4.4 by 10-meter VE, displaying a magical forest and featuring various gamified everyday activities in different game designs. Using their hands, represented in the VE, the participants could interact with the virtual objects placed throughout the VE and trigger visual and auditory feedback. Symptoms of cybersickness were checked, and usability, user experience, and acceptability were evaluated using customized questionnaires with a visual analog scale for youths and a 5-point Likert scale for their therapists. RESULTS: None of the participants reported any signs of cybersickness after 20 minutes of VR HMD exposure time. They rated comfort (median 10/10) and movement ability (median 10/10) with the VR HMD as high. The VE was perceived as being really there by the majority (median 8/10), and the participants had a strong feeling of spatial presence in the VE (median 9.5/10). They enjoyed exploring the virtual world (median 10/10) and liked this new therapy approach (median 10/10). Therapists' acceptance of the VR HMD was high (4/5). There were 5 patients that needed more support than usual, mainly for supervision, when moving around with the VR HMD. Otherwise, therapists felt that the VR HMD hardly affected their patients' movement behavior (median 4.75/5), whereas it seemed to increase their level of therapy engagement (median 4/5) compared to conventional physiotherapy sessions. CONCLUSIONS: This study demonstrates the usability of an immersive VE delivered through a VR HMD to engage youths in the training of everyday walking activities. The participants' and therapists' positive ratings on user experience and acceptance further support the promising application of this technology as a future therapeutic tool in pediatric neurorehabilitation.

Test-retest reliability of upper limb robotic exoskeleton assessments in children and youths with brain lesions.

In children with congenital or acquired brain lesions, impaired upper limb function can affect independence. Assessing upper limb function is important for planning and evaluating neurorehabilitative interventions. Robotic devices increase measurement-objectivity and enable measuring parameters reflecting more complex motor functions. We investigated the relative and absolute test-retest reliability of assessments to measure upper limb functions in children and adolescents with brain lesions with the exoskeleton ChARMin. Thirty children (9 females, mean age ± SD = 12.5 ± 3.3 years) with congenital brain injuries (n = 15), acquired (n = 14), both (n = 1) and impaired upper limb function participated. They performed the following ChARMin assessments and repeated them within three to seven days: active and passive Range of Motion (ROM), Strength, Resistance to Passive Movement, Quality of Movement, Circle, and Workspace. We calculated the systematic difference, Intraclass Correlation Coefficient (ICC) and Smallest Real Difference (SRD) for each parameter. Six parameters of three assessments showed systematic errors. ICCs ranged from little to very high and SRD values varied considerably. Test-retest reliability and measurement errors ranged widely between the assessments. Systematic differences indicated that random day-to-day variability in performance would be responsible for reduced reliability of those parameters. While it remains debatable whether robot-derived outcomes should replace certain routine assessments (e.g., ROM, strength), we recommend applying certain technology-based assessments also in clinical practice.Trial registration: This study was registered prospectively at ClinicalTrials.gov (identifier: NCT02443857) on May 14, 2015.

DO YOUTHS WITH NEUROMOTOR DISORDER AND THEIR THERAPISTS PREFER A MIXED OR VIRTUAL REALITY HEAD-MOUNTED DISPLAY?

OBJECTIVE: To evaluate the usability of 2 head-mounted displays in youths undergoing neurorehabilitation; a mixed reality head-mounted display and a virtual reality head-mounted display. DESIGN: Observational cross-sectional study. PATIENTS: Thirteen youths (age range 7.8-16.5 years) with neuromotor disorder. METHODS: Youths wore a mixed reality or a virtual reality head-mounted display while being verbally guided through a scene with virtual objects. Differences between the 2 systems, regarding usability, user experience, and acceptability, were evaluated using standardized questions for the youths and their therapists. System preferences and symptoms of cybersickness were noted. RESULTS: Both head-mounted displays were easy to mount and adjust to the children's heads, but the mixed reality system was unstable in 40% of the youths. Participants stated that they could move naturally with both devices. Object appearance scored higher with the virtual reality system, while therapists rated youths' movement execution and needed additional support in favour of the mixed reality system. Most youths preferred the virtual reality device, mainly due to the more distinct appearance of objects and the objects' richer colours. Therapists' preferences were balanced. Two children reported minimal signs of cybersickness. CONCLUSION: Youths and therapists accepted both systems well, with advantages regarding usability, user experience, and preference for the virtual reality, and acceptability for the mixed reality head-mounted display.

Contextual interference in children with brain lesions: a pilot study investigating blocked vs. random practice order of an upper limb robotic exergame.

INTRODUCTION: Evidence about contextual interference in children with brain lesions when practising motor tasks is lacking. Our main objective was to evaluate the feasibility of a randomised controlled trial (RCT) comparing blocked with random practice order of an upper limb robotic exergame to improve reaching in children with neuromotor disorders with a pilot trial. METHODS: We recruited children with brain lesions and impaired upper limb functions who underwent a 3-week schedule that consisted of baseline assessments, intervention period (participants were randomised to a blocked or random order group), and follow-up assessment. We evaluated ten feasibility criteria, including the practicability of the inclusion/exclusion criteria, recruitment rate, feasibility of randomisation, scheduling procedure, and the participants' programme adherence. RESULTS: The inclusion/exclusion criteria were not completely feasible as patients who were not able to perform the exergames were included. Twelve participants were recruited, and six datasets were used for analysis. The scheduling and randomisation procedures were generally feasible, but the procedure was only partially feasible for the participants, as some sessions were aborted due to lack of motivation and fatigue. CONCLUSION: An RCT following this study protocol is not feasible. We formulated suggestions for future studies that aim to investigate contextual interference as in this pilot study. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02443857 , registered on May 14, 2015.

Contextual interference in children with brain lesions: protocol of a pilot study investigating blocked vs. random practice order of an upper limb robotic exergame.

BACKGROUND: If adults practice several motor tasks together, random practice leads to better transfer and retention compared to blocked practice. Knowledge about this contextual interference (CI) effect could be valuable to improve neurorehabilitation of children. We present the protocol of a randomised controlled pilot study investigating the feasibility of blocked practice vs. random practice of robot-assisted upper limb reaching in children with brain lesions undergoing neurorehabilitation. METHODS: Children with affected upper limb function due to congenital or acquired brain lesions undergoing neurorehabilitation will be recruited for a randomised controlled pilot study with a 3-week procedure. In the control week (1), two assessment blocks (robot-assisted reaching tasks, Melbourne assessment 2, subscale fluency), 2 days apart, take place. In the practice week (2), participants are randomly allocated to blocked practice or random practice and perform 480 reaching and backward movements in the horizontal and vertical plane using exergaming with an exoskeleton robot per day during three consecutive days. Assessments are performed before, directly after and 1 day after the practice sessions. In the follow-up week (3), participants perform the assessments 1 week after the final practice session. The primary outcome is the immediate transfer of the Melbourne Assessment 2, subscale fluency. Secondary outcomes are the immediate retention, 1-day and 1-week delayed transfer and retention and acquisition during the practice sessions. We will evaluate the feasibility of the inclusion criteria, the recruitment rate, the scheduling procedure, the randomisation procedure, the procedure for the participants, the handling of the robot, the handling of the amount of data, the choice of the outcome measures and the influence of other therapies. Furthermore, we will perform a power calculation using the data to estimate the sample size for the main trial. DISCUSSION: The protocol of the pilot study is a first step towards a future main randomised controlled trial. This low risk pilot study might induce some benefits for the participants. However, we need to place its results into perspective, especially concerning the generalisability, as it remains questionable whether improving reaching constrained within a robotic device will ameliorate daily life reaching tasks. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT02443857.

From the Representativeness of Sampled Odors to an Olfactive Camera.

Although representativeness is often a prerequisite when sampling odors, the methods used have never been assessed from the analytical and sensory perspective simultaneously. We validate several critical innovations in the methods used to sample odors, starting with a previously developed static-and-trapped headspace (S&T-HS) cell, to minimize sorptive biases and allow for thermodesorption of trapped odors. The addition of a desorption oven allows for restoration and testing of odors sampled by not only S&T-HS but also other techniques (solid-phase microextraction, headspace sorptive extraction, purge-and-trap headspace). The S&T-HS cell exhibits satisfactory representativeness, much higher than the three other techniques. This allows, for the first time, a proposal to use this technology as an olfactive camera to capture and restore an odor. The method was tested on a sample of a complex fresh ashtray odor.

First validation of a novel assessgame quantifying selective voluntary motor control in children with upper motor neuron lesions.

The question whether novel rehabilitation interventions can exploit restorative rather than compensatory mechanisms has gained momentum in recent years. Assessments measuring selective voluntary motor control could answer this question. However, while current clinical assessments are ordinal-scaled, which could affect their sensitivity, lab-based assessments are costly and time-consuming. We propose a novel, interval-scaled, computer-based assessment game using low-cost accelerometers to evaluate selective voluntary motor control. Participants steer an avatar owl on a star-studded path by moving the targeted joint of the upper or lower extremities. We calculate a target joint accuracy metric, and an outcome score for the frequency and amplitude of involuntary movements of adjacent and contralateral joints as well as the trunk. We detail the methods and, as a first proof of concept, relate the results of select children with upper motor neuron lesions (n = 48) to reference groups of neurologically intact children (n = 62) and adults (n = 64). Linear mixed models indicated that the cumulative therapist score, rating the degree of selectivity, was a good predictor of the involuntary movements outcome score. This highlights the validity of this assessgame approach to quantify selective voluntary motor control and warrants a more thorough exploration to quantify changes induced by restorative interventions.

Robot-assisted arm assessments in spinal cord injured patients: a consideration of concept study.

Robotic assistance is increasingly used in neurological rehabilitation for enhanced training. Furthermore, therapy robots have the potential for accurate assessment of motor function in order to diagnose the patient status, to measure therapy progress or to feedback the movement performance to the patient and therapist in real time. We investigated whether a set of robot-based assessments that encompasses kinematic, kinetic and timing metrics is applicable, safe, reliable and comparable to clinical metrics for measurement of arm motor function. Twenty-four healthy subjects and five patients after spinal cord injury underwent robot-based assessments using the exoskeleton robot ARMin. Five different tasks were performed with aid of a visual display. Ten kinematic, kinetic and timing assessment parameters were extracted on joint- and end-effector level (active and passive range of motion, cubic reaching volume, movement time, distance-path ratio, precision, smoothness, reaction time, joint torques and joint stiffness). For cubic volume, joint torques and the range of motion for most joints, good inter- and intra-rater reliability were found whereas precision, movement time, distance-path ratio and smoothness showed weak to moderate reliability. A comparison with clinical scores revealed good correlations between robot-based joint torques and the Manual Muscle Test. Reaction time and distance-path ratio showed good correlation with the "Graded and Redefined Assessment of Strength, Sensibility and Prehension" (GRASSP) and the Van Lieshout Test (VLT) for movements towards a predefined position in the center of the frontal plane. In conclusion, the therapy robot ARMin provides a comprehensive set of assessments that are applicable and safe. The first results with spinal cord injured patients and healthy subjects suggest that the measurements are widely reliable and comparable to clinical scales for arm motor function. The methods applied and results can serve as a basis for the future development of end-effector and exoskeleton-based robotic assessments.

Effect of rifampicin on the pharmacokinetics of imatinib mesylate (Gleevec, STI571) in healthy subjects.

OBJECTIVE: This study was carried out to investigate the influence of CYP3A induction with rifampicin on imatinib (Gleevec) exposure. METHODS: The study employed a single center, single-sequence design. A group of 14 healthy male and female subjects received imatinib as a single 400 mg oral dose on two occasions: on study day 1 and on study day 15. Rifampicin treatment (600 mg once daily) for CYP4503A induction was initiated on study day 8 and maintained until day 18. Imatinib pharmacokinetics were determined up to 96 h after dosing on day 1 (no induction) and on days 15-18 (during concomitant rifampicin). Plasma concentrations of imatinib and its main metabolite CGP74588 were determined using a LC/MS/MS method. The ratio of 6beta-hydroxycortisol to cortisol excreted in the urine was measured to monitor the induction of CYP3A. RESULTS: During concomitant rifampicin administration, the mean imatinib C(max), AUC(0-24) and AUC(0- infinity ) decreased by 54% (90% CI: 48-60%), 68% (64-70%) and 74% (71-76%), respectively. The increase in clearance (Cl/f) was 385% (348-426%) during rifampicin treatment. The mean C(max) and AUC(0-24) of the metabolite CGP74588 increased by 88.6% (68.3%-111.4%) and 23.9% (13.5%-35.2%) after rifampicin pretreatment. However, the AUC(0- infinity ) decreased by 11.7% (3.3-19.4%). All subjects demonstrated a marked induction of hepatic microsomal CYP3A analyzed by the excretion ratio of 6beta-hydroxycortisol to cortisol from a mean baseline concentration of 5.6 U to 50.5 U. CONCLUSION: Concomitant use of imatinib and rifampicin or other potent inducers of CYP4503A may result in subtherapeutic plasma concentrations of imatinib. In patients in whom rifampicin or other CYP3A inducers are prescribed, alternative therapeutic agents with less potential for enzyme induction should be selected.

Bioequivalence, safety, and tolerability of imatinib tablets compared with capsules.

PURPOSE: Imatinib (Glivec) has been established as a highly effective therapy for chronic myeloid leukemia and gastrointestinal tumors. The recommended daily dosage of 400-600 mg requires simultaneous intake of up to six of the current 100-mg capsules. Due to the need to swallow multiple capsules per dose, there is a potential negative impact on treatment adherence; therefore, a new imatinib 400-mg film-coated tablet has been developed. To improve dosing flexibility, particularly with regard to the pediatric population and the management of adverse events, a scored 100-mg film-coated tablet has also been introduced. EXPERIMENTAL DESIGN: A group of 33 healthy subjects were randomly assigned to one of six treatment sequences, in which they received imatinib as 4 x 100-mg capsules (reference), 4 x 100-mg scored tablets (test), and 1 x 400-mg tablet (test). Blood sampling was performed for up to 96 h after dosing, followed by a 10-day washout period prior to the next sequence. After the third dosing, subjects were monitored to assess delayed drug-related adverse events. Pharmacokinetic parameters were assessed using concentration-time curves for plasma imatinib and its metabolite CGP74588. RESULTS: Median Tmax was 2.5 h for capsules and tablets. Mean AUC((0-inf)) values were 27,094, 26,081 and 25,464 ng.h/ml for 4 x 100-mg capsules, 4 x 100-mg tablets, and 1 x 400-mg tablets, respectively. Cmax values were 1748, 1638 and 1606 ng/ml, and t(1/2) values were 15.8, 15.9 and 15.7 h. The test/reference ratios for AUC((0-inf)), AUC((0-96) (h)), and C(max) were 0.98, 0.98 and 0.95 for 4 x 100-mg tablets versus 4 x 100-mg capsules, and 0.95, 0.95 and 0.92 for 1 x 400-mg tablet versus 4 x 100-mg capsules. The 95% confidence intervals were fully contained within the interval (0.80, 1.25). Eight mild and one moderate adverse event considered to be drug related were reported. These events showed no clustering by type of dosage form and were of little to no clinical significance. CONCLUSIONS: Film-coated 100-mg (scored) and 400-mg tablet dose forms of imatinib are bioequivalent to the commercial 100-mg hard-gelatin capsule, and are as safe and well tolerated.

Assist-as-needed path control for the PASCAL rehabilitation robot.

Adults and children with neurological disorders often require rehabilitation therapy to improve their arm motor functions. Complementary to conventional therapy, robotic therapy can be applied. Such robots should support arm movements while assisting only as much as needed to ensure an active participation of the patient. Different control strategies are known to provide arm support to the patient. The path controller is a strategy that helps the patient's arm to stay close to a given path while allowing for temporal and spatial freedom. In this paper, an assist-as-needed path controller is presented that is implemented in the end-effector-based robot PASCAL, which was designed for children with cerebral palsy. The new control approach is a combination of an existing path controller with additional speed restrictions to support, when the arm speed is too slow, and to resist, when the speed is too fast. Furthermore, a target position gain scheduling is introduced in order to reach a target position with a predefined precision as well as an adaptable direction-dependent supportive flux that supports along the path. These path control features were preliminarily tested with a healthy adult volunteer in different conditions. The presented controller covers the range from a completely passive user, who needs full support to an actively performed movement that needs no assistance. In close future, the controller is planned to be used to enable reaching in children as well as in adults and help to increase the intensity of the rehabilitation therapy by assisting the hand movement and by provoking an active participation.

Estimating the patient's contribution during robot-assisted therapy.

Robot-assisted therapy has become increasingly common in neurorehabilitation. Sophisticated controllers have been developed for robots to assist and cooperate with the patient. It is difficult for the patient to judge to what extent the robot contributes to the execution of a movement. Therefore, methods to comprehensively quantify the patient's contribution and provide feedback are of key importance. We developed a method comprehensively to estimate the patient's contribution by combining kinematic measures and the motor assistance applied. Inverse dynamic models of the robot and the passive human arm calculate the required torques to move the robot and the arm and build, together with the recorded motor torque, a metric (in percentage) that represents the patient's contribution to the movement. To evaluate the developed metric, 12 nondisabled subjects and 7 patients with neurological problems simulated instructed movement contributions. The results are compared with a common performance metric. The estimation shows very satisfying results for both groups, even though the arm model used was strongly simplified. Displaying this metric to patients during therapy can potentially motivate them to actively participate in the training.

Versatile robotic interface to evaluate, enable and train locomotion and balance after neuromotor disorders.

Central nervous system (CNS) disorders distinctly impair locomotor pattern generation and balance, but technical limitations prevent independent assessment and rehabilitation of these subfunctions. Here we introduce a versatile robotic interface to evaluate, enable and train pattern generation and balance independently during natural walking behaviors in rats. In evaluation mode, the robotic interface affords detailed assessments of pattern generation and dynamic equilibrium after spinal cord injury (SCI) and stroke. In enabling mode,the robot acts as a propulsive or postural neuroprosthesis that instantly promotes unexpected locomotor capacities including overground walking after complete SCI, stair climbing following partial SCI and precise paw placement shortly after stroke. In training mode, robot-enabled rehabilitation, epidural electrical stimulation and monoamine agonists reestablish weight-supported locomotion, coordinated steering and balance in rats with a paralyzing SCI. This new robotic technology and associated concepts have broad implications for both assessing and restoring motor functions after CNS disorders, both in animals and in humans.