Psychological state estimation from physiological recordings during robot-assisted gait rehabilitation

Understanding task "challenge" in stroke rehabilitation: an interdisciplinary concept analysis

PURPOSE

Rehabilitation plays a critical role in minimising disability after stroke, with the concept of "challenge" proposed to be essential to rehabilitation efficacy and outcomes. This review unpacks how challenge is conceptualised in stroke rehabilitation literature from the perspectives of physiotherapy, occupational therapy, speech-language therapy and people with stroke. A secondary purpose was to provide a definition of challenge that is applicable to stroke rehabilitation.

METHODS

Principle-based concept analysis was utilised to examine challenge within the stroke rehabilitation literature. Forty-two papers were included. Data analysis involved immersion, analytical questioning, coding and synthesis to elicit the conceptual components of challenge.

RESULTS

Challenge was understood as a multidimensional and dynamic concept with three facets: nominal, functional and perceived challenge. Functional and perceived challenge were integral to optimal challenge. Optimal challenge was central to enhancing the outcomes and experiences of people with stroke, in rehabilitation and everyday life.

CONCLUSIONS

Challenge is a key concept which, when carefully optimised to the person's ability and experience, may positively influence their learning, recovery and engagement after stroke. This review lays a conceptual foundation for better understanding, operationalisation and advancement of challenge, offering important implications for addressing the growing burden of stroke disability, through rehabilitation.

Cooperative Goal Generation for Reaching Tasks in Robot-Assisted Rehabilitation

Robot-assisted neurorehabilitation requires automated generation of goal positions for reaching tasks in functional movement therapy. In state-of-the-art solutions, these positions are determined by a motivational therapy game either through constraints on the end-effector (2D or 3D games), or individual arm joints (1D games). Consequently, these positions cannot be adapted to the patients' specific needs by the therapist, and the effectiveness of the training is reduced. We solve this issue by generating goal positions using Gaussian Mixture Models and probability density maps based on the active range of motion of the patient and desired activities, while being compliant with existing game constraints. Therapists can modify the goal generation via an intuitive difficulty and activity parameter. The pipeline was tested on the upper-limb exoskeleton ANYexo 2.0. We have shown that the range of motion exploration rate could be altered from 0.39% to 5.9% per task and that our method successfully generated a sequence of reaching tasks that matched the range of motion of the selected activity, up to an inlier accuracy of 78.9%. Results demonstrate that the responsibilities of the therapy game (i.e., motivating the patient) and the therapists (i.e., individualizing the training) could be distributed properly. We believe that with our pipeline, effective cooperation between the involved agents is achieved, and the provided therapy can be improved.

Tailoring Upper-Limb Robot-Aided Orthopedic Rehabilitation on Patients' Psychophysiological State

Physical therapy keeps exploiting more and more the capabilities of the robot of adapting the treatments to patients' needs. This paper aims at presenting a psychophysiological-aware control strategy for upper limb robot-aided orthopedic rehabilitation. The main features are the capability of i) generating point-to-point trajectories inside an adaptable workspace, ii) providing assistance in guiding the patients' limbs in accomplishing the assigned task allowing them to freely move with a certain degree of spatial and temporal autonomy and iii) tuning the control parameters according to the patients' kinematics performance and psychophysiological state. The implemented control strategy is validated in a real clinical setting on eight orthopedic patients undergoing twenty daily robot-aided rehabilitation sessions. The psychophysiological-aware control strategy evidenced a positive impact on the enrolled participants since they are effectively conducted in a calmer condition with respect to the patients who did not receive the psychophysiological adaptation. Moreover, clinical performance indicators suggest that the proposed tailored control strategy improves motor functions.

Technology Assisted Rehabilitation Patient Perception Questionnaire (TARPP-Q): development and implementation of an instrument to evaluate patients' perception during training

BACKGROUND

The introduction of technology-assisted rehabilitation (TAR) uncovers promising challenges for the treatment of motor disorders, particularly if combined with exergaming. Patients with neurological diseases have proved to benefit from TAR, improving their performance in several activities. However, the subjective perception of the device has never been fully addressed, being a conditioning factor for its use. The aims of the study were: (a) to develop a questionnaire on patients' personal experience with TAR and exergames in a real-world clinical setting; (b) to administer the questionnaire to a pilot group of neurologic patients to assess its feasibility and statistical properties.

METHODS

A self-administrable and close-ended questionnaire, Technology Assisted Rehabilitation Patient Perception Questionnaire (TARPP-Q), designed by a multidisciplinary team, was developed in Italian through a Delphi procedure. An English translation has been developed with consensus, for understandability purposes. The ultimate version of the questionnaire was constituted of 10 questions (5 with multiple answers), totalling 29 items, exploring the patient's performance and personal experience with TAR with Augmented Performance Feedback. TARPP-Q was then administered pre-post training in an observational, feasible, multi-centric study. The study involved in-patients aged between 18 and 85 with neurological diseases, admitted for rehabilitation with TAR (upper limb or gait). FIM scale was run to control functional performance.

RESULTS

Forty-four patients were included in the study. All patients answered the TARPP-Q autonomously. There were no unaccounted answers. Exploratory factor analyses identified 4 factors: Positive attitude, Usability, Hindrance perception, and Distress. Internal consistency was measured at T0. The values of Cronbach's alpha ranged from 0.72 (Distress) to 0.92 (Positive attitude). Functional Independence Measure (FIM®) scores and all TARPP-Q factors (Positive attitude, Usability, Hindrance perception, except for Distress (p = 0.11), significantly improved at the end of the treatment. A significant positive correlation between Positive attitude and Usability was also recorded.

CONCLUSIONS

The TARPP-Q highlights the importance of patients' personal experience with TAR and exergaming. Large-scale applications of this questionnaire may clarify the role of patients' perception of training effectiveness, helping to customize devices and interventions.

Reduced Parasympathetic Reactivation during Recovery from Exercise in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome

Although autonomic nervous system (ANS) dysfunction in Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) has been proposed, conflicting evidence makes it difficult to draw firm conclusions regarding ANS activity at rest in ME/CFS patients. Although severe exercise intolerance is one of the core features of ME/CFS, little attempts have been made to study ANS responses to physical exercise. Therefore, impairments in ANS activation at rest and following exercise were examined using a case-control study in 20 ME/CFS patients and 20 healthy people. Different autonomous variables, including cardiac, respiratory, and electrodermal responses were assessed at rest and following an acute exercise bout. At rest, parameters in the time-domain represented normal autonomic function in ME/CFS, while frequency-domain parameters indicated the possible presence of diminished (para)sympathetic activation. Reduced parasympathetic reactivation during recovery from exercise was observed in ME/CFS. This is the first study showing reduced parasympathetic reactivation during recovery from physical exercise in ME/CFS. Delayed HR recovery and/or a reduced HRV as seen in ME/CFS have been associated with poor disease prognosis, high risk for adverse cardiac events, and morbidity in other pathologies, implying that future studies should examine whether this is also the case in ME/CFS and how to safely improve HR recovery in this population.

The influence of psychological and cognitive states on error-related negativity evoked during post-stroke rehabilitation movements

Background

Recently, error-related negativity (ERN) signals are proposed to develop an assist-as-needed robotic stroke rehabilitation program. Stroke patients’ state-of-mind, such as motivation to participate and active involvement in the rehabilitation program, affects their rate of recovery from motor disability. If the characteristics of the robotic stroke rehabilitation program can be altered based on the state-of-mind of the patients, such that the patients remain engaged in the program, the rate of recovery from their motor disability can be improved. However, before that, it is imperative to understand how the states-of-mind of a participant affect their ERN signal.

Methods

This study aimed to determine the association between the ERN signal and the psychological and cognitive states of the participants. Experiments were conducted on stroke patients, which involved performing a physical rehabilitation exercise and a questionnaire to measure participants' subjective experience on four factors: motivation in participating in the experiment, perceived effort, perceived pressure, awareness of uncompleted exercise trials while performing the rehabilitation exercise. Statistical correlation analysis, EEG time-series and topographical analysis were used to assess the association between the ERN signals and the psychological and cognitive states of the participants.

Results

A strong correlation between the amplitude of the ERN signal and the psychological and cognitive states of the participants was observed, which indicate the possibility of estimating the said states using the amplitudes of the novel ERN signal.

Conclusions

The findings pave the way for the development of an ERN based dynamically adaptive assist-as-needed robotic stroke rehabilitation program of which characteristics can be altered to keep the participants’ motivation, effort, engagement in the rehabilitation program high. In future, the single-trial prediction ability of the novel ERN signals to predict the state-of-mind of stroke patients will be evaluated.

Application of the Caprini risk assessment model for deep vein thrombosis among patients undergoing laparoscopic surgery for colorectal cancer

Application of the Caprini risk assessment model was explored in patients with deep vein thrombosis (DVT) after laparoscopic colorectal cancer surgery.This study was a prospective study. The risk factors for DVT were assessed with a survey at baseline and on the morning of surgery, first day after surgery and sixth day by using repeated blood vessels on color Doppler ultrasound of the lower limbs, and the intraoperative and postoperative conditions were recorded.Among 148 surgical patients, 24.3% had asymptomatic DVT. According to the risk stratification, the incidence of DVT was related to the Caprini score (P < .001). The area under the curve of the Caprini model was 0.701 ± 0.047 (95% CI: 0.609-0.793, P<.001). The Youden index was 0.368, while the critical point was 10.5 in the Caprini model, corresponding to a sensitivity of 0.806 and a specificity of 0.563. Age, cardiovascular disease, intraoperative blood loss, postoperative fever, preoperative preparation, and hospital stay were higher in DVT patients than in patients without DVT. Moreover, the incidence of DVT in patients with a lithotomy position was higher than that in patients with a scissors position. In binary logistic regression analysis, the independent risk factors for DVT development were age, intraoperative blood loss, and preoperative preparation time.The Caprini model can be used for the prediction of venous thromboembolism in laparoscopic colorectal cancer surgery patients. The thrombosis risk assessment model must be established in line with patients undergoing endoscopic malignant tumor surgery.

Effects of Presence and Challenge Variations on Emotional Engagement in Immersive Virtual Environments

Serious games and immersive virtual reality promote emotional engagement during learning tasks, mostly by providing (1) skill-adapted challenges with performance feedback (for trial and error learning) and (2) enhanced presence (further reactions to multimodal stimuli), respectively. However, it is still unclear how each of these two strategies independently influence emotional states to engage subjects to a task. This study assessed the dimensions of emotion (valence-arousal-dominance) of 87 healthy subjects in a virtual game, assigned to 2 groups that were exposed to a different set of 5 trials: Group A experienced game variations by virtual factors affecting user's presence, whereas group B experienced levels of difficulty, affecting challenge. Emotional reports and 26 features extracted from physiological signals were statistically analyzed. Results showed that presence-based experimental conditions were able to modify the sense of arousal, whereas valence and dominance responded to challenge variation, i.e. were positively correlated with game score. Arousal is likely to increase with low sense of coexistence (social presence) and decrease with low scenario realism (physical presence). Faster breathing and higher skin conductance (SC) were detected at high challenge, whereas heart rate variability and SC increased with higher arousal. The evidence from this study suggests that both strategies can be used to separately influence dimensions of emotion, pointing out the customization of presence-based factors as a promising method to adjust emotional engagement by impacting arousal. Further research should be undertaken to identify the independent effect of single presence factors on emotional states.

Classification of Rehabilitation Participation in Elderly In-patients with Mild Cognitive Impairments Utilizing Physiological Responses

This study investigated the possibility of utilising physiological responses and machine learning techniques to determine the degree of participation of in-patients with mild cognitive impairment at rehabilitation institutions. Physiological signals related to autonomic functions, cardio-activity, sweat gland activation, and skin surface temperature were obtained, and machine learning classifiers were used to classify rehabilitation participation levels as higher or lower participation when participants were required to perform a VR-based rehabilitative task. Classifiers such as a decision tree or support vector machine can effectively determine two different levels of participation suggesting the proposed approach can help therapists assess an important aspect of client satisfaction.

Brain-machine interfaces for controlling lower-limb powered robotic systems

OBJECTIVE

Lower-limb, powered robotics systems such as exoskeletons and orthoses have emerged as novel robotic interventions to assist or rehabilitate people with walking disabilities. These devices are generally controlled by certain physical maneuvers, for example pressing buttons or shifting body weight. Although effective, these control schemes are not what humans naturally use. The usability and clinical relevance of these robotics systems could be further enhanced by brain-machine interfaces (BMIs). A number of preliminary studies have been published on this topic, but a systematic understanding of the experimental design, tasks, and performance of BMI-exoskeleton systems for restoration of gait is lacking.

APPROACH

To address this gap, we applied standard systematic review methodology for a literature search in PubMed and EMBASE databases and identified 11 studies involving BMI-robotics systems. The devices, user population, input and output of the BMIs and robot systems respectively, neural features, decoders, denoising techniques, and system performance were reviewed and compared.

MAIN RESULTS

Results showed BMIs classifying walk versus stand tasks are the most common. The results also indicate that electroencephalography (EEG) is the only recording method for humans. Performance was not clearly presented in most of the studies. Several challenges were summarized, including EEG denoising, safety, responsiveness and others.

SIGNIFICANCE

We conclude that lower-body powered exoskeletons with automated gait intention detection based on BMIs open new possibilities in the assistance and rehabilitation fields, although the current performance, clinical benefits and several key challenging issues indicate that additional research and development is required to deploy these systems in the clinic and at home. Moreover, rigorous EEG denoising techniques, suitable performance metrics, consistent trial reporting, and more clinical trials are needed to advance the field.

Bringing Psychological Strategies to Robot-Assisted Physiotherapy for Enhanced Treatment Efficacy

Robotic technologies offer a range of functions to augment clinical rehabilitation practice. However, compliance with robot-assisted rehabilitation techniques has not been optimally achieved. Traditional approaches to improving the treatment efficacy are focusing more on the system function, while psychological factors have not been integrated comprehensively. In this perspective paper, eight key factors reflecting three conceptions-robot design, function design, and patients' expectations have been evaluated and analyzed. Clinical results with 28 therapists and 84 patients indicate that integrating psychological strategies into robot-assisted physiotherapy may promote better trust and acceptance of rehabilitation robots.

Improving Challenge/Skill Ratio in a Multimodal Interface by Simultaneously Adapting Game Difficulty and Haptic Assistance through Psychophysiological and Performance Feedback

In order to harmonize robotic devices with human beings, the robots should be able to perceive important psychosomatic impact triggered by emotional states such as frustration or boredom. This paper presents a new type of biocooperative control architecture, which acts toward improving the challenge/skill relation perceived by the user when interacting with a robotic multimodal interface in a cooperative scenario. In the first part of the paper, open-loop experiments revealed which physiological signals were optimal for inclusion in the feedback loop. These were heart rate, skin conductance level, and skin conductance response frequency. In the second part of the paper, the proposed controller, consisting of a biocooperative architecture with two degrees of freedom, simultaneously modulating game difficulty and haptic assistance through performance and psychophysiological feedback, is presented. With this setup, the perceived challenge can be modulated by means of the game difficulty and the perceived skill by means of the haptic assistance. A new metric (FlowIndex) is proposed to numerically quantify and visualize the challenge/skill relation. The results are contrasted with comparable previously published work and show that the new method afforded a higher FlowIndex (i.e., a superior challenge/skill relation) and an improved balance between augmented performance and user satisfaction (higher level of valence, i.e., a more enjoyable and satisfactory experience).

An analysis of physiological signals as a measure of task engagement in a multi-limb-coordination motor-learning task

There is widespread agreement in the physical rehabilitation community that task engagement is essential to effective neuromuscular recovery. Despite this, there are no clear measures of such task engagement. This paper assesses the extent to which certain physiological measurements might provide a measure of task engagement. In previous studies, correlations between mental focus and certain physiological measurements have been observed in subjects performing tasks requiring mental effort. In this study, the authors analyzed whether these signals showed similar correlation when subjects performed a multi-limb-coordination motor-learning task. Subjects played a video game which required the use of both arms and one leg to play a simplified electronic drum set with varying difficulty. Heart rate (HR), skin conductance level (SCL), and facial electromyogram (EMG) were recorded while the subjects played. Analysis of the recordings showed statistically significant correlations relating task difficulty to SCL, HR and EMG amplitude in corrugator supercilii. No statistically significant correlation was observed between task difficulty and EMG in frontalis.

Psychophysiological response to cognitive workload during symmetrical, asymmetrical and dual-task walking

Walking with a lower limb prosthesis comes at a high cognitive workload for amputees, possibly affecting their mobility, safety and independency. A biocooperative prosthesis which is able to reduce the cognitive workload of walking could offer a solution. Therefore, we wanted to investigate whether different levels of cognitive workload can be assessed during symmetrical, asymmetrical and dual-task walking and to identify which parameters are the most sensitive. Twenty-four healthy subjects participated in this study. Cognitive workload was assessed through psychophysiological responses, physical and cognitive performance and subjective ratings. The results showed that breathing frequency and heart rate significantly increased, and heart rate variability significantly decreased with increasing cognitive workload during walking (p<.05). Performance measures (e.g., cadence) only changed under high cognitive workload. As a result, psychophysiological measures are the most sensitive to identify changes in cognitive workload during walking. These parameters reflect the cognitive effort necessary to maintain performance during complex walking and can easily be assessed regardless of the task. This makes them excellent candidates to feed to the control loop of a biocooperative prosthesis in order to detect the cognitive workload. This information can then be used to adapt the robotic assistance to the patient's cognitive abilities.

Auto-adaptive robot-aided therapy using machine learning techniques

This paper presents an application of a classification method to adaptively and dynamically modify the therapy and real-time displays of a virtual reality system in accordance with the specific state of each patient using his/her physiological reactions. First, a theoretical background about several machine learning techniques for classification is presented. Then, nine machine learning techniques are compared in order to select the best candidate in terms of accuracy. Finally, first experimental results are presented to show that the therapy can be modulated in function of the patient state using machine learning classification techniques.

Physiological responses and energy cost of walking on the Gait Trainer with and without body weight support in subacute stroke patients

Background

Robotic-assisted walking after stroke provides intensive task-oriented training. But, despite the growing diffusion of robotic devices little information is available about cardiorespiratory and metabolic responses during electromechanically-assisted repetitive walking exercise. Aim of the study was to determine whether use of an end-effector gait training (GT) machine with body weight support (BWS) would affect physiological responses and energy cost of walking (ECW) in subacute post-stroke hemiplegic patients.

Methods

Participants: six patients (patient group: PG) with hemiplegia due to stroke (age: 66 ± 15y; time since stroke: 8 ± 3 weeks; four men) and 6 healthy subjects as control group (CG: age, 76 ± 7y; six men).

Interventions: overground walking test (OWT) and GT-assisted walking with 0%, 30% and 50% BWS (GT-BWS0%, 30% and 50%). Main Outcome Measures: heart rate (HR), pulmonary ventilation, oxygen consumption, respiratory exchange ratio (RER) and ECW.

Results

Intervention conditions significantly affected parameter values in steady state (HR: p = 0.005, V’E: p = 0.001, V'O₂: p < 0.001) and the interaction condition per group affected ECW (p = 0.002). For PG, the most energy (V’O₂ and ECW) demanding conditions were OWT and GT-BWS0%. On the contrary, for CG the least demanding condition was OWT. On the GT, increasing BWS produced a decrease in energy and cardiac demand in both groups.

Conclusions

In PG, GT-BWS walking resulted in less cardiometabolic demand than overground walking. This suggests that GT-BWS walking training might be safer than overground walking training in subacute stroke patients.

Prediction of gait recovery in spinal cord injured individuals trained with robotic gait orthosis

Background

Motor impairment is a major consequence of spinal cord injury (SCI). Earlier studies have shown that robotic gait orthosis (e.g., Lokomat) can improve an SCI individual’s walking capacity. However, little is known about the differential responses among different individuals with SCI. The present longitudinal study sought to characterize the distinct recovery patterns of gait impairment for SCI subjects receiving Lokomat training, and to identify significant predictors for these patterns.

Methods

Forty SCI subjects with spastic hypertonia at their ankles were randomly allocated to either control or intervention groups. Subjects in the intervention group participated in twelve 1-hour Lokomat trainings over one month, while control subjects received no interventions. Walking capacity was evaluated in terms of walking speed, functional mobility, and endurance four times, i.e. baseline, 1, 2, and 4 weeks after training, using the 10-Meter-Walking, Timed-Up-and-Go, and 6-Minute-Walking tests. Growth Mixture Modeling, an analytical framework for stratifying subjects based on longitudinal changes, was used to classify subjects, based on their gait impairment recovery patterns, and to identify the effects of Lokomat training on these improvements.

Results

Two recovery classes (low and high walking capacity) were identified for each clinical evaluation from both the control and intervention groups. Subjects with initial high walking capacity (i.e. shorter Timed-Up-and-Go time, higher 10-Meter-Walking speed and longer 6-Minute-Walking distance) displayed significant improvements in speed and functional mobility (0.033 m/s/week and–0.41 s/week respectively); however no significant change in endurance was observed. Subjects with low walking capacity exhibited no significant improvement. The membership in these two classes—and thus prediction of the subject’s gait improvement trajectory over time—could be determined by the subject’s maximum voluntary torque at the ankle under both plantar-and dorsi-flexion contractions determined prior to any training.

Conclusion

Our findings demonstrate that subjects responded to Lokomat training non-uniformly, and should potentially be grouped based on their likely recovery patterns using objective criteria. Further, we found that the subject’s ankle torque can predict whether he/she would benefit most from Lokomat training prior to the therapy. These findings are clinically significant as they can help individualize therapeutic programs that maximize patient recovery while minimizing unnecessary efforts and costs.

Influence of psychologic features on rehabilitation outcomes in patients with subacute stroke trained with robotic-aided walking therapy

OBJECTIVE

The aim of this study was to investigate whether the rehabilitation outcomes with robotic-aided gait therapy may be affected by patients' and caregivers' psychologic features after subacute stroke.

DESIGN

This is a controlled, longitudinal, observational pilot study conducted on 42 patients divided in robotic-assisted gait training plus conventional physical therapy group, robotic-assisted gait training dropout group, and conventional physical therapy group. The outcome measures were walking ability (Functional Ambulation Category) and independency in activities of daily living (Barthel Index) measured before and after intervention. Psychologic features were measured before intervention using the Hospital Anxiety and Depression Scale, the Eysenck Personality Questionnaire, and recovery locus of control in the patients and the State-Trait Anxiety Inventory and the Beck Depression Inventory in the caregivers.

RESULTS

Patient anxiety was significantly higher in those who refused/abandoned robotic therapy (P = 0.002). In the subjects allocated to the robotic group, the recovery of walking ability was significantly affected by the perceived recovery locus of control (P = 0.039, odds ratio = 14); and the recovery of independency in activities of daily living, by anxiety (P = 0.018, odds ratio = 0.042). Conversely, psychologic factors did not significantly affect the outcomes of conventional rehabilitation.

CONCLUSIONS

Psychologic features, particularly recovery locus of control and anxiety, affected the rehabilitative outcomes of the patients involved in robotic treatment more than those in conventional rehabilitation.

Auditory display as a prosthetic hand sensory feedback for reaching and grasping tasks

Upper limb amputees have to rely extensively on visual feedback in order to monitor and manipulate successfully their prosthetic device. This situation leads to high consciousness burden, which generates fatigue and frustration. Therefore, in order to enhance motor-sensory performance and awareness, an auditory display was used as a sensory feedback system for the prosthetic hand's spatio-temporal and force information in a complete reaching and grasping setting. The main objective of this study was to explore the effects of using the auditory display to monitor the prosthetic hand during a complete reaching and grasping motion. The results presented in this paper point out that the usage of an auditory display to monitor and control a robot hand improves the temporal and grasping performance greatly, while reducing mental effort and improving their confidence.

Cortical response to psycho-physiological changes in auto-adaptive robot assisted gait training

Robot-assisted treadmill training improves motor function and walking ability in neurologically impaired patients. However, despite attention having been shown to play a role in training success, psychological responsiveness to task difficulty and motivational levels at task onset have not been measured. Seven healthy subjects participated in a robot-assist treadmill training task. Subjects engaged in a virtual task with varying difficulty levels that was shown to induce a feeling of being bored, excited and over-stressed. The participants' mental engagement was measured using the ECG-based heart rate variability in real time, during gait training as a proxy for EEG and psychological test batteries. Heart rate variability (HRV), which has been shown to reflect cortical engagement for both cognitive and physical tasks, was measured using nonlinear measures obtained from the Poincaré plot. We show that the cortical response to the task measured with HRV varies in relation to the level of mental engagement in response to the difficulty level of the virtual task. From these results we propose that nonlinear measures quantify cortical response / motivational level to robot-assist motor learning tasks and that the adaptation to the task is dependent on the level of motivation.

Adaptive Hierarchical Control for the Muscle Strength Training of Stroke Survivors in Robot-Aided Upper-Limb Rehabilitation

Muscle strength training for stroke patients is of vital importance for helping survivors to progressively restore muscle strength and improve the performance of their activities in daily living (ADL). An adaptive hierarchical therapy control framework which integrates the patient's real biomechanical state estimation with task-performance quantitative evaluation is proposed. Firstly, a high-level progressive resistive supervisory controller is designed to determine the resistive force base for each training session based on the patient's online task-performance evaluation. Then, a low-level adaptive resistive force triggered controller is presented to further regulate the interactive resistive force corresponding to the patient's real-time biomechanical state – characterized by the patient's bio-damping and bio-stiffness in the course of one training session, so that the patient is challenged in a moderate but engaging and motivating way. Finally, a therapeutic robot system using a Barrett WAM™ compliant manipulator is set up. We recruited eighteen inpatient and outpatient stroke participants who were randomly allocated in experimental (robot-aided) and control (conventional physical therapy) groups and enrolled for sixteen weeks of progressive resistance training. The preliminary results show that the proposed therapy control strategies can enhance the recovery of strength and motor control ability.

Real-time closed-loop control of cognitive load in neurological patients during robot-assisted gait training

Cognitively challenging training sessions during robot-assisted gait training after stroke were shown to be key requirements for the success of rehabilitation. Despite a broad variability of cognitive impairments amongst the stroke population, current rehabilitation environments do not adapt to the cognitive capabilities of the patient, as cognitive load cannot be objectively assessed in real-time. We provided healthy subjects and stroke patients with a virtual task during robot-assisted gait training, which allowed modulating cognitive load by adapting the difficulty level of the task. We quantified the cognitive load of stroke patients by using psychophysiological measurements and performance data. In open-loop experiments with healthy subjects and stroke patients, we obtained training data for a linear, adaptive classifier that estimated the current cognitive load of patients in real-time. We verified our classification results via questionnaires and obtained 88% correct classification in healthy subjects and 75% in patients. Using the pre-trained, adaptive classifier, we closed the cognitive control loop around healthy subjects and stroke patients by automatically adapting the difficulty level of the virtual task in real-time such that patients were neither cognitively overloaded nor under-challenged.