Converging Clinical and Engineering Research on Neurorehabilitation II: Proceedings of the 3rd International Conference on NeuroRehabilitation (ICNR2016), October 18-21, 2016, Segovia, Spain
Abstract
The book reports on advanced topics in the areas of neurorehabilitation research and practice. It focuses on new methods for interfacing the human nervous system with electronic and mechatronic systems to restore or compensate impaired neural functions. Importantly, the book merges different perspectives, such as the clinical, neurophysiological, and bioengineering ones, to promote, feed and encourage collaborations between clinicians, neuroscientists and engineers. Based on the 2016 International Conference on Neurorehabilitation (ICNR 2016) held on October 18-21, 2016, in Segovia, Spain, this book covers various aspects of neurorehabilitation research and practice, including new insights into biomechanics, brain physiology, neuroplasticity, and brain damages and diseases, as well as innovative methods and technologies for studying and/or recovering brain function, from data mining to interface technologies and neuroprosthetics. In this way, it offers a concise, yet comprehensive reference guide to neurosurgeons, rehabilitation physicians, neurologists, and bioengineers. Moreover, by highlighting current challenges in understanding brain diseases as well as in the available technologies and their implementation, the book is also expected to foster new collaborations between the different groups, thus stimulating new ideas and research directions.
Chapters (100)
Improving motor function after stroke is an in an important area of research in neurorehabilitation. Clinical trials using non-invasive brain stimulation (NIBS) to improve rehabilitation outcome after stroke showed modest effect sizes or even lack of efficacy [1–3]. One important reason for this limited therapeutic success may be too simplistic “one hat fits it all” strategies, e.g. aiming at increasing excitability in the ipsilesional primary motor cortex [4] that disregard high interindividual variability in responses to NIBS protocols, even in healthy subjects [5]. Several strategies that have been recently developed to improve therapeutic effect size of NIBS during stroke neurorehabilitation will be detailed in this presentation.
By continuous pairing of user intend (interpreted with a Brain-computer-interface—BCI system) and artificial reproduction of that movement BCIs for neuromodulation follow the principle of Hebbian association that underlies motor learning leading to neuroplasticity and associated functional changes. In the current study, movement-related cortical potentials (MRCPs) were detected using elecroencephalography (EEG) during repetitions of foot dorsiflexion. Detection triggered a either a robotic actuator or a functional electrical stimulator timed so that the resulting afferent volley arrived at the peak negative phase of the MRCP. The tibialis anterior motor evoked potentials increased significantly following both types of interventions (p = 0.006) and there was no difference between their effectiveness (p = 0.344). Results have implications for the design of BCIs intended for neuromodulation.
The neurophysiological techniques that can induce plasticity or simply modulate cortical excitability or produce interference with normal brain activity and behavior are known as neuromodulation techniques. The interest in using neuromodulation techniques in neurorehabilitation has sharply increased in the last years. Recently we described that transcranial application of static magnetic field is able to interfere with brain activity.
Severe Spinal Cord Injury (SCI) alters the communication between supra-spinal centers and the sensorimotor networks coordinating limb movements, which are usually located below the injury. Epidural electrical stimulation of lumbar segments has shown the ability to enable descending motor control of the lower limbs in rodents and humans with severe paralysis. Using computational models and in vivo experiments in rodents, we found that EES facilitates motor control through the recruitment of muscle spindle feedback circuits. Stimulation protocols targeting these circuits allowed the selective modulation of synergistic muscle groups, both in rodents and non-human primates. This framework supports the design of stimulation strategies for humans.
The activity of corticospinal pathways is important in movement control and its plasticity is essential for learning motor skills and re-learning them after spinal cord injury (SCI) and other CNS disorders. After SCI, corticospinal excitability and connectivity diminish, as SCI disturbs supraspinal connections. However, this can be reversed. Thus, if there is a way to enhance the function of corticospinal pathways, motor function recovery may be improved. The central hypothesis of this study is that operant conditioning of the corticospinal motor evoked potential can strengthen corticospinal connectivity and thereby improve motor function recovery in people after incomplete SCI.
The aim of this study was to investigate the effects of a rehabilitation exercise for people with incomplete Spinal Cord Injury (iSCI), based on cycling and combined afferent electrical stimulation (ES-cycling), to normalize spinal activity in response to a plantar cutaneous stimulation. We studied Soleus H-reflex excitability following ipsilateral plantar electrical stimulation applied at 25–100 ms inter-stimulus intervals (ISI’s), on 13 non-injured subjects and 10 subjects with iSCI. Reflexes were tested before and after a 10 min session of ES-cycling to evaluate the effects of the exercise. Plantar-conditioned H-reflex modulation increased in the iSCI group after ES-cycling, compared to the limited modulation observed before the exercise. Conversely, the non-injured group presented pronounced modulation both before and after the exercise. We conclude that ES-cycling improved plantar-conditioned spinal neuronal excitability in subjects with iSCI. Results could be used in the design of more effective leg-cycling therapies, to promote central neuroplasticity and rehabilitation in lower limb muscle activity following iSCI.
Human spinal cord injuries (SCI) disrupt the pathways between the brain and spinal cord, resulting in substantial impairment and loss of function. We recorded surface electromyogram signals (sEMG) using grids of electrodes (8 × 8) applied on Biceps Brachii and Triceps Brachii muscles. We aimed to identify dysfunctional muscle activation in individuals with incomplete injuries of the cervical cord. We recorded sEMG and force from one SCI individual (Chronic, C5-C7, ASIA score D) and from a neurologically intact person during the generation of an isometric sinusoidal force trajectory (15s elbow flexion + 15s elbow extension). We found that the SCI subject was not able to follow the target force during elbow extension as precisely as in elbow flexion. Failure in tracking force was quantified using the root mean squared error between the target and generated forces. Our data suggest that C7 was the most affected spinal segment while the anatomical level had been diagnosed C5-C7. These data show the potential use of sEMG grid recording for localizing the motor lesion level within the spinal cord. Additional confirmatory studies are necessary to validate our results.
Magnetic stimulation techniques, either repetitive TMS (rTMS) or Static Magnetic Fields, allow to modulate brain activity through the skull in a non invasive and painless way. When rTMS is used, low frequencies of stimulation (≤1 Hz) produce inhibitory changes in excitability whilst higher rates (above 5 Hz) appear to produce increasing excitability Pascual-Leone et al. (J. Clin. Neurophysiol. 15(4):333–343, 1998) [1]. By using two different experimental approaches (anaestethetized cat and monkey) to study the early visual system, we show here that rTMS applied at low and high frequency has opposing effects on the EEG. These effects can be detected locally but also in a wider spatial extent. Further, we report data supporting the suppressive nature of the static magnetic stimulation. It supports the idea that static magnets could be used for different purposes ranging from experimental studies to clinical applications.
Brain stimulation therapies involve activating or touching the brain directly with electricity, magnets, or implants to treat different brain traumas or disorders. However, most studies do not take in account the relation between these therapies and the corresponding morphological and molecular changes in neurons. These changes depend on the external stimulation frequency and intensity, and are different in each subset of neurons, meaning an absence of neuronal excitability control. In this sense, the Axon Initial Segment is place of action potential initiation. This axonal domain has the property of changing its density of voltage gated ion channels, its position and length in response to different levels of stimuli. In this presentation we will review the molecular mechanisms that control neuronal excitability at the Axon Initial Segment and the different types of AIS plasticity that may explain the success or failure of therapies involving electrical stimulation of neurons.
We introduce and validate a novel measure of motor unit action potential (MUAP) variability in surface electromyograms (EMG) that are recorded during dynamic muscle contractions. This measure is fully automatic, builds on the motor unit spike trains as estimated by previously introduced Convolution Kernel Compensation method and allows tracking of MUAP variability for each individual motor unit separately. Preliminary tests on synthetic surface EMG signals demonstrate its high accuracy and capability of identifying cyclostationary changes of MUAP shapes. This measure represents the first, but very important step towards motor unit identification in dynamic muscle contractions.
Essential tremor and Parkinson’s disease cause abnormal oscillatory activity in a variety of brain structures that is transmitted to spinal motoneurons and generates tremor. Because the motoneuron pool integrates synaptic inputs from descending and spinal circuits, the decoding of its activity provides a view on all the neural pathways involved in tremor generation. We investigated tremor mechanisms by analyzing the behavior of populations of motoneurons within a single muscle, across antagonist muscle pairs, and in relation to cortical activity. We observed that tremor is caused by a common cortical input projected to all motoneurons. We also found that spinal reflex pathways contribute fundamentally to shaping tremor properties. We posit that although ET and PD tremor are centrally generated, tremor properties are strongly determined by the interaction between descending and afferent inputs to the motoneuron pool.
We investigated the possibility to identify motor units (MUs) with high-density surface electromyography (HDEMG) over experimental sessions in different days. 10 subjects performed submaximal knee extensions across three sessions in three days separated by one week, while EMG was recorded from the vastus medialis muscle with high-density electrode grids. The shapes of the MU action potentials (MUAPs) over multiple channels extracted from HDEMG decomposition were matched across sessions by cross-correlation. Forty and twenty percent of the MUs decomposed could be tracked across two and three sessions, respectively (average cross correlation 0.85 ± 0.04). The estimated properties of the matched motor units were similar across the sessions. For example, mean discharge rate and recruitment thresholds were measured with an intra-class correlation coefficient (ICCs) >0.80. These results strongly suggest that the same MUs were indeed identified across sessions. This possibility will allow monitoring changes in MU properties following interventions or during the progression of neuromuscular disorders.
This presentation will cover the methods used to investigate neuronal circuitries between peripheral receptors and skeletal muscles in human subjects. There are a number of problems regarding reflex studies using experimental animals. There are also problems in the recording and analysis aspects of these experiments. To overcome these problems we have utilized precisely-controlled mechanical or electrical stimuli to activate receptors and single motor units from human muscles. We also used classical and novel methods to analyze the results to indicate neuronal networks.
In this study, we aimed to use both the probability- and frequency-based analyses methods simultaneously to examine cutaneous silent period induced by strong electrical currents. Subjects were asked to contract their hand muscles so that single motor unit discharged at a rate of approximately 8 Hz. Strong electrical stimuli were delivered to the back of the hand and induced cutaneous silent period in all units. It was found that the duration of the cutaneous silent period (CSP) was significantly longer when the same data were analysed using frequency-based analysis method compared with the probability-based methods.
In some pathologies with impaired control gait, the dual task execution even worse these gait alterations, as in Parkinson’s Disease (PD) occurs. The aim of this paper is to present the effects of a rehabilitation program that integrates dual tasks in gait rehabilitation and prepare patients to move functionally in complex environments in a person with PD. 3D photogrammetry system and two force platform were used for the registering of kinematic and kinetic parameters, respectively. The main variables that have improved after the proposed program are: speed gait (54.87 %), step length (52.01 %), stride length (43.48 %), double support time (37.36 %), flexo-extension hip range of motion (36.08 %) and vertical reaction force at the heel contact moment (23 %). The inclusion of complex environments representative of daily life activities in gait rehabilitation is useful for improving the performance of walking when parkinsonian gait impairment is suffered.
Analysis of Ground Reaction Forces is a non-invasive useful tool for the analysis of gait in post-stroke patients. The analysis of ground reactions forces in post-stroke patients is difficult due to the large variability of the shape of the force curves, making very difficult an analysis of the different events of the gait pattern, as has been common practice for years. Functional Principal Components, allows the analysis of the fitting coefficients to a sample of curves better describing the variability of the whole set. In this contribution, this type of analysis has been made, and the reliability of the measurements has been obtained, demonstrating the usefulness of the approach.
We aim to optimize post-stroke motor control rehabilitation therapy via integration of cardiovascular activity. Although stroke therapy ranges from months to years for some stroke patients, the majority of neurological recovery occurs within the first three months. Afterwards, neurological recovery occurs at a reduced rate for a period of up to one year. Current studies report that cardiovascular activity increases synaptic plasticity by affecting synaptic structure and potentiating synaptic strength, strengthening neurogenesis, metabolism and vascular function. Integrating cardiovascular activity in post-stroke motor control rehabilitation therapy may restore two-way communication between the central nervous system and extremities via growth of alternative central nervous system pathways; thus, resulting in improved motor control in both upper and lower extremities. We hypothesize that the neurotrophic factors engendered by cardiovascular activity significantly fortify descending motor pathways. In part one, functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) quantify the level of brain activity resulting from upper extremity movement and lower extremity movement in two cohorts, each cohort consisting of four members. In part two, cohort one will perform 30 min of cardiovascular activity prior to 30 min of upper and lower extremity strength training five times a week for a total of three months. During this period, cohort two will perform only 30 min of upper and lower extremity strength training five times a week. In part three, fMRI and EEG will quantify the level of brain activity resulting from upper extremity movement and lower extremity movement in both cohorts. After data collection, juxtaposition of functional magnetic resonance images in conjunction with electroencephalograms during day one and during day ninety will occur in order to quantify the significance of cardiovascular activity integration in post-stroke motor control rehabilitation therapy.
This paper investigates different possible stability criteria of human walking and their correlation to the nature of the real motion. The analysis is performed using human motion capture data and a precise dynamic 3D model of human walking with 34 DOF and physical contact constraints.
Despite frequent application of ankle-foot orthoses (AFOs), little scientific evidence is available to guide AFO-provision early after stroke. A randomized controlled trial was conducted to study the effects of AFO-provision in (sub-) acute stroke patients. Primary aim: to study effects of the actual provision of AFOs on functional outcomes. Secondary aim: to study whether the point in time at which an AFO is provided (early (week 1) or delayed (week 9)), influences these effects. Thirty-three subjects were included and walking speed, balance (Berg Balance Scale, BBS) and independence of walking (Functional Ambulation Categories, FAC) were measured. Positive effects of AFO-provision were found two weeks after provision, both when provided early (significant effects on all outcome measures) or late (BBS p = 0.011, FAC p = 0.008). Comparing the early and delayed group showed that early provision resulted in extra improvements on BBS (+5.1 points, p = 0.002) compared to late provision.
We developed a robot, CUREs (Chulalongkorn University Rehabilitation robotic Exoskeleton system), for upper extremity rehabilitation. Five subacute stroke patients participated in this pilot study. All patients had severe upper extremity weakness (Brunnstrom stage I and II, Fugl–Meyer Assessment Upper Extremity score, motor function, are from 4 to 10). They participated in 30 min of conventional upper extremity training and 30 min of robotic training, 5 days per week for 2 consecutive weeks. The Fugl–Meyer Assessment Upper Extremity Scale was improved after 2 weeks training in all participants. However, the Motor Assessment Scale was not changed. In the future, we plan to conduct a 4 week randomized–controlled trial study to compare the rehabilitation outcome between CUREs robot with conventional therapy in stroke patient.
A novel neuromechanical model for investigating patient-specific lower limb movement dynamics after spinal cord injury (SCI) is presented. The model is designed in joint space and takes into account muscle neuromechanics. Application of the model is demonstrated in the application of a balancing task.
Exo-Glove is a soft wearable robotic hand to assist hand function of people who have paralysis of the hands. Exo-Glove is compactly structured with soft fabrics and adapts an under-actuation concept. Pinch performance is defined, and the variation of the pinching performance with Exo-Glove with respect to the tendon route design is shown through an experiment. A subject with spinal cord injury participated in the experiment. As shown by the experimental result, Exo-Glove provides adequate pinching performance, and the tendon routing of Exo-Glove severely affects its pinching performance.
This work presents the translation from a humanoid robotic hand to a prosthetic prototype and its first evaluation in a set of 9 persons with amputation. The Pisa/IIT SoftHand is an underactuated hand built on the neuroscientific principle of motor synergies enabling it to perform natural, human-like movements and mold around grasped objects with minimal control input. These features motivated the development of the SoftHand Pro, a prosthetic version of the SoftHand built to interface with a prosthetic socket. The results of the preliminary testing of the SoftHand Pro showed it to be a highly functional design with an intuitive control system. Present results warrant further testing to develop the SoftHand Pro.
This paper presents a wearable rehabilitation exoskeleton for the elbow joint with two degrees of freedom (DOF), flexion-extension and pronation-supination, actuated with Shape Memory Alloy (SMA) based actuators. Due to the actuation system, the proposed exoskeleton presents a light weight, low noise, and is designed in a simple structure. The number of actuators and the preliminary designed was calculated based on a biomechanical simulation with a specific category of patients.
In this paper we validate a gravity compensation algorithm for a novel shoulder-elbow exoskeleton, aimed at compensating the device weight during the execution of rehabilitation exercises. Along with the description of the exoskeleton, we present the results of the validation of the algorithm on data acquired in static and dynamic trials, in unloaded conditions. Results showed good performance of the algorithm in calculating the gravity torque of each joint, suggesting the possibility to implement rehabilitation exercises in which a specific net amount of assistive torque is provided to the user’s joints.
This paper reports two experiments that we recently conducted. The first established that upper-limb rhythmic and discrete movements are differentially affected after a stroke, reinforcing the statement that these two movements form two fundamentally different motor primitives. The second focused on the development of a specific robot-assisted therapy for rhythmic movements, targeting assistance-as-needed. We claim that rhythmic movements deserve such as specific therapy, since they are part of the daily motor repertoire and belong to a different class of movements as the discrete ones.
Upper limb motor impairment often causes long-term disability in stroke patients and implies limitations in activities of daily living. Several studies tested robotic devices for proximal or distal upper limb rehabilitation and reported results principally focused on specific treated district without significant global effects. We propose a novel approach that integrates a hand distal effector in task-oriented arm training for upper limb functional rehabilitation. Four chronic stroke patients underwent to an intensive rehabilitative treatment using a robotic device that provides arm weight support and assistance of the hand closing/opening within specific setting in virtual reality. After treatment improvements in upper limb functional scales and in kinematic and pressure assessments were observed, highlighting effects on global upper limb motor performance and distal motor control. Furthermore a decrease in resting motor threshold and a reduction in silent period recorded from unaffected hemisphere were evident, suggesting a potential cortical reorganization.
Stroke survivors often experience long-term upper extremity impairment. This can greatly impair activities of daily living. Repetitive practice of task-oriented training is thought to be beneficial for rehabilitation, but task practice can be limited by poor hand motor control. The objective of this study is to use a soft robotic glove to facilitate hand therapy of stroke survivors. Participants, all of whom will have experienced a stroke less than one year prior to enrollment, will complete 18 one-hour sessions, consisting of 20 min of passive stretching followed by 40 min of active task practice while wearing the device. To date two subjects have finished all training sessions. The device was well tolerated and improvements were observed. The study is ongoing.
When a lower extremity is lost the primary and obvious effect is the loss of the mechanical support of weight bearing when standing and walking. Additionally there are loses of sensation for the lost leg, the sensation for interaction of the leg with its environment as well the central nervous system motor control of the leg in response to these sensations and for the purpose of locomotion. All of these losses affect quality of live negatively in many different ways, a parameter that is accepted valid for economic decision making since its improvements are highly associates with increased mobility. Reestablishing the central nervous system connection of bionic prosthesis is the next lower limb prosthetic challenge. Here we present the first cases of amputees gaining voluntary control of commercially available microprocessor controlled lower limb prosthesis using implantable myoelectric sensors.
In an attempt to overcome the several limitations of currently available/investigated human-machine interfaces (HMI) for the control of robotic hand prostheses, we propose a new HMI exploiting the magnetic field produced by magnets implanted in the muscles. As a magnet is implanted in a muscle it will travel with it, and its localization could provide a direct measure of the contraction/elongation of that muscle, which is voluntarily controlled by the individual. Here we present a proof of concept of a single magnet localizer, which computes on-line the position of a magnet in a certain workspace. In particular, the system comprises a pair of magnetic field sensors mounted on custom printed circuit boards, and an algorithm that resolves the inverse magnetic problem using the magnetic dipole model. The accuracy and the repeatability of our system were evaluated using six miniature magnets. Ongoing results suggest that the envisioned system is viable.
User Centered Design of bionic and assistive devices is growing in importance as many technologies are now moving from lab concepts to certified medical products for use in daily life. The enthusiasm to develop new technologies often focuses on the scientific requirements but often very practical user requirements are over looked. This presentation discusses the role of user centered design in bionics development and how this relates to usability in use. The presentation frames the importance of usability and user centered design on technology acceptance, generally by users, but also with focus on technology acceptance and adoption by older age adults.
Powered lower limb prosthesis are facing energy and efficiency challenges. This article presents an investigation into reducing the energy losses and increasing the efficiencies of energy regeneration for a powered prosthetic knee during level ground walking. The results showed that the regeneration and overall system efficiencies would dramatically increase if the negative mechanical load in the braking quadrants are within the regenerative zone of the motor. This approach reduced the energy losses in the stance and swing phases and increased the possibility of harvesting more negative mechanical energy during level ground walking.
A neuromuscular model (NMC) presented by H. Geyer and extended by S. Song shows very interesting similarities with real human locomotion. The model uses a combination of reflex loops to generate stable locomotion and is able to cope with external disturbances and adapt to different conditions. However, to our knowledge no works exist on the capability of the model to handle sensory noise. In this paper, we present a method for designing Central Pattern Generators (CPG) as feedback predictors, which can be used to handle large amount of sensory noise. We show that the whole system (NMC + CPG) is able to cope with a very large amount of noise, much larger than what the original system (NMC) could handle.
This study combines experimental-based and model-based methodologies for accessing in vivo musculoskeletal function in healthy individuals. We use ultrasound and dynamometer technologies to derive subject-specific muscle parameters including muscle isometric force, optimal fiber length and tendon slack length. We then assess the impact of subject-specificity on the electromyography-driven simulation of walking of the composite musculoskeletal system.
Lower-limb exoskeletons are a promising tool for restoring or augmenting locomotion performance. While engineering advances have led to marked improvements on the machine side of the human machine interface, fundamental aspects of the physiological response of the human user remain unknown—especially at the level of individual leg muscles. One complication is that it is difficult to make direct measurements from muscles in humans without being invasive. Here we offer a novel benchtop approach by introducing a ‘smart’ robotic interface into the framework of biological muscle-tendon work loop experiments in order to simulate the local dynamical environment muscles experience in vivo during locomotion with exoskeleton assistance. Using this framework we demonstrate that providing force in parallel with a muscle-tendon using an ‘exo-tendon’ can have unintended consequences, disrupting the ‘tuned’ spring-like mechanics of the underlying biological muscle tendon unit.
Positive assistive and therapeutic effects of FES have been proved for post-stroke subjects suffering from foot drop. However, the published studies are very heterogeneous in terms of methodology, therapy duration, session duration and session frequency, where most studies rely on intensive FES-based therapy. In this document a clinical protocol is proposed for analyzing the effect of medium and low intensity of FES-based therapy. The protocol is designed to be used with a surface multi-field FES system and it is based on available studies on literature and preliminary results with chronic post-stroke subjects. The proposed clinical trials could help determining the minimum necessary FES-based therapy intensity for obtaining positive therapeutic results.
This paper introduces a quasi-static method for controlling reaching movements of a paralyzed human arm using FES. A subject-specific model was estimated from experimental data collected from a single subject with a high cervical spinal cord injury who uses a functional electrical stimulation neuroprosthesis. We tested a controller based on this model that maps desired shoulder and elbow joint positions to muscle stimulation commands. The controller tracked the desired lateral direction of hand movements well but did not track forward/backward or vertical movements well. The lateral position of the hand was not greatly affected by the direction of movement given identical stimulation commands, which supports the hypothesis that velocity information is not required for tracking slow hand movements.
The combined use of functional electrical stimulation and robotic exoskeleton in a hybrid rehabilitation system represents a promising research field for rehabilitation of the motor functions after stroke. In this work, we report the results obtained in a study carried out with a hybrid robotic system for reaching rehabilitation. The system was tested in two sessions with one chronic stroke subject.
The walking duration of gait restoration systems that use functional electrical stimulation (FES) is severely limited by the rapid onset of muscle fatigue. Alternatively, fully actuated orthoses can also be employed to restore walking in paraplegia. However, due to the high power consumption of electric motors the walking duration of such devices are limited by the charge of the batteries. This paper proposes that a hybrid system, which uses FES and an actuated orthosis, is capable of achieving greater walking durations than an FES only system and more energetically efficient than a lower-limb exoskeleton. This is illustrated through results of optimizations of a musculoskeletal gait model for three actuation cases: FES only, electric motors only, and a hybrid system. The presented results illustrate that a hybrid system may be capable of greater walking durations than FES-based systems while using half the energy of a lower-limb exoskeleton.
In this communication we will present some general approaches developed in our team and recent results involving inertial measurement units (IMU) in FES-controllers for lower-limb movement assistance in different situations of sensory-motor deficiencies. We will discuss some of the challenges to met in order to achieve robust adaptive controllers.
This contribution describes a method for real-time analysis of muscle activity over the gait cycle while simultaneously applying Functional Electrical Stimulation (FES) to the assessed muscles. Inertial sensors at the foot are used for real-time gait phase detection in order to synchronize the stimulation with the gait. An EMG analysis has been performed to cancel out stimulation artifacts, to filter the data and to extract the voluntary EMG activity. This corrected EMG signal has been rectified and low-pass filtered to produce an envelope that was parameterized as a function of the Gait Cycle Percentage (GCP). The volitional EMG activity profile has been averaged over five strides and can be presented as a moving average over the exercise duration. Initial evaluation with healthy subjects showed that this procedure is feasible to detect the expected volitional muscle activity profiles also during active electrical stimulation.
This paper presents a control approach of a lower limb exoskeleton by modulating the original impedance of the wearer’ swinging leg to reduce the muscular efforts. The proposed method ensures compensation of the damping and gravity effects of the wearer’s leg to desired levels offering though the possibility to adapt the lower limb wearer’s impedance as a function of the gait phase evolution. The performance of the proposed approach is evaluated by carrying out experiments with two healthy subjects. The EMG activities of the extensor muscles spanning the knee-joint are used as assesment criteria. The results show that the muscular activities required to perform the same flexion/extension movements are effectively reduced by providing lower desired joint damping and stiffness with exoskeleton’s assistance.
Phantom limb pain (PLP) is a frequent consequence of amputation, and it is notoriously difficult to treat. Despite isolated reports of success, no medical/non-medical treatments have been beneficial on more than a temporary basis. Recent evidence suggests that the pathophysiological mechanism of PLP is related to neuroplastic changes in the cortex. While the majority of the treatments currently offered seek to actively suppress the pain, the EU consortium ‘EPIONE’ will challenge the status-quo of PLP treatment by actively creating natural, meaningful sensations that will restore the neuroplastic changes in the cortex and thereby control and alleviate pain. The consortium will develop dedicated, technological solutions and test these in multi-center clinical trials within Europe and the US.
Transversal intrafascicular multichannel electrodes (TIME) have been developed to interface with peripheral nerves after upper limb amputation. Intended use is the electrical stimulation of the median and ulnar nerve to deliver sensory feedback during phantom limb pain treatment and artificial hand control. Miniaturized electrode arrays were developed on polyimide substrates with thin film metallization using sputtered iridium oxide as electrode coating. Here, we report on the essential requirements including biocompatibility, mechanical and stimulation stability that have been investigated before permission was granted by the legal authorities to conduct subchronic first-in-man clinical trials. Explants have been investigated to identify possible first failure points and optimize the devices for chronic implantation.
The aim of this study was to evaluate neural coupling patterns in schizophrenia (SCH) patients and healthy controls during an auditory oddball task. Two measures of functional connectivity were applied to 28 SCH patients and 51 healthy controls to characterize electroencephalographic (EEG) activity. Specifically, magnitude squared coherence (MSC) and the imaginary part of coherency (ICOH) were computed for five frequency bands: theta, alpha, beta-1, beta-2 and gamma. The results showed a statistically significant modulation increase in MSC and ICOH for controls with respect to SCH in the theta band, and a decrease in ICOH for the beta-2 band. Furthermore, controls showed more significant changes from the baseline and active task windows than SCH patients. Our findings suggest that SCH patients show coupling abnormalities during an auditory oddball task compared to healthy controls.
High Frequency Oscillations (HFOs, >80 Hz) are events that have been linked to the seizure onset zone (SOZ). Few studies have identified HFOs in noninvasive EEG and MEG signals due to the high signal-to-noise ratio (SNR) required, but beamforming-based virtual sensors (VS) can increase SNR. We computed the beamforming-VS as a grid inside the brain volume model for 200 s MEG signals. Events of interest (EOIs) exceeding a threshold were automatically determined as well as the area of interest, where EOIs occurred more frequently. Finally HFOs inside the area of interest were selected and compared with simultaneous iEEG recordings.
Detection of an increment in stress levels is a step towards improving the quality of people’s lives, especially in the case of people with intellectual disabilities, as they have fewer resources to deal with this situation. This paper presents a biophysical stress classification system that is able to classify the detected stress situations at three intensity levels: low, medium and high. Furthermore, the system distinguishes between continued stress and a momentary alert depending on the subject’s arousal. The system uses two non-invasive physiological signals for the classification: the galvanic skin response and the heart rate variability. The experiment shows that the proposed system is able to detect and classify the different stress states achieving an accuracy of 97.5 % with a 0.9 % FN rate.
A system has been developed to detect postures and movements of people, using the skeleton information provided by the OpenNI library. A supervised learning approach has been used for generating static posture classifier models. In the case of movements, the focus has been done in clustering techniques. These models are included as part of the system software once generated, which reacts to postures and gestures made by any user. The automatic detection of postures is interesting for many applications, such as medical applications or intelligent interaction based on computer vision.
Wearable technologies and low-cost hardware platforms for physiological data sensing are rapidly growing. However, biosignal sources pose usability and acceptability challenges, which can negativelly affect the users’ experience and performance when working with systems that incorporate physiological sensing. This paper analyses and presents practical considerations about the use of biosignals, highlighting aspects that practitioners and novice researchers should be aware of when designing their systems, especially in real-world use cases. Common biosignal modalities are characterised according to a proposed multidimensional taxonomy (C.A.O.S.), devised to characterize physiological data sources and balancing expectations around their use in human-computer interaction.
In order to be useful in daily life, lower limb exoskeletons have to be able to provide support not only for nominal situations, such as level ground walking, but also for the recovery from extreme situations. In this paper, we investigate which torques a lower leg exoskeleton would have to produce in order to allow a person to recover from large perturbations or pushes that may occur while walking. We propose a model-based optimization approach that takes into account dynamic models of the human and the exoskeleton as well as experimental data of humans being pushed. Using optimal control and a least squares objective function we compute the joint torques that exoskeletons of different masses and mass distributions would have to produce in order to make the person follow the recorded recovery trajectories of healthy subjects and which loads would occur in the structure. The results of these computations can serve as guidelines for the design of future lower limb exoskeletons.
We summarize the achievements of the EU FP7 funded project CONTRAST on cognitive rehabilitation after stroke. We developed a neuropsychological algorithm to assign patients to specific, personalized neurofeedback training to improve cognitive function, namely attention, declarative memory, inhibitory control, and working memory. Further, BCI technology was integrated into a remote control set-up, such that therapists can supervise simultaneously multiple patients at their home during BCI-based neurofeedback training. Phase I studies with subacute and chronic stroke patients demonstrated the potential of our approach such that patients were able to learn regulation of the respective brain activity and improved in the targeted cognitive function. Phase II studies are necessary to consolidate our findings.
Alzheimer’s Disease (AD) produce important clues in speech, which may be used in pathology monitoring and grading. The study is intended to model which aspects of articulation may be affected more, and to which extent. Comparing speech from AD patients versus control subjects produced in animal naming, certain characteristics may be observed, encoding fluency and formant dynamics. Kullback-Leibler Divergence from formant velocity distributions is sought to work as a possible biomarker in AD monitoring tasks.
This paper investigates cortical responses to vibro-tactile stimuli. EEG was recorded in two conditions: when vibrations were applied focally on the muscle during relaxation and during muscle contraction. Mu and beta waves analysis of the EEG signals suggest that vibrations applied before the contraction increases the stretch of the muscle, thus improving its output performance. Further analysis of the vibrations applied during the muscle contraction shows cortical activation while modulating vibro-tactile stimuli to stabilise muscle performance.
The role of attention in formulating the input-signals to the CNS toward enhancing the motor-control ability in human is unclear. Here we hypothesized that the distance between the arms in alignment to the frontal center of a person, and the voluntary shifting of his visual attention play roles in enhancing the internal model and the body-control ability. To examine this, six participants were introduced to dual-steering-device. Using the device, we can modulate the participant’s visual attentions and arms distance while performing various tasks. Major muscles and brain activities of the participants were monitored using EMG, and fNIRS. The results were compatible with our hypothesis: users could inhibit muscular activities in the passive movements with increasing distance of the arms and with a visual focus on the inhibited arm. We believe that this study can add important contributing factors in designing rehabilitation program by adjusting the possible input-combination to enhance the internal-model.
Motor imagery based Brain-Computer Interface (BCI) utilizes an electrophysiological phenomenon of EEG power decrease in alpha frequency band, but its larger inter-subject variability limits the practical use. Here we tested three types of visual feedback objects in BCI from abstract to realistic scenarios during motor imagery to see its effect on self-induced changes of EEG power decrease. Double case study in hemiplegic stroke participants was also conducted to check its feasibility as neuro-facilitatory technique on the motor system. We found that a first person perspective of realistic visual feedback, which copies the participant’s mental image, assisted the user to perform motor imagery resulting in generation of large EEG power decrease. The same result was found also in hemiplegic stroke patients. This study has clear implications for both the mechanism of mental process of motor imagery and the influence of feedback type on BCI performance.
In this work the authors investigated whether the muscle synergies concept could improve the isometric hand force estimation. Electromyographic (EMG) activity from 9 arm muscles and hand forces applied at the Light-Exos Exoskeleton end-effector were recorded during isometric contractions in several workspace points lying on the parasagittal plane crossing the shoulder joint. The muscle synergies were extracted in two different ways according to the statements that the muscle primitives are ‘Arm Pose Related’ or ‘Arm Pose Shared’. From the pre-processed EMG signals the authors then estimated the hand forces using three methods. The results showed that the muscle synergy concept improves the isometric force estimation paving the way for a synergy-based myoelectric control.
Muscle synergy interprets the neural strategy adopted by the central nervous system (CNS) to simplify the coordination of muscles recruitments when performing useful movements. The computational mechanism of defining the optimal muscle combinations, however, still debatable. Muscle synergy deals with muscle activations pattern and time-dependent variables. The synergy space defines the suitable combinations of muscles, and time-dependent variables vary in lower-dimensional space to drive the behavior. In this study, we investigated the role of the CNS to optimize muscle patterns when performing skilled behavior. We introduced two synergy indices: the synergy stability index that indicates the similarity of the recruited synergies, and the synergy coordination index that indicates the size of the synergy space. The results on automatic posture response experiments on seven healthy participants show that both indices are positively correlated with the overall balance skill of the participants. Results suggest the optimal mechanisms adopted by the CNS to recruit muscles.
Recent studies endorse the use of robotic and virtual reality (VR) systems for rehabilitation. Myoelectric (EMG) signals have been used for prosthetic control but their application to rehabilitation has been limited so far. Here we present a novel approach using an EMG controlled VR interface to test the synergistic organization of the neural control of arm movements in healthy subjects. EMG control offers the possibility to manipulate visual feedback according to the subject’s muscle activity and to test effects of simulated interventions on the human neuromuscular system that are either compatible or incompatible with the synergies. Such EMG controlled VR interface may open up new possibilities for rehabilitation as it offers the possibility to provide assistance tailored to the individual changes in synergistic organization.
Correct classification of motor imagery tasks is not the only requirement of a Brain-Computer Interface (BCI) based on Sensorimotor Rhythms (SMR). Indeed, a SMR-BCI controlling an external device (e.g., robotic prostheses) needs to reliably detect even if the user is in the so-called Intentional Non-Control (INC) state. In this work, we propose a novel approach to online detect INC and thus, to reduce undesired delivered commands during SMR-BCI operations. Results with six healthy subjects show that the proposed INC detection framework does not affect the online BCI performance and, more importantly, it reduces the number of unintentionally delivered BCI commands with respect to a traditional approach (in average 42.7 ± 13.76 % less).
The recruitment of large nerve fibers before small fibers is an issue in many neuromodulation applications. It may, e.g., cause the larynx to be activated by vagus nerve stimulation (VNS) at lower thresholds than required for activation of clinically relevant fibers, such as baroreceptors. We here present results from two animals, indicating that a custom waveform, combining known techniques, can activate the baroreflex with suppressed laryngeal activation, as compared to rectangular pulses. Such selective baroreflex activation with tolerable side-effects could potentially help advance VNS as a treatment modality in resistant hypertension.
The purpose of this work is to evaluate the effect of the electrode size on recorded High Density surface electromyography (HDsEMG) maps. We recorded the sEMG signals using a grid of electrodes (16 × 8) placed on biceps brachii (5 mm inter electrode distance, IED). Each instantaneous map was interpolated using the 2D Sinc function to obtain a “continuous” map (CM, with 10,000 samp/m). To simulate acquisition with different electrode sizes, each CM was filtered with a circular averaging filter (2–10 mm diameter) and re-sampled with 10 mm IED. Inverse transfer function to compensate the effect of electrode size was applied on each map. The results suggest that greater electrode size implies higher error of power (RMS2) of each map values with respect to the CM. The error of power estimation introduced by electrode size is smaller than the error due to IED.
Performance dependent reward activates the striatum, a key region of the reward system. However, stroke patients were identified to show reduced brain activations to rewarding feedback in cognitive tasks when compared to healthy age-matched controls. This was reflected in impaired reinforcement learning. Whether their response to reward derived from preceding motor performance is also reduced, is, however, still unknown. Using functional magnetic resonance imaging, striatal activity linked to performance dependent monetary reward was measured during the training of a repetitive arc-tracking task. Pilot results of nine stroke patients and nine age-matched healthy individuals point towards a tendency for reduced responsiveness of ventral parts of the striatum in stroke patients, while the dorsal striatum, although to a smaller extent, shows an opposite trend. This is of particular interest as ventral striatal activation was found to be the key factor for successful overnight consolidation in an earlier study using a similar task.
Individuals with motor impairments typically walk at much slower speeds than their unimpaired counterparts, yet their gait data is still evaluated against the relatively faster gait of healthy subjects. Therefore a good understanding of unimpaired gait at extremely slow speeds is needed for comparison. Studies have shown that walking at very slow speeds is quantitatively different from self-selected walking speed. These modifications can be observed at different levels (kinetic, kinematic, electromyographic). In order to better understand the changes in walking at extremely slow speeds, we recorded seven subjects walking at their preferred speed and at speeds ranging from 0.11 m/s to 0.61 m/s. In this study, we analyzed changes in muscle activations and quantified their variability using the Pearson correlation coefficient. Confirming previous observation, we show that both the inter- and intra-subject variability of muscle activities increases with decreases in walking speed, with a more pronounced effect for proximal muscles. The inter-subject correlation of muscle activities also suggests a modular organization of muscle activities in three functional blocks at normal speed. This modular organization vanishes with decreasing walking speed following a proximo-distal gradient.
In recent years, robot-assisted exercise has been used as a therapeutic intervention to improve upper limb function of stroke survivors. The purpose of this study is to investigate the relation between functional movement and kinematics in robot assisted reach exercise for chronic stroke survivors. The robot assisted reach exercise was performed in a test bed during 40 min for 16 chronic stroke survivors. The training test bed consisted of one Whole Arm Manipulator (WAM) and one projective display device. The Action Research Arm Test (ARAT) was used to measure functional movement and kinematics (movement time) was analyzed based on the performance of a reaching movement toward 3 targets. ARAT was positively correlated with movement time in 3 targets. In addition, stepwise linear regression analysis revealed that the movement time toward a contralateral target (target 1) was the explanatory variable closely associated with the functional movement, i.e., ARAT.
The Activities of Daily Living (ADL) are used to refer the daily self care activities. Stroke survivors usually experience an impairment in the functionality limbs being affected their independent life. A complete assessment of a patient implies functional and analytic evaluation. However, the joints range cannot be always measured due to the complexity of the patient limbs. The aim of this paper is to present the results acquired from a new tool being able to quantify objectively the functional level of joints movement during ADLs. Twelve healthy subjects had participated in this trial. Four ADL were performed to measure and evaluate the maximum range reached in those activities.
This work investigates adaptation of stepping responses when neurologically impaired subject is subjected to repeated perturbations while walking. Balance assessment robot in combination with treadmill (BAR-TM) and Optitrack camera were used to deliver well repeated perturbations and to track foot placement respectively. Results show that when neurologically impaired subject is repeatedly subjected to identical perturbations we may expect adaptation period at the beginning of session when gait parameters change before they settle. We could speculate that in order to evoke repeatable postural responses during balance training patients should be subjected to limited number of exercises that would allow them to surpass adaptation period and focus on repeating the movement after it stabilizes.
This work studied different electrode configurations and processing windows for detecting the intention of pedaling initiation. Furthermore, data were pseudo-online analyzed. The main goal was to find alterations in the mu and beta frequency bands where event-related synchronization and desynchronization (ERS/ERD) is produced. The results show an improvement using time before and after the movement onset rather than until the movement onset.
The optimal number of EEG channels is a controversial issue for motor imagery based BCIs for stroke rehabilitation. In this study, we compared the BCI performance with 63, 27 and 16 channels of EEG on three stroke patients across 10 to 24 sessions, and demonstrated that the 16 channels montage yields similar classification error (21.3 ± 11.6, 10.5 ± 6.6 and 16.0 ± 12.6 % for these patients respectively) to montages with larger number of channels. This is important for practical applications in stroke rehabilitation, since fewer channels means lower cost, less preparation time and easier maintenance.
Is presented a non-supervised method for feature selection based on similarity index, which is applied in a brain-computer interface (BCI) to recognize gait preparation/stops. Maximal information compression index is here used to obtain redundancies, while representation entropy value is employed to find the feature vectors with high entropy. EEG signals of six subjects were acquired on the primary cortex during walking, in order to evaluate this approach in a BCI. The maximum accuracy was 55 and 85 % to recognize gait preparation/stops, respectively. Thus, this method can be used in a BCI to improve the time delay during dimensionality reduction.
Brain-Computer Interfaces (BCI’s) aim to create a channel of communication between a person and a device without any physical action on the environment by the user. There are several BCI systems, some of them focusing on motor actions by the user. Various techniques exist for such BCI systems, such as extraction of the power in different frequency bands. These techniques have proven to be useful but require extensive training by the end user and the creation of new models every time other user intends to use the system. In this paper we present a new method based on spectral entropy to detect changes in motor area and their possible application in the detection of imagined movement. The successes obtained with this technique is about 76 %.
Changes in cortical signals related to motor planning processes have been widely studied in the past. However, no studies so far have investigated the intra-subject differences in these signals between analytical and coordinated upper limb and lower limb cue-based movements. Here, data from healthy subjects is analyzed to research this aspect. The statistical analysis carried out with data from 7 subjects indicates that statistically significant differences were observed between premotor cortical activities of upper and lower limb movements. Specifically, higher amplitudes of the contingent negative variation pattern were observed for lower-limb tasks. Such results may be due to complexity in movement task planning. BCI devices could take advantage and be improved with the knowledge provided.
Linear decoders have been successfully applied to extract human limbs kinematics from low-frequency cortical modulations. In this, intermediate descending motor pathways are absorbed in the regression. Here we propose the use of linear decoders to map cortical function to the spinal function (muscle primitive-level), thus shortening the transmission distance and reducing the dimensionality of the decoding of a large number of muscles. Our first results show that it is possible to accurately reconstruct muscle primitives computed from knee flexion-extension and to successfully detect muscle activity during repetitive cyclic movements.
Defining rehabilitation robots behavior during training exercises is necessary for their optimum performance. In this work, a comprehensive training mode for an upper limb rehabilitation robot, the UHP (Universal Haptic Pantograph) is presented. The proposed mode, which is divided in three phases, focuses on upper limb extension allowing the task to be adapted to the recovery state of the patient and ensuring exercise completion. Experimental validation of the training mode is carried out with the upper limb rehabilitation robot UHP.
In the field of rehabilitation robotics, few researchers have been focusing on the problem of controlling motor coordination in post-stroke patients. Studies on coordination learning, when the robotic devices act at the joint level on multiple interaction points, as in the case of exoskeletons, are lacking. For this reason, we studied on 10 healthy subjects the possibility of learning a non-natural inter-joint coordination while performing a pointing task. This coordination was induced by a 4-DOF robotic exoskeleton, applying resistive force fields at the joint level. Preliminary results showed the capability of our controller to modify human healthy natural coordination after exposition to the fields and generalization of these effects to movements which were never exposed to these constraints.
Stroke is one of the leading causes of long-term disability in the United States with many survivors unable to participate in rehabilitation. We have developed a novel rehabilitation paradigm for the upper extremity (UE) that uses a Tongue Drive System (TDS) to control a robotic device (HandMentor (HM)) while engaging with a game-like user interface for stroke survivors. This pilot study demonstrates that the TDS-HM intervention elicits improvements in motor performance that transfer to reduced upper extremity impairments and improvements in quality of life for moderately to severely impaired stroke survivors.
Upper limb rehabilitation is critical for patients affected by spinal cord injury (SCI). Currently, robotics and Virtual Reality (VR) have changed the way in which rehabilitation therapies are provided. However, a still unreached precondition for these systems is the precise and practical estimation of limb posture and an objective evaluation of patient’s improvement. In this manuscript we present an upper limb rehabilitation platform combining VR, patient posture estimation and objective kinematic indices. This manuscript describes the software platform and criteria which integrate the modules of the system. We report preliminary results of the kinematic indices and platform usability by practitioners.
In this study, a portable prototype system of FES foot drop correction and gait measurements was tested with a hemiplegic subject to develop FES rehabilitation system for gait of hemiplegic subjects. The prototype system consisted of a tablet-type device and 2 inertial sensors and an electrical stimulator. The system worked properly during walking with FES foot drop correction and foot inclination angles, stride time and gait event timings were obtained. Measurement of gait information during rehabilitation training can be useful to evaluate the training effect.
Recovery from impaired gait of stroke patients is in increasing need. Conventional robotic gait rehabilitation systems based on treadmill restrict pelvic horizontal motion and lack of presenting natural gait patterns with actual foot contact on the ground. To overcome these limitations, we proposed a novel robotic walker for gait rehabilitation. It consists of an active omni-directional mobile platform and a body weight support (BWS) unit to assist gait motions. The control algorithm automatically executes force assistance during gait. 3 healthy young subjects were recruited to evaluate force assistance of robotic walker. Recorded assistive forces of the walker to pelvis showed beneficial influence.
In 2012 the neurologic controlled exoskeleton Hybrid assistive limb (HAL) was introduced in the spinal cord injury unit of the university hospital Bochum. Until now 20 acute and 40 chronic injured SCI patients (AIS A-D) have been treated for a three-month body weight supported treadmill training paradigm wearing HAL with a daily intervention. Here we report the feasibility, safety and outcome aspects. All subjects improved due to their functional walking abilities without wearing HAL. The training was feasible, safe and suitable in acute and chronic SCI patients. No severe adverse events were recorded.
In this paper, we develop methods to assist people with disability to control robot using brain computer interface. Using a new technique based on action grammar, robot is able to carry out tasks such as opening a door knob simply by recognizing the intention of the user. To successfully implement this concept we have developed techniques that help articulated robot to become spatially aware. We provide a set of actions which can be combined using action-grammar that is modelled using stationary Markov decision process. We demonstrate our methodology using two tasks (i) screw-insertion and (ii) door-opening.
In the ongoing ironHand (iH) project, a wearable soft-robotic glove, the iH system, is being developed to support the hand during daily functioning. This study gives a first insight in the potential effect of the glove in daily life. Preliminary results show that participants can increase their pinch grip strength with assistance of the glove, but functional tasks were performed slower with the glove compared to performing these tasks without the glove. Overall, usability of the iH system was perceived as good. More research is needed to determine the impact of the iH system in daily life.
In this study, a wearable soft-robotic glove that is connected to a computer with therapeutic software to train hand function (the ironHand therapeutic system, iH TS) is introduced. This study explored usability of the iH TS after first use without receiving instructions from researchers. The results on the System Usability Scale (SUS) are promising for acceptance of the iH TS in daily life (mean SUS score = 66.4). More research is needed to determine user acceptance and the effects of the therapeutic hand exercises after a longer acquaintance period.
Stroke is a devastating condition with profound implications for health economics and resources worldwide. Recent works showed that the use of brain-machine interfaces (BMI) could help movement improvements in severely affected chronic stroke patients. This work shows the feasibility and use of a Soft Orthotic Physiotherapy Hand Interactive Aid (SOPHIA) system, able to provide more intense rehabilitation sessions and facilitate the supervision of multiple patients by a single Physiotherapist. The SOPHIA device is controlled by a BMI system and has a lightweight design and low cost. Tests with researchers showed that the system presents a reliable and stable control, besides being able to actively open the volunteers’ hands.
Predictions of human movements to study control strategies or to investigate human-orthosis interaction require the solution of the kinematic redundancy by solving a dynamic optimization problem. Direct collocation is a promising method to solve these problems. However, the convergence of collocation methods is sensitive to problem formulation. We therefore compare different optimal control problem formulations to predict the motion of a planar two-link model with foot-ground contact. The use of implicit dynamic formulations minimizing accelerations or jerks facilitated the convergence over other formulations. The use of automatic differentiation and an appropriate time scale had a large influence on computation time.
This study presents the assessment of ankle-foot gait abnormalities and estimation of neuromuscular control for maintaining gait dynamic stability and avoid falls. Control signals are modelled as the rate of change in the body COM acceleration as an input and the COP velocity as an output. Experiments show that the toe foot condition is least stable than inverted and normal walk at loading phase. However, the overdamped motor output response, equally stable for the three undamped input instabilities, shows the robustness of our proposed motor controller. Results show that our novel neuromotor inspired controller, based on behavioral I/O signals, is robust and suitable for the assessment of exoskeletal stability and control of wearable soft robotic applications.
The combined use of the functional electrical stimulation (FES) with robotic devices known as Hybrid Robotic System emerges as a promising solution to improve rehabilitation therapies after neurological injuries (e.g. stroke). This work presents a first step towards the implementation of a Hybrid Robotic System simulation platform that could allow exploring several control strategies to improve its performance. The results show the feasibility of the platform to deploy several control strategies that combines FES and robotic devices.
The design and test of a Microsoft Kinect-based system for automatic evaluation of the Modified Jebsen Test of Hand Function (MJT) is presented. The MJT was administered to 11 chronic stroke patients (both the dominant and non-dominant hand). MJT completion times were evaluated by a therapist using a stopwatch and automatically by use of the Kinect-based system. The ground truth times were assessed based on visual inspection of video-recordings. Analysis of the agreement between the MJT times estimated by the two methods generally showed better agreement between the ground truth times and the times estimated by the Kinect-based system compared to the agreement between the ground truth times and the times obtained by the therapist.
Many stroke and spinal cord injured (SCI) patients suffer from a paretic arm movement, which can be characterized by a limited shoulder flexion. We consider a possibility to assist the patient in slow arbitrary arm flexions within a large range of motion. To address this issue, we propose a shoulder flexion dependent weight support during robot-assisted therapy of the upper limb. Inverse static models of the cable-driven robotics and the passive human arm are used to estimate the required forces at the ropes to flex the upper arm in order to compensate a given percentage of the arm weight. Our results show that conventional constant rope forces during a therapy may produce an over- or undercompensated weight support, whereas the proposed adaptive approach achieves a desired larger range of motion.
Autonomous robots that are interfaced with virtual or augmented reality gaming are increasingly being developed to provide repetitive intensive practice to promote increased compliance and facilitate better outcomes in neurorehabilitation therapies. These therapist robots, equipped with a set of sensors and actuators for monitorizing the environment, allows health professionals to supervise the recovery of patients with serious disability. In this paper, a new system for supervising neuro-rehabilitation therapies using autonomous robots is presented. The therapy explained in this work is based on a set of virtual reality games developed by using robotics technologies, such as RGB-D camera and depth image processing. Three different virtual reality games have been developed in the application to gain better outcomes during the therapy, each one focuses on a typical exercise: ‘Touch the apple’, ‘Follow the path’ and ‘Imitate the dance’. Both, the virtual reality games and the main robotics technologies for their development, are explained in this paper.
Cutaneomuscular afferent information is essential for voluntary motor tasks such as gait and balance. After spinal cord injury (SCI) changes in sensorimotor activity are involved in recovery of limited motor function. Here we present a review of clinical neurophysiological measures that quantify sensorimotor dysfunction and which have the potential to benchmark the therapeutic effect of cutaneous stimulation during SCI neurorehabilitation. Specifically we will show that long-latency cutaneous reflex and cutaneomuscular conditioned H-reflex techniques can quantify spinal neuronal activity and the presence of either adaptive or maladaptive motor control mechanisms after incomplete SCI. In conclusion, the development of neurorehabilitation programs in combination with cutaneomuscular stimulation protocols is a viable strategy not only to promote motor recovery but to prevent maladaptive neuroplasticity such as spasticity after SCI.
This century is about brain. Not surprisingly, two megaprojects, the “Brain Activity Map Project” (USA) and the “Human Brain Project” (EU) with a total budget over a few billions euros, have been initiated across the Atlantic and mobilized many of the best and most renowned neuroscientists. They both aims to answer open questions for Neuroscience, such as: Is there an underlying functional architecture to the brain’s networks?, What is the functional connectivity diagram of a circuit?, What are the long-range interactions that underlie cognitive functions and behaviour? or What are the paths of information flow?. Such questions are related to the so-called “functional connectivity”, it reflects the statistical interdependencies between two physiological signals, providing information about functional interactions between the corresponding brain regions. Over the last years it has been increasingly used in neuroscience. Specifically, in the study of electrophysiological recordings such as Magnetoencephalography (MEG) and Electroencephalography (EEG).
Patients with physical disabilities can benefit from robotic rehabilitation since it can provide more control, accuracy and variety of training modes. This improves the efficiency of the patient’s rehabilitation. This work presents a multimodal platform for acquisition and processing of EEG and EMG signal with inertial sensors data in order to detect lower limb movement intentions. A pseudo-online technique was also developed. Experiments were conducted with 5 healthy subjects performing lower limb motor tasks and an experimental protocol is proposed. In the future, this interface will be integrated in an active knee orthosis for robotic rehabilitation. The results obtained show that the system is capable to acquire, process and classify the signals synchronously. The movement intention was evaluated taking into account EEG and EMG signal together with a OR logic, detecting \(60.0\pm 21.2\,\%\) of movement intentions. The movement anticipation achieved \(881.5\pm 136.3\) ms based on EEG signal and \(137.6\pm 77.5\) ms based on EMG signals.
We focus in developing and applying simplified low-cost technologies that allow an approach to rehabilitation process and make it accessible to the most number of subjects. We present two cases: the mechanical equipment for gait stimulation, and the computer system for gait analysis. In both cases we have developed technologies that are being applied with patiens with terapeutical results in concordance with literature. Statistical validation is being performed. These experiences provide access to rehabilitation technologies to a significant amount of patients, professionals and institutions.
Cerebral Palsy (CP) is the most common cause of permanent serious physical disability in childhood. New strategies are needed to help promote, maintain, and rehabilitate the functional capacity of children with severe level of impairment. The main objective of this work is to present a Human-Robot interaction strategy for overground rehabilitation to support novel robotic-based therapies for CP rehabilitation. This strategy is implemented in a new Wearable Robotic Walker named CPWalker. In our approach, legs’ kinematics information obtained from a Laser Range Finder (LRF) sensor is used to detect the user’s locomotion intentions and drive the robotic platform. During a preliminary validation we observed that this approach enabled the robot to continuously follow the human velocity and provided body weight support during gait.
Is presented a non-supervised method for feature selection based on similarity index, which is applied in a brain-computer interface (BCI) to recognize gait preparation/stops. Maximal information compression index is here used to obtain redundancies, while representation entropy value is employed to find the feature vectors with high entropy. EEG signals of six subjects were acquired on the primary cortex during walking, in order to evaluate this approach in a BCI. The maximum accuracy was 55 % and 85 % to recognize gait preparation/stops, respectively. Thus, this method can be used in a BCI to improve the time delay during dimensionality reduction.
One of the most used signals in Brain Machine Interfaces (BMI) is the Steady State Visually Evoked Potentials (SSVEP). In a SSVEP-based BMI, a visual stimulus that flickers in a constant frequency is presented to the user, and the system has to detect if the user is gazing the stimulus. Usually the stimulus is a rectangular signal and there are no clear criteria for select the duty cycle, which is generally fixed to 50 %. We propose a model for SSVEP that links the phase and amplitude variations in function of the duty cycle for a specific frequency. This model can be adjusted using only the phase of the SSVEP signal and it could improve the SSVEP-based BMI by selecting the duty cycle. The model was fixed for SSVEP responses in a man who is 39 years old. The mean absolute error below 0.3 rad shows that the model predicts the phase in the majority of the used frequencies.
In this paper a method to detect velocity-dependant spasticity is shown. The implementation and test were performed using the ankle joints of a six degrees of freedom exoskeleton. In this first stage of the project all subjects are healthy and have no difficulty in walking. The subjects left the ankle joint relaxed while sat down and did not rest the heel on the floor. Force sensors are used in order to measure the interaction forces between the exoskeleton and de human limb. The sensors were located in the foot sole under the subject’s forefoot and in the instep brace. A variable oscillator is used to generate the angle reference signal and to vary the frequency thereof. The presence of spasticity was detected based on the interaction force information and the joint movement was decreased in speed.
Exoskeletons are becoming one of the most promising devices to improve quality of life to injured patients to regain ability to walk. Bioinspired designs in exoskeletons could increase adaptability as well as minimal interference to perform gait movements. An important issue regarding the design of this devices is the hip joint. This work presents a design of a bioinspired hip exoskeleton for enhanced physical interaction, which is based on the motion analysis model, taking into account bioinspired design criterion, and also concepts of wearable robots. As future work, both, the 3D prototyping of this device to evaluate the gait performance, and the actuators selection are considered.
Self mobility plays a key role in early childhood development. Several devices has been proposed but most of them are prohibitive for low income families in development countries. Includ.ING project proposes the inclusive development of a low cost robotic vehicle for children with motor disabilities. Four operation modes and several sensing functions were defined for the platform. This paper present the design and early results with the developed platform.
... From the literature review out of the 14 available channels of Emotive Pro headset, 4 of the frontal channels F3, F4, FC5 and FC6 are selected to be analyzed (Fig. 3). The selected channels are the nearest to the motor cortex region and have been previously used to acquiring MI data in various references [15,17,18]. ...
... The musculotendon unit models described in the previous section require the quantification of a set of parameters, namely (i) optimal fiber length, (ii) tendon slack length, (iii) muscle pennation angle, and (iv) maximal isometric force [87]. Some of these parameters can be directly measured via ultrasound [348][349][350] or MR imaging [92]. However, some others, such as the optimal muscle length or tendon slack, are difficult to measure [104]. ...
The rising importance of movement analysis led to the development of more complex biomechanical models to describe in detail the human motion patterns. The models scaled from simplistic two-dimensional to three-dimensional representations of body including detailed joint, muscle, tendon, and ligament models. Different computational methodologies have been proposed to extend traditional kinematic and dynamic analysis to include not only the evaluation of muscle forces but also the action of the central nervous system. Hence, a large number of models varying in complexity and target application are available in literature. This narrative review aims to provide an overview of the modeling of biomechanical systems used for the analysis of human movement within the framework of multibody dynamics, for those enrolled in engineering, clinical, rehabilitation and sports applications. The review includes detailed and generic models, as well as the main methodologies applied to model muscle activation and contraction dynamics. Numerous skeletal, musculoskeletal and neuromusculoskeletal models with variable degrees of complexity, accuracy and computational efficiency were identified. An important remark is that the most suitable model depends on the study objectives, detail level of the depicted anatomical structures, target population or performed motion. Summarizing, biomechanical systems have evolved remarkably during the last decades. Such advances allowed to gain a deep knowledge on how the human nervous system controls the movement during different activities, which has been used not only to optimize motor performance but also to develop solutions that allow impaired people to regain motor function in cases of disability, among other applications.
... These electrodes may be penetrating or non-penetrating. While penetrating electrodes offer the potential of higher selectivity [27][28][29], non-penetrating nerve cuff electrodes (NCEs) on nerves in the upper [2,4,5,30] and lower [3,26,[31][32][33][34][35] extremities have been operational and stable for more than 11 years postimplantation [31] in terms of stimulation threshold and functional output over time [2,26]. To date, the effects of both implantation and chronic use of NCEs on neurophysiology and muscle innervation have not been examined with established clinical measures, including both nerve conduction studies (NCS) and needle electromyography (EMG). ...
Background:
Peripheral nerve stimulation with implanted nerve cuff electrodes can restore standing, stepping and other functions to individuals with spinal cord injury (SCI). We performed the first study to evaluate the clinical electrodiagnostic changes due to electrode implantation acutely, chronic presence on the nerve peri- and post-operatively, and long-term delivery of electrical stimulation.
Methods:
A man with bilateral lower extremity paralysis secondary to cervical SCI sustained 5 years prior to enrollment received an implanted standing neuroprosthesis including composite flat interface nerve electrodes (C-FINEs) electrodes implanted around the proximal femoral nerves near the inguinal ligaments. Electromyography quantified neurophysiology preoperatively, intraoperatively, and through 1 year postoperatively. Stimulation charge thresholds, evoked knee extension moments, and weight distribution during standing quantified neuroprosthesis function over the same interval.
Results:
Femoral compound motor unit action potentials increased 31% in amplitude and 34% in area while evoked knee extension moments increased significantly (p < 0.01) by 79% over 1 year of rehabilitation with standing and quadriceps exercises. Charge thresholds were low and stable, averaging 19.7 nC ± 6.2 (SEM). Changes in saphenous nerve action potentials and needle electromyography suggested minor nerve irritation perioperatively.
Conclusions:
This is the first human trial reporting acute and chronic neurophysiologic changes due to application of and stimulation through nerve cuff electrodes. Electrodiagnostics indicated preserved nerve health with strengthened responses following stimulated exercise. Temporary electrodiagnostic changes suggest minor nerve irritation only intra- and peri-operatively, not continuing chronically nor impacting function. These outcomes follow implantation of a neuroprosthesis enabling standing and demonstrate the ability to safely implant electrodes on the proximal femoral nerve close to the inguinal ligament. We demonstrate the electrodiagnostic findings that can be expected from implanting nerve cuff electrodes and their time-course for resolution, potentially applicable to prostheses modulating other peripheral nerves and functions.
Trial registration:
ClinicalTrials.gov NCT01923662 , retrospectively registered August 15, 2013.
Eye contact is one of the main human skills, and a prerequisite of verbal language. However, children with Autism Spectrum Disorder (ASD) often have an important deficit in this skill, compromising their entire cognitive and social development. This work shows the development of two Serious Games (SGs), based on the concept of Child-Robot Interaction (CRI), for the exercise and improvement of eye contact and visual attention, as well as concepts such as imitation and emotion recognition. For face detection and eye movement monitoring, an open source development framework of machine learning named MediaPipe was used. Tests with children with ASD will be conducted, and it is expected these SGs have a positive impact regarding the improvement of both eye contact and visual attention for these children.
Engineered feature extraction can compromise the ability of Atrial Fibrillation (AFib) detection algorithms to deliver near real-time results. Autoencoders (AEs) can be used as an automatic feature extraction tool, tailoring the resulting features to a specific classification task. By coupling an encoder to a classifier, it is possible to reduce the dimension of the Electrocardiogram (ECG) heartbeat waveforms and classify them. In this work we show that morphological features extracted using a Sparse AE are sufficient to distinguish AFib from Normal Sinus Rhythm (NSR) beats. In addition to the morphological features, rhythm information was included in the model using a proposed short-term feature called Local Change of Successive Differences (LCSD). Using single-lead ECG recordings from two referenced public databases, and with features from the AE, the model was able to achieve an F1-score of 88.8%. These results show that morphological features appear to be a distinct and sufficient factor for detecting AFib in ECG recordings, especially when designed for patient-specific applications. This is an advantage over state-of-the-art algorithms that need longer acquisition times to extract engineered rhythm features, which also requires careful preprocessing steps. To the best of our knowledge, this is the first work that presents a near real-time morphological approach for AFib detection under naturalistic ECG acquisition with a mobile device.
Objective
To support the design of assistive devices and prostheses, we investigated the changes in upper-limb muscle synergies during the practice of a myoelectric controlled game using proportional-sequential control.
Methods
We evaluated 1) whether individual muscle synergies change in their structure; 2) variability; 3) distinctiveness; and 4) whether individuals become more similar with practice. Ten individuals practiced a myoelectric-controlled serious game for ten consecutive days (25 min/day) and one day after one week without training (retention).
Results
The results showed that individuals decreased the number of synergies employed and modified their flexor synergies structure, becoming more similar as a group with practice. Nevertheless, within-individual synergies’ variability and distinctiveness did not change.
Conclusion
These results point out that individuals do not demonstrate muscle patterns less variable or differentiable after practice. However, participants increased performance and became more attuned to the task dynamics.
Significance
The present findings indicate that, depending on the task requirements, individuals converge to more similar muscle activation patterns – a feature that should be further explored in prosthetic design.
This article addresses the potential clinical value of techniques based on surface electromyography (sEMG) in rehabilitation medicine with specific focus on neurorehabilitation. Applications in exercise and sport pathophysiology, in movement analysis, in ergonomics and occupational medicine, and in a number of related fields are also considered. The contrast between the extensive scientific literature in these fields and the limited clinical applications is discussed. The “barriers” between research findings and their application are very broad, and are longstanding, cultural, educational, and technical. Cultural barriers relate to the general acceptance and use of the concept of objective measurement in a clinical setting and its role in promoting Evidence Based Medicine. Wide differences between countries exist in appropriate training in the use of such quantitative measurements in general, and in electrical measurements in particular. These differences are manifest in training programs, in degrees granted, and in academic/research career opportunities. Educational barriers are related to the background in mathematics and physics for rehabilitation clinicians, leading to insufficient basic concepts of signal interpretation, as well as to the lack of a common language with rehabilitation engineers. Technical barriers are being overcome progressively, but progress is still impacted by the lack of user-friendly equipment, insufficient market demand, gadget-like devices, relatively high equipment price and a pervasive lack of interest by manufacturers. Despite the recommendations provided by the 20-year old EU project on “Surface EMG for Non-Invasive Assessment of Muscles (SENIAM),” real international standards are still missing and there is minimal international pressure for developing and applying such standards. The need for change in training and teaching is increasingly felt in the academic world, but is much less perceived in the health delivery system and clinical environments. The rapid technological progress in the fields of sensor and measurement technology (including sEMG), assistive devices, and robotic rehabilitation, has not been driven by clinical demands. Our assertion is that the most important and urgent interventions concern enhanced education, more effective technology transfer, and increased academic opportunities for physiotherapists, occupational therapists, and kinesiologists.
ResearchGate has not been able to resolve any references for this publication.