|Track 1. Prosthetics|
|SS1. Bayesian models for motor rehabilitation and control
Organizer: Luca Citi, University of Essex, UK
Bayesian statistics is a branch of statistics that provides a principled way of calculating how prior knowledge can be combined with observations in a statistically optimal way. Numerous studies have shown that many aspects of human behaviour (e.g. sensorimotor control) are close to “Bayes’ optimal”. Learning (or re-learning in the case of rehabilitation) can also be seen as the process of updating our priors based on new observations. It seems therefore natural to employ such models not only to describe neural processes but also to the decode them for the purpose of improving the effectiveness of rehabilitation or for the control of a prosthetic device. From an implementation point of view, Bayesian models are also appealing because they may be better suited than other techniques for dealing with the smaller data sets that are typically available in biomedical engineering applications.
This session’s goal is to promote and showcase cutting edge research in modeling human neuromotor control for the purpose of improving rehabilitation techniques or designing better control strategies for actuated prostheses. In particular, this special session is dedicated to understanding the potential benefits and limitations of applying Bayesian techniques to such problems. We envision the session will provide an interdisciplinary platform bringing together scientists working on theoretical models and engineers working on rehabilitation and restoration devices. Example topics may include but are not limited to: Bayesian model averaging, Markov-Chain Monte Carlo, sequential Monte Carlo, dynamic causal modeling, state-space modeling, hierarchical Bayesian modeling, non-parametric Bayesian inference, and their application to the aforementioned biomedical engineering problems.
SS2. Translating research prototypes to bedside: the lesson-learnt of the RETRAINER EU project
In the rehabilitation field, every year innovative high-tech research prototypes are developed and sometimes pilot-tested in small groups of patients but only few of them are really made available to patients with a proved demonstration of their efficacy. The most ambitious aim of the EU project RETRAINER is to propose a novel roadmap to translate research prototypes to bedside passing through randomized controlled trials (RCT) and industrial exploitation.
Starting from the results of a previous FP7 project, MUNDUS, the RETRAINER project developed two hybrid robotic systems for arm training (RETRAINER-S1) and for hand functions recovery (RETRAINER-S2) in stroke survivors. RETRAINER-S1 supports arm movements by the combined action of a passive exoskeleton for weight relief and Functional Electrical Stimulation (FES), which amplitude is controlled based on the residual EMG activities of the same stimulated muscles. Residual functionalities are thus trained and improved rather than replaced by the device. RETRAINER S2 consists of a wearable modular system with multiple electrode arrays, in which the stimulation patterns can be tuned to elicit functional grasp, to obtain whole muscle conditioning and to produce open-loop or closed-loop grasp control. Both systems benefit from the use of interactive objects, i.e. daily life objects equipped with RFID tags. A reader embedded in the robotic system recognizes the selected object or target among several ones and drives the rehabilitation exercises. Both RETRAINER-S1 and RETRAINER-S2 have been extensively tested in two parallel multi-center RCTs involving 68 stroke patients each.
The special session will describe the two RETRAINER systems, the preliminary results of the two RCTs and the perspectives of passive exoskeletons, upper limb FES, hand neuroprostheses and smart objects for upper limb stroke rehabilitation. Finally, the point of view of the big industry will be reported by Ottobock representatives.
SS3. Computer Models in the Design of Neurotechnologies and Rehabilitation Tools
Neurological disorders affect a growing number of individuals in modern societies, and represent a titanic burden in terms of economic and humane costs. However, the development of new therapies for these conditions is hindered by a chronic lack of knowledge in the organization and function of the nervous system. Computer models could potentially overcome this limitation by constraining the design of novel technologies and restrict experiments to meaningful measures. In this workshop we explore how models can be used to synthetize scientific concepts into in-silico representations of the nervous system and neuromuscular architecture and instruct the design of effective therapies in neurological impairments.
SS4. New perspectives in upper limb prosthetics: from the robotics laboratory to clinical use
|Track 2. Rehabilitation Robotics|
|SS5. Improving Strategies for Human-Robot Interaction for Rehabilitation Robotics applications
Organizers: Giacomo Severini, School of Electrical and Electronic Engineering, University College Dublin; Donal Holland, School of Mechanical and Materials Engineering, University College Dublin.
Robotic systems have become, in recent years, common tools in neurorehabilitation practice. Robots, defined as devices capable of manipulating physical objects and responding to their physical environment through mechanical actuation and sensing, are attractive as assistive and training technologies due to their ability to mimic, assist and perform the complex motions involved in human movement. Research in robotics has yielded, in the past two decades, new technologies, advances in rehabilitation, and an improved understanding of biomechanics and neuromechanics. This new knowledge has allowed the development of robotic training systems that can facilitate high-intensity training after neurological injuries, and of more versatile and effective robotic assistive devices. In this Special Session we will give a broad perspective of new solutions and technologies aimed at improving interaction between human and robots for applications in neurorehabilitation and assistance. We will cover the topic from different technological points of view, spanning the development of new control strategies, the exploitation of advanced modeling tools and the design of new hardware solutions. Our aim is to give to the delegates a broad overview of current advancements in the design and development of new solutions in rehabilitation robotics aiming at improving Human-Robot Interaction.
SS6. Increasing the exercise intensity during gait training
Exercise intensity is important in the rehabilitation of patients with central neurological disorders. At one hand it is determined by technical aspects of the rehabilitation methods and at the other hand by physiological and psychological factors of the patient. Rehabilitation psychology offers a number of theoretical models, related to motivational factors, that can be implemented in rehabilitation practice in order to increase the active participation by the patient.Stroke guidelines recommend exercising large-muscle activity, such as walking, 3 to 5 times per week for 20 minutes at moderate intensity. Current evidence suggests that the exercise intensity during gait training is below moderate intensity – especially bodyweight-supported and robot-assisted walking. Currently, only walking unassisted or with assistive devices was able to induce moderateintensity. Therefore, how to increase this exercise intensity is a crucial question. First, the implementation of virtual reality during (robot-assisted) gait training is an option. It can provide a safe, purposeful and enriched environment to practice. Furthermore, it has the potential to give “real-time” feedback and rewards to improve the performance of the user. Using an immersive head-mounted-display and incorporating features relevant for recovery and motor learning, could increase the exercise intensity.
Second, the use of augmented reality systems (e.g. projecting virtual objects on a treadmill or the use of games on a TV screen) is another approach. Although the media are very different as well as the methods for interacting with them, they all provide motivational cues in order to increase compliance and adherence to the rehabilitation schemes, all with the goal of eliciting large number of repetitions.
Third, a patient controlled haptic treadmill can be used. At the moment, one of the biggest constrains of treadmill based devices is that users cannot change the gait speed at will, based on their own intention of movement. This has also had a considerable impact on serious game designers, as they are very limited when it comes to design environments and advanced ways of interaction.
All these topics will be discussed during this special session, data of own studies will be presented and a patient controlled haptic treadmill will be introduced.
SS7. Shaping robotic training to maximize patient outcome: new trends and perspectives
In robot-assisted neurorehabilitation, matching the task difficulty level to the patient’s needs and abilities, both initially and as the recovery process progresses, may enhance the effectiveness of training and improve patients’ motivation and outcome. In this special session, we will present new perspectives for shaping robot-assisted training in order to maximize patient outcome. Improved recovery can be obtained through training modalities that incorporate the predictions of computational models, or by targeting both motor and sensory impairments. Information about the neural correlates of movement at peripheral and central level – for instance, muscle synergies and EEG activity – may provide additional ways to optimize training and to promote plasticity. Finally, the use forms of reward that promote motivation will be discussed from a sensory perspective.
SS8. Neurorehabilitation from clinical perspective and robotic perspective: Contradictions and Integrations
Facilitating neural plasticity is the key to give a hope of functional recovery following neurological injuries. Effective, as well as, practical assessments of neural plasticity and/or motor recovery should be highly essential when considering applying rehabilitation approach. Although robotics rehabilitations technologies have been developing rapidly and remarkably in recent years, the question that remain unanswered is how much can these technologies be reliable alternative to traditional clinical approaches? What are the points of convergence and points of contention? This session will discuss various topics concerning technology-based assessments and compare it with traditional clinical assessments trying to highlight the gap between the two approaches.
In this special session, the initial part of the talks will be about the possibility of utilizing technologies to stimulate neural plasticity through motivating post-stroke patients along the rehabilitation: The first talk will be about the impact of using interaction gaming and virtual reality for post-stroke rehabilitation . The second talk will be about the possibility of utilizing machine learning to predict recovery direction of the patient . The next part of the talks will be about the influence of home-setting vs clinical-setting rehabilitation . In , the author will be presenting a home rehabilitation robot that can be as efficient as therapy in clinical settings. Developing such as technology may increase recovery chance of the patient due the factor of repetitive and intensive training, as well as, decreasing on economic pressures. In , a medical doctor will be highlighting the neuroscientific principles of rehabilitation and discussing the possibility to overcome dichotomy between traditional and technological intervention. In , the author will discuss the perspectives of clinicians and developers regarding what helps or hinders uptake of rehabilitation technologies.
SS9. Balance control during walking-related motor tasks
This proposal plans to organize a special session with emphasis on works concerning balance control during walking-related motor tasks.
Bipedal walking is an inherently unstable and complex motor task that requires progression of the body towards the intended destination while keeping an upright posture. It consists of three primary components: locomotion, balance, and ability to adapt to the environment. Noticeably, balance and postural control are achieved by implementing sophisticated motor control strategies to ensure stability and prevent falls. Normal aging and different neurological diseases are accompanied by an alteration in the capacity to ambulate in a stable manner, thus increasing the probability of falling during the activities of daily living. Gait disorders and subsequent falls are among the major causes of chronic disability in the elderly and subjects with neuro-musculo-skeletal disorders. Accordingly, there is an urgent need for novel therapeutic and technological solutions that can prevent falls and gait disorders.
Research on balance control during walking-related motor tasks could lead to effective methods to early identify and remediate gait deficits and develop methods and technologies for the augmentation and restoration of gait function. Early identification of people most at risk for gait disorders may facilitate preventative intervention to reduce the likelihood of falls or other complications. Augmented sensory feedback can be used to remediate compromised sensorial integration, essential for proper gait function. An increasing amount of exoskeletons are currently used to remediate compromised musculoskeletal function and restore walking in individuals with gait disorders. Innovate robotic devices and control strategies are helping in the development of new gait training protocols.
This special session will bring together globally recognized roboticists, neuroscientists and engineers to initiate a multi-disciplinary discussion on this rapidly evolving field.
SS10. The use of ambulant technology in stroke rehabilitation
Ambulant measurement technologies, such as Inertial Measurement Units (IMUs) and force shoes have become available for use in daily clinical practice. This development has a huge potential as it allows measurements of more complex gait tasks, which can greatly increase our understanding of the influence of stroke on gait biomechanics. Ambulant technologies can also be used to quantify the gait activities of an individual and, thus, be used for the control of (soft) exoskeletons. This special session will provide an overview of the current state of the art of how ambulant technologies are used quantify stroke and control (soft) exoskeletons.
SS11. Redundancy and modularity in motor control: neuroscience, prosthetic, rehabilitative and assistive approaches
The central nervous system shows a remarkable ability in coordinating a high number of muscles acting across one or more joints. To produce a wide range of purposeful movements, the CNS exhibits extremely complex patterns of coordination, observable both at the neuromuscular level and at the joint kinetic and kinematic level. This complexity, together with the high dimensionality, makes it challenging to understand the underlying mechanisms and to use them in practical applications. The concept of motor synergism is becoming more and more attractive in the neurorehabilitation scenario, due to its ability to describe motor coordination in a compact and intuitive way. After many years of research in the field of neuroscience and bioengineering, we are now observing an increasing number of applications in the fields of rehabilitation robotics, neuro-prosthetics, neuro-musculo-skeletal modelling and clinical assessment.
In the field of neuroscience, investigation of motor coordination has unveiled some important aspects underlying neural control strategies by the nervous system, which led to a better understanding of human movement in health and pathology. Synergy based indicators have been identified, able to objectively quantify motor impairment in a wide range of pathologies with motor symptoms; this increased knowledge has led to the development of devices and neurorehabilitation therapies based on the concept of motor synergies.
In the field of upper limb prosthetics, the concept of synergetic motion offers a powerful tool to optimize the design of mechanisms towards compact, lightweight yet functional devices, i.e., that minimize compensatory movements in the users. The application of this concept results in mechanisms that produce non-holonomic combinations of joint angles, which is desirable to reduce the number of actuators towards increased robustness and controllability. The concept of synergy has also been used to facilitate the definition of myoelectric control strategies, through the dimensionality reduction of high-density EMG signals from upper limb muscles.
This special session wants to gather all these new ideas for translating the synergy analysis into the future of rehabilitation practice, encompassing methodologies, innovative devices, and approaches based on motor control theories to deepen our knowledge on the design of functional prostheses and effective neuro-rehabilitation therapies.
|Track 3. Neuroscience|
|SS13. Neural Signal Analysis: Novel Approaches to Understanding Brain Diseases
Organizers: Roberto Hornero (PhD), Biomedical Engineering Group, University of Valladolid.
Among the many diseases affecting human health, disorders of the brain are major causes of morbidity, mortality and impaired quality of life. Electrophysiological and neuroimaging techniques – including electroencephalogram (EEG), magnetoencephalogram (MEG), functional magnetic resonance imaging (fMRI), diffusion tensor imaging (DTI) or positron-emission tomography (PET), among others – are a window to the brain. However, substantial challenges remain in interpreting the results of such large-scale quantitative data. Novel methods and approaches in the field of signal processing open up new avenues for better understanding brain function, as well as diagnosing and treating brain diseases. This special session is focused on reporting cutting-edge research in disease-related brain disorders and discussing the upcoming challenges to be faced. Specifically, this session aims to gather participants in a multidisciplinary forum for presenting, sharing and discussing current and future research trends on the application of different brain signal processing techniques to further understand different pathological conditions affecting the brain.
SS14. New Frontiers in Movement Analysis: from assessment to rehabilitation
Movement analysis allows for providing to neuroscientists important information about how human beings move their body, in terms of kinematics, kinetics and energies, and to clinicians a quantitative description and assessment of motor functions and deficits of a patient. The first use of instrumented movement analysis dates back to the end of XIX century, since then, the use of this technology has made enormous strides. Now, instrumented movement analysis is ready for a new jump forward thanks to the diffusion of wearable devices and the diffusion of robots and virtual reality systems with which it could be integrated. Hence, movement analysis seems ready to be diffused in the rehabilitation gyms or at home of patients or even worn by community dwelling people for a continuous monitoring. This special session aims at facing this new challenge for instrumented human movement analysis.
SS15. Modeling Joint Neuromechanics and Its Applications: System Identification Approach
Joint neuromechanics deals with the relationship between torques (forces) applied to a joint and the resulting movements. Forces, and resulting joint torques, are exerted by muscles via tendons, passive tissue (e.g. ligaments), and the external environment through limbs. Muscle forces are both passive (i.e. due to purely mechanical mechanisms), and active, generated as a result of neural activation from central and spinal responses mediated by proprioceptive feedback.
Accurate modelling of joint neuromechanics is challenging since they arise from a complex interaction of dynamic, nonlinear elements; e.g. muscles, tendons, and neural circuits. Also, the system has limited non-invasive observability and exhibits time-varying behaviour during functional tasks. Nonetheless, it has been studied for over four decades using reductionist and holistic (system-level) approaches. Reductionist models simulate contact and individual muscle/tendon forces and have applications in surgery and artificial joint design. However, reductionist models are too detailed for building biomimetic prostheses, controlling orthoses, and quantitative assessment of peripheral neuromuscular diseases. For these applications, system-level models capturing the net dynamic behaviour of all neuromuscular elements are more useful.
These models are estimated by applying system identification methods to data recorded in-vivo. They have three forms depending on input/output definition: 1) Stiffness: Dynamic relation between joint position as input and the torque acting about it; 2) Compliance: Inverse of stiffness; and 3) impedance relating joint velocity to torque. They often consist of two distinct components: a) Mechanical due to limb inertia, viscoelasticity of the muscle-tendon complex, and active properties of contracting muscle; b) Neural arising from changes in muscle activation in response to, e.g., changing muscle length and force.
This special session first presents state-of-the-art identification methods to model nonlinear joint stiffness in quasi-/non-stationary conditions of posture/movement. It then presents five papers from distinguished researchers in human motor control, neuromotor rehabilitation, prostheses, and orthoses that utilize or benefit from system identification in their research.
SS16. Machine Learning in NeuroRehabilitation
In recent years, neurorehabilitation technologies have proliferated to include a wide variety of activity sensing, stimulation, feedback, and assistive devices. While such devices have greatly expanded the capacity for neuro-kinetic recording, analysis, and augmentation, the personalization and adaptation of device output to maximize functional recovery based on the needs of individual patients remains an intractable challenge. The complex, multi-factorial nature of this problem makes it well suited for machine learning-based optimization. Accordingly, this session will explore the applications, methods, capabilities, remaining challenges, and opportunities for machine learning approaches in neurorehabilitation.
SS17. Non-Invasive Stimulation At Different Level Of Nervous System In Neurorehabilitation
Non-invasive cerebral and spinal cord stimulations have shown some potential for promoting plasticity, which can enhance the functional recovery causing more robust plastic changes in the sensory-motor system following spinal cord or brain injury. Electrical or magnetic stimulation of the motor cortex or central nervous system has been shown to promote motor plasticity Repetitive transcranial magnetic stimulation (rTMS) is a noninvasive and painless procedure that modulates excitability of cortical motor areas, inducing long-lasting changes in the descending corticospinal tract and trans-synaptically at distant sites. The procedure consists of the delivery of short intense electrical current over the scalp through an insulated coil (magnetic coil). Depending on stimulation parameters, TMS can upregulate or downregulate to different extents the excitability of the neural structures under the stimulating coil. The mechanisms underlying the aftereffect of non-invasive brain stimulation are likely a number of interacting factors such as a synaptic plasticity throughout mechanisms of long-term potentiation and depression; and/or by enhancing the expressions of neurotransmitters and neurotrophins, such as glutamate, N-methyl-daspartate, and brain-derived neurotrophic factor (BDNF).
Trancotaneous spinal cord stimulation (tSCS): Over the last past years, a number of experiments especially in animals have shown the plasticity of the spinal cord, and proven that following SCI and under the appropriate intervention, the brain-to-spinal connectome possess the malleability to render to some extent functional recovery. Using tSCI, which provides additional excitation and engages spinal reflex circuits to the cyclic sensory feedback during robotic-driven stepping in complete SCI patients. Recent studies in human patients have shown that epidural stimulation of the lumbosacral spinal cord in paraplegic patients leads to voluntary movements of the legs and postural control for a short period of time.
Focal or generalized vibratory stimuli in neurorehabilitation is well tolerated, effective and easy to use, and it could be used to reduce spasticity, to promote motor activity and motor learning within a functional activity, even in gait training, independent from etiology of neurological pathology. iT It can be used the neurological diseases or disorders like stroke, spinal cord injury, multiple sclerosis, Parkinson’s’ disease and dystonia. It can improve functional improvement such as gait following stroke.
SS18. Cognitive approaches for rehabilitation of patients with neurological disorders
Top-down or cognitive approaches accept that the motor requirements for any task are variable and that motor control for a particular task becomes more adequate when subject understand its purpose. These methods empathised during therapy the importance of assisting the patients to identify, develop and utilize cognitive strategies to carry out daily life activities more effectively. This special session aims to illuminate about the cognitive methods used currently during the rehabilitation of the subjects with neurological conditions.
|Track 4. BMI|
|SS19. Multimodal neural interfaces for rehabilitation and assistance of people with disability
Organizers: Eduardo López-Larraz (University of Tübingen) and Jaime Ibáñez (University College London).
Brain-machine interfaces and other multimodal neural interfaces directly or indirectly accessing the central nervous system constitute a potential tool to boost the efficacy of rehabilitative and assistive technologies for patients with neural conditions.
Despite the first validations of closed-loop brain-machine interfaces with healthy subjects and patients with neural conditions were presented almost two decades ago, their use is still limited, and always in very specific and well-controlled conditions. Before neuroprosthetics and neural interfaces become a standard tool to rehabilitate or substitute the motor function of patients with paralysis, these systems have to prove their reliability out of the lab. In addition, further basic research needs to be conducted to study the functional role of brain activities in the generation of movements, and to understand how these can be processed and harnessed to improve currently existing neural rehabilitation protocols based on the stimulation of the lesioned neural system.
SS20. Application of Functional Electrical Stimulation (FES) to lower limb movement assistance
This special session intends to give an overview of on-going research on the applicationof FES to assist lower limb movements for both training and functional objectives. Recent results on gait assistance in post-stroke hemiplegic individuals, transfer assistance in complete SCI individuals, cycling in post-stroke hemiplegia and rowing in people with SCI will be discussed. Engineering and clinical issues will be considered through interdisciplinary presentations.
SS21. Uncovering neural mechanisms of post-stroke recovery using clinical imaging tools
Over the last decades, the amount and quality of rehabilitative treatments have increased, capitalizing on recent technological developments. Regardless of those developments, recovery from a neurological condition is still very challenging and most patients struggle to translate their improvements to daily life activities. In this regard, a better understanding of the mechanisms underlying motor recovery is necessary to further improve the existing therapies. In this special session, we will define robust biomarkers of recovery through various techniques such as functional magnetic resonance imaging, electroencephalography and diffusion tensor imaging, which allow exploring the neural processes of post-stroke rehabilitation. We believe that understanding these mechanisms will pave the way towards innovative therapies efficiently harnessing plasticity, hence optimally engaging the spared brain connections and networks.
SS22. Pattern Recognition Techniques for assessment, training and rehabilitation
This special session is focused on the area of the application of pattern recognition and intelligent data mining techniques to support bioengineering applications from the acquisition of the data, to the model of the experiments and then to the phase of evaluating and interpreting the final information. In this context, machine learning methods and feature extraction algorithms are playing a key role to design, implement and validate innovative protocols, because they can support the best decision making strategies to collect, treat, process, interpret and present data acquired by different devices or sensors in several scenarios.
SS23. Brain machine interfaces to restore locomotion after spinal cord injury
Spinal cord injury (SCI) profoundly affects patients’ quality of life and functional independence, and has tremendous financial consequences as a result of health care costs and lost productivity. Improved neurorehabilitation to address these needs after SCI is vital. In the recent years, brain machine interfaces (BMIs) have become promising technical means to improve or restore locomotion after SCI. BMIs acquire, decode and then transform brain signals into control commands that can drive robotic systems such as exoskeletons, or even functional electrical stimulation (FES) that allows voluntary use of the patient’s own muscles. This Special Session will present several different types of BMIs and discuss their advantages and disadvantages.
SS24. Array electrode for the assessment of muscle functions; When, where and why?
In late nineties of the XX century 16 European groups have contributed to the SENIAM project (Surface EMG for Non-Invasive Assessment of Muscles): a concerted action founded by the European Commission within the BIOMED 2 program. Recommendations for sensor and sensor placement procedures have been developed and guidelines for processing and reporting sEMG data have been set. The high-density electrode arrays, miniature wireless data acquisition systems open new and innovative fields for sEMG applications. Among many applications the special session will discuss the following: monitoring muscular activation, control of powered artificial extremities and exoskeletons and biofeedback as part of the exercise and sports. The basic advantage that the presenters in the session will discuss is that the electrophysiological signal recorded via an array provides a temporal and spatial map related to the motor systems underlying the area covered by the array. Array electrodes eliminate the problem of the positioning over the muscle belly or other muscle parts. Array electrode require multi-channel recorder, many leads connector to the recorder, and much more complex analysis of signals. Invited speakers are asked to present their experiences in the application of array electrodes to initiate the discussion what are the bottlenecks in the wider use of EMG maps in motor control and rehabilitation. The session also includes a presentation of an implantable array for control of artificial hand.
SS25. Reshaping Perception and Action in Human-Machine Interfaces
After stroke, spinal cord injury (SCI) or loss of limb, one become dependent upon a variety of technological instruments. These could be assistive devices such as wheelchairs or exoskeletons, or prosthetic limbs. Other devices, such as robotic manipulators, are therapeutic tools to perform targeted exercises. In all such cases, the interactions between the user and the machine is mediated by interfaces, re-mapping the users’ actions onto their sensory consequences: the motion of a visible cursor, of a wheelchair or of an artificial limb. In this session, the speakers will discuss how by programming this mapping it is possible to reorganize motor behavior, recover lost skills or acquire new ones. Different forms of feedback, visual tactile and kinaesthetic will be considered in their ability to generate stable changes in perception and behaviour. These changes occur as the user is learning to control the assistive or therapeutic device by forming an internal model of its operation. The engineering of sensory motor interactions will be considered as a means to form new stable motor memories through the combination visual, tactile and kinaesthetic feedback.
SS26. Brain-state dependent non-invasive neuromodulation of human cortex
A variety of adjuvant protocols have been tested to enhance the spontaneous biological recovery in the acute and subacute phases following a stroke. However, no significant benefits have been reported with respect to classic therapeutic interventions. A relatively novel approach uses the instantaneous brain-state to trigger non-invasive peripheral nerve or cortical stimulation. A current important question is to what extent repetitive pairing of peripheral or central stimulation with different brain states would lead to a lasting change of corticospinal excitability. Answering this question poses an inherent challenge due to the strong influence of preceding muscle activity on stimulation effects and due to the large inter-trial variability in spontaneous and movement-related brain activity. In order to face this unpredictability of intrinsic brain states, neurofeedback devices could provide a viable solution and may help in developing new closed-loop interventions based on the current brain state in order to facilitate cortico-spinal plasticity, e.g. in the context of neurorehabilitation. This session provides an opportunity for researchers, engineers and clinicians to present and discuss novel neurorehabilitation paradigms based on the targeted neuromodulation of cortical and spinal level networks using brain-state dependent stimulation paradigms.