Advances in neuroprosthetic learning and control

Volitional modulation of neuronal activity in the external globus pallidus by engagement of the cortical-basal ganglia circuit

Volitional modulation of neural activity is not confined to the cortex but extends to various brain regions. Yet, it remains unclear whether neurons in the basal ganglia structure, the external globus pallidus (GPe), can be volitionally controlled. Here, we employed a volitional conditioning task to compare the volitional modulation of GPe and primary motor cortex (M1) neurons as well as the underlying circuits and control mechanisms. The results revealed that the volitional modulation of GPe neuronal activity engaged both M1 and substantia nigra pars reticulata (SNr) neurons, indicating the involvement of the cortex-GPe-SNr loop. In contrast, the volitional modulation of M1 neurons primarily occurred through the engagement of M1 local circuitry. Furthermore, lesioning M1 neurons did not affect the volitional learning or volitional control signal in GPe, whereas lesioning of GPe neurons impaired the learning process for the volitional modulation of M1 neuronal activity at the intermediate stage. Additionally, lesion of GPe neurons enhanced M1 neuronal activity when performing the volitional control task without reward delivery and a random reward test. Taken together, our findings demonstrated that GPe neurons could be volitionally controlled by engagement of the cortical-basal ganglia circuit and inhibit learning process for the volitional modulation of M1 neuronal activity by regulating M1 neuronal activity. Thus, GPe neurons can be effectively harnessed for independent volitional modulation for neurorehabilitation in patients with cortical damage. KEY POINTS: The cortical-basal ganglia circuit contributes to the volitional modulation of GPe neurons. Volitional modulation of M1 neuronal activity mainly engages M1 local circuitry. Bilateral GPe lesioning impedes volitional learning at the intermediate stages. Lesioning of GPe neurons inhibits volitional learning process by regulating M1 neuronal activity.

Transitioning from global to local computational strategies during brain-machine interface learning

When learning to use a brain-machine interface (BMI), the brain modulates neuronal activity patterns, exploring and exploiting the state space defined by their neural manifold. Neurons directly involved in BMI control (i.e., direct neurons) can display marked changes in their firing patterns during BMI learning. However, the extent of firing pattern changes in neurons not directly involved in BMI control (i.e., indirect neurons) remains unclear. To clarify this issue, we localized direct and indirect neurons to separate hemispheres in a task designed to bilaterally engage these hemispheres while animals learned to control the position of a platform with their neural signals. Animals that learned to control the platform and improve their performance in the task shifted from a global strategy, where both direct and indirect neurons modified their firing patterns, to a local strategy, where only direct neurons modified their firing rate, as animals became expert in the task. Animals that did not learn the BMI task did not shift from utilizing a global to a local strategy. These results provide important insights into what differentiates successful and unsuccessful BMI learning and the computational mechanisms adopted by the neurons.

Impact of dataset size and long-term ECoG-based BCI usage on deep learning decoders performance

Introduction

In brain-computer interfaces (BCI) research, recording data is time-consuming and expensive, which limits access to big datasets. This may influence the BCI system performance as machine learning methods depend strongly on the training dataset size. Important questions arise: taking into account neuronal signal characteristics (e.g., non-stationarity), can we achieve higher decoding performance with more data to train decoders? What is the perspective for further improvement with time in the case of long-term BCI studies? In this study, we investigated the impact of long-term recordings on motor imagery decoding from two main perspectives: model requirements regarding dataset size and potential for patient adaptation.

Methods

We evaluated the multilinear model and two deep learning (DL) models on a long-term BCI & Tetraplegia (ClinicalTrials.gov identifier: NCT02550522) clinical trial dataset containing 43 sessions of ECoG recordings performed with a tetraplegic patient. In the experiment, a participant executed 3D virtual hand translation using motor imagery patterns. We designed multiple computational experiments in which training datasets were increased or translated to investigate the relationship between models' performance and different factors influencing recordings.

Results

Our results showed that DL decoders showed similar requirements regarding the dataset size compared to the multilinear model while demonstrating higher decoding performance. Moreover, high decoding performance was obtained with relatively small datasets recorded later in the experiment, suggesting motor imagery patterns improvement and patient adaptation during the long-term experiment. Finally, we proposed UMAP embeddings and local intrinsic dimensionality as a way to visualize the data and potentially evaluate data quality.

Discussion

DL-based decoding is a prospective approach in BCI which may be efficiently applied with real-life dataset size. Patient-decoder co-adaptation is an important factor to consider in long-term clinical BCI.

Converging Robotic Technologies in Targeted Neural Rehabilitation: A Review of Emerging Solutions and Challenges

Recent advances in the field of neural rehabilitation, facilitated through technological innovation and improved neurophysiological knowledge of impaired motor control, have opened up new research directions. Such advances increase the relevance of existing interventions, as well as allow novel methodologies and technological synergies. New approaches attempt to partially overcome long-term disability caused by spinal cord injury, using either invasive bridging technologies or noninvasive human-machine interfaces. Muscular dystrophies benefit from electromyography and novel sensors that shed light on underlying neuromotor mechanisms in people with Duchenne. Novel wearable robotics devices are being tailored to specific patient populations, such as traumatic brain injury, stroke, and amputated individuals. In addition, developments in robot-assisted rehabilitation may enhance motor learning and generate movement repetitions by decoding the brain activity of patients during therapy. This is further facilitated by artificial intelligence algorithms coupled with faster electronics. The practical impact of integrating such technologies with neural rehabilitation treatment can be substantial. They can potentially empower nontechnically trained individuals-namely, family members and professional carers-to alter the programming of neural rehabilitation robotic setups, to actively get involved and intervene promptly at the point of care. This narrative review considers existing and emerging neural rehabilitation technologies through the perspective of replacing or restoring functions, enhancing, or improving natural neural output, as well as promoting or recruiting dormant neuroplasticity. Upon conclusion, we discuss the future directions for neural rehabilitation research, diagnosis, and treatment based on the discussed technologies and their major roadblocks. This future may eventually become possible through technological evolution and convergence of mutually beneficial technologies to create hybrid solutions.

Wireless recording from unrestrained monkeys reveals motor goal encoding beyond immediate reach in frontoparietal cortex

System neuroscience of motor cognition regarding the space beyond immediate reach mandates free, yet experimentally controlled movements. We present an experimental environment (Reach Cage) and a versatile visuo-haptic interaction system (MaCaQuE) for investigating goal-directed whole-body movements of unrestrained monkeys. Two rhesus monkeys conducted instructed walk-and-reach movements towards targets flexibly positioned in the cage. We tracked 3D multi-joint arm and head movements using markerless motion capture. Movements show small trial-to-trial variability despite being unrestrained. We wirelessly recorded 192 broad-band neural signals from three cortical sensorimotor areas simultaneously. Single unit activity is selective for different reach and walk-and-reach movements. Walk-and-reach targets could be decoded from premotor and parietal but not motor cortical activity during movement planning. The Reach Cage allows systems-level sensorimotor neuroscience studies with full-body movements in a configurable 3D spatial setting with unrestrained monkeys. We conclude that the primate frontoparietal network encodes reach goals beyond immediate reach during movement planning.

A Practical Approach to Artificial Intelligence in Plastic Surgery

Understanding the intersection of technology and plastic surgery has been and will be essential to positioning plastic surgeons at the forefront of surgical innovation. This account of the current and future applications of artificial intelligence (AI) in reconstructive and aesthetic surgery introduces us to the subset of issues amenable to support from this technology. It equips plastic surgeons with the knowledge to navigate technical conversations with peers, trainees, patients, and technical partners for collaboration and to usher in a new era of technology in plastic surgery. From the mathematical basis of AI to its commercially viable applications, topics introduced herein constitute a framework for design and execution of quantitative studies that will better outcomes and benefit patients. Finally, adherence to the principles of quality data collection will leverage and amplify plastic surgeons’ creativity and undoubtedly drive the field forward.

Bypassing stroke-damaged neural pathways via a neural interface induces targeted cortical adaptation

Regaining the function of an impaired limb is highly desirable in paralyzed individuals. One possible avenue to achieve this goal is to bridge the interrupted pathway between preserved neural structures and muscles using a brain–computer interface. Here, we demonstrate that monkeys with subcortical stroke were able to learn to use an artificial cortico-muscular connection (ACMC), which transforms cortical activity into electrical stimulation to the hand muscles, to regain volitional control of a paralysed hand. The ACMC induced an adaptive change of cortical activities throughout an extensive cortical area. In a targeted manner, modulating high-gamma activity became localized around an arbitrarily-selected cortical site controlling stimulation to the muscles. This adaptive change could be reset and localized rapidly to a new cortical site. Thus, the ACMC imparts new function for muscle control to connected cortical sites and triggers cortical adaptation to regain impaired motor function after stroke.

Monkeys were trained to use an artificial cortico-muscular connection (ACMC) to regain control over a paralyzed hand following subcortical stroke. Control over the paralyzed hand was accompanied by the appearance of localized high-gamma modulation in the cortex, which could be rapidly reset and relocalized to a different cortical site to reactivate motor control.

Rapid calibration of an intracortical brain-computer interface for people with tetraplegia

OBJECTIVE

Brain-computer interfaces (BCIs) can enable individuals with tetraplegia to communicate and control external devices. Though much progress has been made in improving the speed and robustness of neural control provided by intracortical BCIs, little research has been devoted to minimizing the amount of time spent on decoder calibration.

APPROACH

We investigated the amount of time users needed to calibrate decoders and achieve performance saturation using two markedly different decoding algorithms: the steady-state Kalman filter, and a novel technique using Gaussian process regression (GP-DKF).

MAIN RESULTS

Three people with tetraplegia gained rapid closed-loop neural cursor control and peak, plateaued decoder performance within 3 min of initializing calibration. We also show that a BCI-naïve user (T5) was able to rapidly attain closed-loop neural cursor control with the GP-DKF using self-selected movement imagery on his first-ever day of closed-loop BCI use, acquiring a target 37 s after initiating calibration.

SIGNIFICANCE

These results demonstrate the potential for an intracortical BCI to be used immediately after deployment by people with paralysis, without the need for user learning or extensive system calibration.

Intrinsic Variable Learning for Brain-Machine Interface Control by Human Anterior Intraparietal Cortex

Although animal studies provided significant insights in understanding the neural basis of learning and adaptation, they often cannot dissociate between different learning mechanisms due to the lack of verbal communication. To overcome this limitation, we examined the mechanisms of learning and its limits in a human intracortical brain-machine interface (BMI) paradigm. A tetraplegic participant controlled a 2D computer cursor by modulating single-neuron activity in the anterior intraparietal area (AIP). By perturbing the neuron-to-movement mapping, the participant learned to modulate the activity of the recorded neurons to solve the perturbations by adopting a target re-aiming strategy. However, when no cognitive strategies were adequate to produce correct responses, AIP failed to adapt to perturbations. These findings suggest that learning is constrained by the pre-existing neuronal structure, although it is possible that AIP needs more training time to learn to generate novel activity patterns when cognitive re-adaptation fails to solve the perturbations.

A New Frontier: The Convergence of Nanotechnology, Brain Machine Interfaces, and Artificial Intelligence

A confluence of technological capabilities is creating an opportunity for machine learning and artificial intelligence (AI) to enable "smart" nanoengineered brain machine interfaces (BMI). This new generation of technologies will be able to communicate with the brain in ways that support contextual learning and adaptation to changing functional requirements. This applies to both invasive technologies aimed at restoring neurological function, as in the case of neural prosthesis, as well as non-invasive technologies enabled by signals such as electroencephalograph (EEG). Advances in computation, hardware, and algorithms that learn and adapt in a contextually dependent way will be able to leverage the capabilities that nanoengineering offers the design and functionality of BMI. We explore the enabling capabilities that these devices may exhibit, why they matter, and the state of the technologies necessary to build them. We also discuss a number of open technical challenges and problems that will need to be solved in order to achieve this.

Robust Closed-Loop Control of a Cursor in a Person with Tetraplegia using Gaussian Process Regression

Intracortical brain computer interfaces can enable individuals with paralysis to control external devices through voluntarily modulated brain activity. Decoding quality has been previously shown to degrade with signal nonstationarities-specifically, the changes in the statistics of the data between training and testing data sets. This includes changes to the neural tuning profiles and baseline shifts in firing rates of recorded neurons, as well as nonphysiological noise. While progress has been made toward providing long-term user control via decoder recalibration, relatively little work has been dedicated to making the decoding algorithm more resilient to signal nonstationarities. Here, we describe how principled kernel selection with gaussian process regression can be used within a Bayesian filtering framework to mitigate the effects of commonly encountered nonstationarities. Given a supervised training set of (neural features, intention to move in a direction)-pairs, we use gaussian process regression to predict the intention given the neural data. We apply kernel embedding for each neural feature with the standard radial basis function. The multiple kernels are then summed together across each neural dimension, which allows the kernel to effectively ignore large differences that occur only in a single feature. The summed kernel is used for real-time predictions of the posterior mean and variance under a gaussian process framework. The predictions are then filtered using the discriminative Kalman filter to produce an estimate of the neural intention given the history of neural data. We refer to the multiple kernel approach combined with the discriminative Kalman filter as the MK-DKF. We found that the MK-DKF decoder was more resilient to nonstationarities frequently encountered in-real world settings yet provided similar performance to the currently used Kalman decoder. These results demonstrate a method by which neural decoding can be made more resistant to nonstationarities.

Data-Driven Transducer Design and Identification for Internally-Paced Motor Brain Computer Interfaces: A Review

Brain-Computer Interfaces (BCIs) are systems that establish a direct communication pathway between the users' brain activity and external effectors. They offer the potential to improve the quality of life of motor-impaired patients. Motor BCIs aim to permit severely motor-impaired users to regain limb mobility by controlling orthoses or prostheses. In particular, motor BCI systems benefit patients if the decoded actions reflect the users' intentions with an accuracy that enables them to efficiently interact with their environment. One of the main challenges of BCI systems is to adapt the BCI's signal translation blocks to the user to reach a high decoding accuracy. This paper will review the literature of data-driven and user-specific transducer design and identification approaches and it focuses on internally-paced motor BCIs. In particular, continuous kinematic biomimetic and mental-task decoders are reviewed. Furthermore, static and dynamic decoding approaches, linear and non-linear decoding, offline and real-time identification algorithms are considered. The current progress and challenges related to the design of clinical-compatible motor BCI transducers are additionally discussed.

Exploring the role of striatal D1 and D2 medium spiny neurons in action selection using a virtual robotic framework

The basal ganglia have been hypothesized to be involved in action selection, i.e. resolving competition between simultaneously activated motor programs. It has been shown that the direct pathway facilitates action execution whereas the indirect pathway inhibits it. However, as the pathways are both active during an action, it remains unclear whether their role is co-operative or competitive. In order to investigate this issue, we developed a striatal model consisting of D1 and D2 medium spiny neurons (MSNs) and interfaced it to a simulated robot moving in an environment. We demonstrate that this model is able to reproduce key behavioral features of several experiments involving optogenetic manipulation of the striatum, such as freezing and ambulation. We then investigate the interaction of D1- and D2-MSNs. We find that their fundamental relationship is co-operative within a channel and competitive between channels; this turns out to be crucial for action selection. However, individual pairs of D1- and D2-MSNs may exhibit predominantly competition or co-operation depending on their distance, and D1- and D2-MSNs population activity can alternate between co-operation and competition modes during a stimulation. Additionally, our results show that D2-D2 connectivity between channels is necessary for effective resolution of competition; in its absence, a conflict of two motor programs typically results in neither being selected.

The Cybathlon BCI race: Successful longitudinal mutual learning with two tetraplegic users

This work aims at corroborating the importance and efficacy of mutual learning in motor imagery (MI) brain–computer interface (BCI) by leveraging the insights obtained through our participation in the BCI race of the Cybathlon event. We hypothesized that, contrary to the popular trend of focusing mostly on the machine learning aspects of MI BCI training, a comprehensive mutual learning methodology that reinstates the three learning pillars (at the machine, subject, and application level) as equally significant could lead to a BCI–user symbiotic system able to succeed in real-world scenarios such as the Cybathlon event. Two severely impaired participants with chronic spinal cord injury (SCI), were trained following our mutual learning approach to control their avatar in a virtual BCI race game. The competition outcomes substantiate the effectiveness of this type of training. Most importantly, the present study is one among very few to provide multifaceted evidence on the efficacy of subject learning during BCI training. Learning correlates could be derived at all levels of the interface—application, BCI output, and electroencephalography (EEG) neuroimaging—with two end-users, sufficiently longitudinal evaluation, and, importantly, under real-world and even adverse conditions.

Noninvasive brain–computer interface (BCI) based on imagined movements can restore functions lost to disability by enabling spontaneous, direct brain control of external devices without risks associated with surgical implantation of neural interfaces. We hypothesized that, contrary to the popular trend of focusing on the machine learning aspects of BCI training, a comprehensive mutual learning methodology could strongly promote users’ acquisition of BCI skills and lead to a system able to succeed in real-world scenarios such as the Cybathlon event, the first international competition for disabled pilots in control of assistive technology. Two severely impaired participants with chronic spinal cord injury (SCI) were trained following our mutual learning approach to control their avatar in a virtual BCI race game. The evolution of the training process, including competition outcomes (gold medal, tournament record), substantiates the effectiveness of this type of training. Most importantly, the present study provides multifaceted evidence on the efficacy of subject learning during BCI training. Learning correlates could be derived at all levels of the interface—application, BCI output, and electroencephalography—with two end-users, longitudinal evaluation, and, importantly, under real-world and even adverse conditions.

Review: Human Intracortical Recording and Neural Decoding for Brain-Computer Interfaces

Brain-computer interfaces (BCIs) use neural information recorded from the brain for the voluntary control of external devices. The development of BCI systems has largely focused on improving functional independence for individuals with severe motor impairments, including providing tools for communication and mobility. In this review, we describe recent advances in intracortical BCI technology and provide potential directions for further research.

Gesture Decoding Using ECoG Signals from Human Sensorimotor Cortex: A Pilot Study

Electrocorticography (ECoG) has been demonstrated as a promising neural signal source for developing brain-machine interfaces (BMIs). However, many concerns about the disadvantages brought by large craniotomy for implanting the ECoG grid limit the clinical translation of ECoG-based BMIs. In this study, we collected clinical ECoG signals from the sensorimotor cortex of three epileptic participants when they performed hand gestures. The ECoG power spectrum in hybrid frequency bands was extracted to build a synchronous real-time BMI system. High decoding accuracy of the three gestures was achieved in both offline analysis (85.7%, 84.5%, and 69.7%) and online tests (80% and 82%, tested on two participants only). We found that the decoding performance was maintained even with a subset of channels selected by a greedy algorithm. More importantly, these selected channels were mostly distributed along the central sulcus and clustered in the area of 3 interelectrode squares. Our findings of the reduced and clustered distribution of ECoG channels further supported the feasibility of clinically implementing the ECoG-based BMI system for the control of hand gestures.

Evolutionary algorithm optimization of biological learning parameters in a biomimetic neuroprosthesis

Biomimetic simulation permits neuroscientists to better understand the complex neuronal dynamics of the brain. Embedding a biomimetic simulation in a closed-loop neuroprosthesis, which can read and write signals from the brain, will permit applications for amelioration of motor, psychiatric, and memory-related brain disorders. Biomimetic neuroprostheses require real-time adaptation to changes in the external environment, thus constituting an example of a dynamic data-driven application system. As model fidelity increases, so does the number of parameters and the complexity of finding appropriate parameter configurations. Instead of adapting synaptic weights via machine learning, we employed major biological learning methods: spike-timing dependent plasticity and reinforcement learning. We optimized the learning metaparameters using evolutionary algorithms, which were implemented in parallel and which used an island model approach to obtain sufficient speed. We employed these methods to train a cortical spiking model to utilize macaque brain activity, indicating a selected target, to drive a virtual musculoskeletal arm with realistic anatomical and biomechanical properties to reach to that target. The optimized system was able to reproduce macaque data from a comparable experimental motor task. These techniques can be used to efficiently tune the parameters of multiscale systems, linking realistic neuronal dynamics to behavior, and thus providing a useful tool for neuroscience and neuroprosthetics.

Feedback for reinforcement learning based brain-machine interfaces using confidence metrics

OBJECTIVE

For brain-machine interfaces (BMI) to be used in activities of daily living by paralyzed individuals, the BMI should be as autonomous as possible. One of the challenges is how the feedback is extracted and utilized in the BMI. Our long-term goal is to create autonomous BMIs that can utilize an evaluative feedback from the brain to update the decoding algorithm and use it intelligently in order to adapt the decoder. In this study, we show how to extract the necessary evaluative feedback from a biologically realistic (synthetic) source, use both the quantity and the quality of the feedback, and how that feedback information can be incorporated into a reinforcement learning (RL) controller architecture to maximize its performance.

APPROACH

Motivated by the perception-action-reward cycle (PARC) in the brain which links reward for cognitive decision making and goal-directed behavior, we used a reward-based RL architecture named Actor-Critic RL as the model. Instead of using an error signal towards building an autonomous BMI, we envision to use a reward signal from the nucleus accumbens (NAcc) which plays a key role in the linking of reward to motor behaviors. To deal with the complexity and non-stationarity of biological reward signals, we used a confidence metric which was used to indicate the degree of feedback accuracy. This confidence was added to the Actor's weight update equation in the RL controller architecture. If the confidence was high (>0.2), the BMI decoder used this feedback to update its parameters. However, when the confidence was low, the BMI decoder ignored the feedback and did not update its parameters. The range between high confidence and low confidence was termed as the 'ambiguous' region. When the feedback was within this region, the BMI decoder updated its weight at a lower rate than when fully confident, which was decided by the confidence. We used two biologically realistic models to generate synthetic data for MI (Izhikevich model) and NAcc (Humphries model) to validate proposed controller architecture.

MAIN RESULTS

In this work, we show how the overall performance of the BMI was improved by using a threshold close to the decision boundary to reject erroneous feedback. Additionally, we show the stability of the system improved when the feedback was used with a threshold.

SIGNIFICANCE

The result of this study is a step towards making BMIs autonomous. While our method is not fully autonomous, the results demonstrate that extensive training times necessary at the beginning of each BMI session can be significantly decreased. In our approach, decoder training time was only limited to 10 trials in the first BMI session. Subsequent sessions used previous session weights to initialize the decoder. We also present a method where the use of a threshold can be applied to any decoder with a feedback signal that is less than perfect so that erroneous feedback can be avoided and the stability of the system can be increased.

Clinical applications of penetrating neural interfaces and Utah Electrode Array technologies

This paper briefly describes some of the recent progress in the development of penetrating microelectrode arrays and highlights the use of two of these devices, Utah electrode arrays and Utah slanted electrode arrays, in two therapeutic interventions: recording volitional skeletal motor commands from the central nervous system, and recording motor commands and evoking somatosensory percepts in the peripheral nervous system (PNS). The paper also briefly explores other potential sites for microelectrode array interventions that could be profitably pursued and that could have important consequences in enhancing the quality of life of patients that has been compromised by disorders of the central and PNSs.

What Turns Assistive into Restorative Brain-Machine Interfaces?

Brain-machine interfaces (BMI) may support motor impaired patients during activities of daily living by controlling external devices such as prostheses (assistive BMI). Moreover, BMIs are applied in conjunction with robotic orthoses for rehabilitation of lost motor function via neurofeedback training (restorative BMI). Using assistive BMI in a rehabilitation context does not automatically turn them into restorative devices. This perspective article suggests key features of restorative BMI and provides the supporting evidence: In summary, BMI may be referred to as restorative tools when demonstrating subsequently (i) operant learning and progressive evolution of specific brain states/dynamics, (ii) correlated modulations of functional networks related to the therapeutic goal, (iii) subsequent improvement in a specific task, and (iv) an explicit correlation between the modulated brain dynamics and the achieved behavioral gains. Such findings would provide the rationale for translating BMI-based interventions into clinical settings for reinforcement learning and motor rehabilitation following stroke.

Control of Redundant Kinematic Degrees of Freedom in a Closed-Loop Brain-Machine Interface

Brain-machine interface (BMI) systems use signals acquired from the brain to directly control the movement of an actuator, such as a computer cursor or a robotic arm, with the goal of restoring motor function lost due to injury or disease of the nervous system. In BMIs with kinematically redundant actuators, the combination of the task goals and the system under neural control can allow for many equally optimal task solutions. The extent to which kinematically redundant degrees of freedom (DOFs) in a BMI system may be under direct neural control is unknown. To address this question, a Kalman filter was used to decode single- and multi-unit cortical neural activity of two macaque monkeys into the joint velocities of a virtual four-link kinematic chain. Subjects completed movements of the chain's endpoint to instructed target locations within a two-dimensional plane. This system was kinematically redundant for an endpoint movement task, as four DOFs were used to manipulate the 2-D endpoint position. Both subjects successfully performed the task and improved with practice by producing faster endpoint velocity control signals. Kinematic redundancy allowed null movements whereby the individual links of the chain could move in a way that cancels out and does not result in endpoint movement. As the subjects became more proficient at controlling the chain, the amount of null movement also increased. Task performance suffered when the links of the kinematic chain were hidden and only the endpoint was visible. Furthermore, all four DOFs of the joint-velocity control space exhibited task-relevant modulation. The relative usage of each DOF depended on the configuration of the chain, and trials in which the less-prominent DOFs were utilized also had better task performance. Overall, these results indicate that the subjects incorporated the redundant components of the control space into their control strategy. Future BMI systems with kinematic redundancy, such as exoskeletal systems or anthropomorphic robotic arms, may benefit from allowing neural control over redundant configuration dimensions as well as the end-effector.

A Review of Control Strategies in Closed-Loop Neuroprosthetic Systems

It has been widely recognized that closed-loop neuroprosthetic systems achieve more favorable outcomes for users then equivalent open-loop devices. Improved performance of tasks, better usability, and greater embodiment have all been reported in systems utilizing some form of feedback. However, the interdisciplinary work on neuroprosthetic systems can lead to miscommunication due to similarities in well-established nomenclature in different fields. Here we present a review of control strategies in existing experimental, investigational and clinical neuroprosthetic systems in order to establish a baseline and promote a common understanding of different feedback modes and closed-loop controllers. The first section provides a brief discussion of feedback control and control theory. The second section reviews the control strategies of recent Brain Machine Interfaces, neuromodulatory implants, neuroprosthetic systems, and assistive neurorobotic devices. The final section examines the different approaches to feedback in current neuroprosthetic and neurorobotic systems.

Context-aware adaptive spelling in motor imagery BCI

OBJECTIVE

This work presents a first motor imagery-based, adaptive brain-computer interface (BCI) speller, which is able to exploit application-derived context for improved, simultaneous classifier adaptation and spelling. Online spelling experiments with ten able-bodied users evaluate the ability of our scheme, first, to alleviate non-stationarity of brain signals for restoring the subject's performances, second, to guide naive users into BCI control avoiding initial offline BCI calibration and, third, to outperform regular unsupervised adaptation.

APPROACH

Our co-adaptive framework combines the BrainTree speller with smooth-batch linear discriminant analysis adaptation. The latter enjoys contextual assistance through BrainTree's language model to improve online expectation-maximization maximum-likelihood estimation.

MAIN RESULTS

Our results verify the possibility to restore single-sample classification and BCI command accuracy, as well as spelling speed for expert users. Most importantly, context-aware adaptation performs significantly better than its unsupervised equivalent and similar to the supervised one. Although no significant differences are found with respect to the state-of-the-art PMean approach, the proposed algorithm is shown to be advantageous for 30% of the users.

SIGNIFICANCE

We demonstrate the possibility to circumvent supervised BCI recalibration, saving time without compromising the adaptation quality. On the other hand, we show that this type of classifier adaptation is not as efficient for BCI training purposes.

Implications of the Dependence of Neuronal Activity on Neural Network States for the Design of Brain-Machine Interfaces

Brain-machine interfaces (BMIs) can improve the quality of life of patients with sensory and motor disabilities by both decoding motor intentions expressed by neural activity, and by encoding artificially sensed information into patterns of neural activity elicited by causal interventions on the neural tissue. Yet, current BMIs can exchange relatively small amounts of information with the brain. This problem has proved difficult to overcome by simply increasing the number of recording or stimulating electrodes, because trial-to-trial variability of neural activity partly arises from intrinsic factors (collectively known as the network state) that include ongoing spontaneous activity and neuromodulation, and so is shared among neurons. Here we review recent progress in characterizing the state dependence of neural responses, and in particular of how neural responses depend on endogenous slow fluctuations of network excitability. We then elaborate on how this knowledge may be used to increase the amount of information that BMIs exchange with brain. Knowledge of network state can be used to fine-tune the stimulation pattern that should reliably elicit a target neural response used to encode information in the brain, and to discount part of the trial-by-trial variability of neural responses, so that they can be decoded more accurately.

Reinforcement learning of self-regulated sensorimotor β-oscillations improves motor performance

Self-regulation of sensorimotor oscillations is currently researched in neurorehabilitation, e.g. for priming subsequent physiotherapy in stroke patients, and may be modulated by neurofeedback or transcranial brain stimulation. It has still to be demonstrated, however, whether and under which training conditions such brain self-regulation could also result in motor gains. Thirty-two right-handed, healthy subjects participated in a three-day intervention during which they performed 462 trials of kinesthetic motor-imagery while a brain-robot interface (BRI) turned event-related β-band desynchronization of the left sensorimotor cortex into the opening of the right hand by a robotic orthosis. Different training conditions were compared in a parallel-group design: (i) adaptive classifier thresholding and contingent feedback, (ii) adaptive classifier thresholding and non-contingent feedback, (iii) non-adaptive classifier thresholding and contingent feedback, and (iv) non-adaptive classifier thresholding and non-contingent feedback. We studied the task-related cortical physiology with electroencephalography and the behavioral performance in a subsequent isometric motor task. Contingent neurofeedback and adaptive classifier thresholding were critical for learning brain self-regulation which, in turn, led to behavioral gains after the intervention. The acquired skill for sustained sensorimotor β-desynchronization correlated significantly with subsequent motor improvement. Operant learning of brain self-regulation with a BRI may offer a therapeutic perspective for severely affected stroke patients lacking residual hand function.

Robust Brain-Machine Interface Design Using Optimal Feedback Control Modeling and Adaptive Point Process Filtering

Much progress has been made in brain-machine interfaces (BMI) using decoders such as Kalman filters and finding their parameters with closed-loop decoder adaptation (CLDA). However, current decoders do not model the spikes directly, and hence may limit the processing time-scale of BMI control and adaptation. Moreover, while specialized CLDA techniques for intention estimation and assisted training exist, a unified and systematic CLDA framework that generalizes across different setups is lacking. Here we develop a novel closed-loop BMI training architecture that allows for processing, control, and adaptation using spike events, enables robust control and extends to various tasks. Moreover, we develop a unified control-theoretic CLDA framework within which intention estimation, assisted training, and adaptation are performed. The architecture incorporates an infinite-horizon optimal feedback-control (OFC) model of the brain's behavior in closed-loop BMI control, and a point process model of spikes. The OFC model infers the user's motor intention during CLDA-a process termed intention estimation. OFC is also used to design an autonomous and dynamic assisted training technique. The point process model allows for neural processing, control and decoder adaptation with every spike event and at a faster time-scale than current decoders; it also enables dynamic spike-event-based parameter adaptation unlike current CLDA methods that use batch-based adaptation on much slower adaptation time-scales. We conducted closed-loop experiments in a non-human primate over tens of days to dissociate the effects of these novel CLDA components. The OFC intention estimation improved BMI performance compared with current intention estimation techniques. OFC assisted training allowed the subject to consistently achieve proficient control. Spike-event-based adaptation resulted in faster and more consistent performance convergence compared with batch-based methods, and was robust to parameter initialization. Finally, the architecture extended control to tasks beyond those used for CLDA training. These results have significant implications towards the development of clinically-viable neuroprosthetics.

Cortical Spiking Network Interfaced with Virtual Musculoskeletal Arm and Robotic Arm

Embedding computational models in the physical world is a critical step towards constraining their behavior and building practical applications. Here we aim to drive a realistic musculoskeletal arm model using a biomimetic cortical spiking model, and make a robot arm reproduce the same trajectories in real time. Our cortical model consisted of a 3-layered cortex, composed of several hundred spiking model-neurons, which display physiologically realistic dynamics. We interconnected the cortical model to a two-joint musculoskeletal model of a human arm, with realistic anatomical and biomechanical properties. The virtual arm received muscle excitations from the neuronal model, and fed back proprioceptive information, forming a closed-loop system. The cortical model was trained using spike timing-dependent reinforcement learning to drive the virtual arm in a 2D reaching task. Limb position was used to simultaneously control a robot arm using an improved network interface. Virtual arm muscle activations responded to motoneuron firing rates, with virtual arm muscles lengths encoded via population coding in the proprioceptive population. After training, the virtual arm performed reaching movements which were smoother and more realistic than those obtained using a simplistic arm model. This system provided access to both spiking network properties and to arm biophysical properties, including muscle forces. The use of a musculoskeletal virtual arm and the improved control system allowed the robot arm to perform movements which were smoother than those reported in our previous paper using a simplistic arm. This work provides a novel approach consisting of bidirectionally connecting a cortical model to a realistic virtual arm, and using the system output to drive a robotic arm in real time. Our techniques are applicable to the future development of brain neuroprosthetic control systems, and may enable enhanced brain-machine interfaces with the possibility for finer control of limb prosthetics.

A cortical-spinal prosthesis for targeted limb movement in paralysed primate avatars

Motor paralysis is among the most disabling aspects of injury to the central nervous system. Here we develop and test a target-based cortical-spinal neural prosthesis that employs neural activity recorded from pre-motor neurons to control limb movements in functionally paralyzed primate avatars. Given the complexity by which muscle contractions are naturally controlled, we approach the problem of eliciting goal-directed limb movement in paralyzed animals by focusing on the intended targets of movement rather than their intermediate trajectories. We then match this information in real-time with spinal cord and muscle stimulation parameters that produce free planar limb movements to those intended target locations. We demonstrate that both the decoded activities of pre-motor populations and their adaptive responses can be used, after brief training, to effectively direct an avatar’s limb to distinct targets variably displayed on a screen. These findings advance the future possibility of reconstituting targeted limb movement in paralyzed subjects.

Beamforming approaches for untethered, ultrasonic neural dust motes for cortical recording: a simulation study

In this paper, we examine the use of beamforming techniques to interrogate a multitude of neural implants in a distributed, ultrasound-based intra-cortical recording platform known as Neural Dust. We propose a general framework to analyze system design tradeoffs in the ultrasonic beamformer that extracts neural signals from modulated ultrasound waves that are backscattered by free-floating neural dust (ND) motes. Simulations indicate that high-resolution linearly-constrained minimum variance beamforming sufficiently suppresses interference from unselected ND motes and can be incorporated into the ND-based cortical recording system.

Toward Building Hybrid Biological/in silico Neural Networks for Motor Neuroprosthetic Control

In this article, we introduce the Bioinspired Neuroprosthetic Design Environment (BNDE) as a practical platform for the development of novel brain–machine interface (BMI) controllers, which are based on spiking model neurons. We built the BNDE around a hard real-time system so that it is capable of creating simulated synapses from extracellularly recorded neurons to model neurons. In order to evaluate the practicality of the BNDE for neuroprosthetic control experiments, a novel, adaptive BMI controller was developed and tested using real-time closed-loop simulations. The present controller consists of two in silico medium spiny neurons, which receive simulated synaptic inputs from recorded motor cortical neurons. In the closed-loop simulations, the recordings from the cortical neurons were imitated using an external, hardware-based neural signal synthesizer. By implementing a reward-modulated spike timing-dependent plasticity rule, the controller achieved perfect target reach accuracy for a two-target reaching task in one-dimensional space. The BNDE combines the flexibility of software-based spiking neural network (SNN) simulations with powerful online data visualization tools and is a low-cost, PC-based, and all-in-one solution for developing neurally inspired BMI controllers. We believe that the BNDE is the first implementation, which is capable of creating hybrid biological/in silico neural networks for motor neuroprosthetic control and utilizes multiple CPU cores for computationally intensive real-time SNN simulations.

Velocity neurons improve performance more than goal or position neurons do in a simulated closed-loop BCI arm-reaching task

Brain-Computer Interfaces (BCIs) that convert brain-recorded neural signals into intended movement commands could eventually be combined with Functional Electrical Stimulation to allow individuals with Spinal Cord Injury to regain effective and intuitive control of their paralyzed limbs. To accelerate the development of such an approach, we developed a model of closed-loop BCI control of arm movements that (1) generates realistic arm movements (based on experimentally measured, visually-guided movements with real-time error correction), (2) simulates cortical neurons with firing properties consistent with literature reports, and (3) decodes intended movements from the noisy neural ensemble. With this model we explored (1) the relative utility of neurons tuned for different movement parameters (position, velocity, and goal) and (2) the utility of recording from larger numbers of neurons-critical issues for technology development and for determining appropriate brain areas for recording. We simulated arm movements that could be practically restored to individuals with severe paralysis, i.e., movements from an armrest to a volume in front of the person. Performance was evaluated by calculating the smallest movement endpoint target radius within which the decoded cursor position could dwell for 1 s. Our results show that goal, position, and velocity neurons all contribute to improve performance. However, velocity neurons enabled smaller targets to be reached in shorter amounts of time than goal or position neurons. Increasing the number of neurons also improved performance, although performance saturated at 30-50 neurons for most neuron types. Overall, our work presents a closed-loop BCI simulator that models error corrections and the firing properties of various movement-related neurons that can be easily modified to incorporate different neural properties. We anticipate that this kind of tool will be important for development of future BCIs.

Reinforcement learning of self-regulated β-oscillations for motor restoration in chronic stroke

Neurofeedback training of Motor imagery (MI)-related brain-states with brain-computer/brain-machine interfaces (BCI/BMI) is currently being explored as an experimental intervention prior to standard physiotherapy to improve the motor outcome of stroke rehabilitation. The use of BCI/BMI technology increases the adherence to MI training more efficiently than interventions with sham or no feedback. Moreover, pilot studies suggest that such a priming intervention before physiotherapy might—like some brain stimulation techniques—increase the responsiveness of the brain to the subsequent physiotherapy, thereby improving the general clinical outcome. However, there is little evidence up to now that these BCI/BMI-based interventions have achieved operate conditioning of specific brain states that facilitate task-specific functional gains beyond the practice of primed physiotherapy. In this context, we argue that BCI/BMI technology provides a valuable neurofeedback tool for rehabilitation but needs to aim at physiological features relevant for the targeted behavioral gain. Moreover, this therapeutic intervention has to be informed by concepts of reinforcement learning to develop its full potential. Such a refined neurofeedback approach would need to address the following issues: (1) Defining a physiological feedback target specific to the intended behavioral gain, e.g., β-band oscillations for cortico-muscular communication. This targeted brain state could well be different from the brain state optimal for the neurofeedback task, e.g., α-band oscillations for differentiating MI from rest; (2) Selecting a BCI/BMI classification and thresholding approach on the basis of learning principles, i.e., balancing challenge and reward of the neurofeedback task instead of maximizing the classification accuracy of the difficulty level device; and (3) Adjusting the difficulty level in the course of the training period to account for the cognitive load and the learning experience of the participant. Here, we propose a comprehensive neurofeedback strategy for motor restoration after stroke that addresses these aspects, and provide evidence for the feasibility of the suggested approach by demonstrating that dynamic threshold adaptation based on reinforcement learning may lead to frequency-specific operant conditioning of β-band oscillations paralleled by task-specific motor improvement; a proposal that requires investigation in a larger cohort of stroke patients.

Generalized analog thresholding for spike acquisition at ultralow sampling rates

Efficient spike acquisition techniques are needed to bridge the divide from creating large multielectrode arrays (MEA) to achieving whole-cortex electrophysiology. In this paper, we introduce generalized analog thresholding (gAT), which achieves millisecond temporal resolution with sampling rates as low as 10 Hz. Consider the torrent of data from a single 1,000-channel MEA, which would generate more than 3 GB/min using standard 30-kHz Nyquist sampling. Recent neural signal processing methods based on compressive sensing still require Nyquist sampling as a first step and use iterative methods to reconstruct spikes. Analog thresholding (AT) remains the best existing alternative, where spike waveforms are passed through an analog comparator and sampled at 1 kHz, with instant spike reconstruction. By generalizing AT, the new method reduces sampling rates another order of magnitude, detects more than one spike per interval, and reconstructs spike width. Unlike compressive sensing, the new method reveals a simple closed-form solution to achieve instant (noniterative) spike reconstruction. The base method is already robust to hardware nonidealities, including realistic quantization error and integration noise. Because it achieves these considerable specifications using hardware-friendly components like integrators and comparators, generalized AT could translate large-scale MEAs into implantable devices for scientific investigation and medical technology.

Electrical stimuli in the central nervous system microenvironment

Electrical stimulation to manipulate the central nervous system (CNS) has been applied as early as the 1750s to produce visual sensations of light. Deep brain stimulation (DBS), cochlear implants, visual prosthetics, and functional electrical stimulation (FES) are being applied in the clinic to treat a wide array of neurological diseases, disorders, and injuries. This review describes the history of electrical stimulation of the CNS microenvironment; recent advances in electrical stimulation of the CNS, including DBS to treat essential tremor, Parkinson's disease, and depression; FES for the treatment of spinal cord injuries; and alternative electrical devices to restore vision and hearing via neuroprosthetics (retinal and cochlear implants). It also discusses the role of electrical cues during development and following injury and, importantly, manipulation of these endogenous cues to support regeneration of neural tissue.

Closed-loop rehabilitation of age-related cognitive disorders

Cognitive deficits are common in older adults, as a result of both the natural aging process and neurodegenerative disease. Although medical advancements have successfully prolonged the human lifespan, the challenge of remediating cognitive aging remains. The authors discuss the current state of cognitive therapeutic interventions and then present the need for development and validation of more powerful neurocognitive therapeutics. They propose that the next generation of interventions be implemented as closed-loop systems that target specific neural processing deficits, incorporate quantitative feedback to the individual and clinician, and are personalized to the individual's neurocognitive capacities using real-time performance-adaptive algorithms. This approach should be multimodal and seamlessly integrate other treatment approaches, including neurofeedback and transcranial electrical stimulation. This novel approach will involve the generation of software that engages the individual in an immersive and enjoyable game-based interface, integrated with advanced biosensing hardware, to maximally harness plasticity and assure adherence. Introducing such next-generation closed-loop neurocognitive therapeutics into the mainstream of our mental health care system will require the combined efforts of clinicians, neuroscientists, bioengineers, software game developers, and industry and policy makers working together to meet the challenges and opportunities of translational neuroscience in the 21st century.

Brain-machine interfaces in neurorehabilitation of stroke

Stroke is among the leading causes of long-term disabilities leaving an increasing number of people with cognitive, affective and motor impairments depending on assistance in their daily life. While function after stroke can significantly improve in the first weeks and months, further recovery is often slow or non-existent in the more severe cases encompassing 30-50% of all stroke victims. The neurobiological mechanisms underlying recovery in those patients are incompletely understood. However, recent studies demonstrated the brain's remarkable capacity for functional and structural plasticity and recovery even in severe chronic stroke. As all established rehabilitation strategies require some remaining motor function, there is currently no standardized and accepted treatment for patients with complete chronic muscle paralysis. The development of brain-machine interfaces (BMIs) that translate brain activity into control signals of computers or external devices provides two new strategies to overcome stroke-related motor paralysis. First, BMIs can establish continuous high-dimensional brain-control of robotic devices or functional electric stimulation (FES) to assist in daily life activities (assistive BMI). Second, BMIs could facilitate neuroplasticity, thus enhancing motor learning and motor recovery (rehabilitative BMI). Advances in sensor technology, development of non-invasive and implantable wireless BMI-systems and their combination with brain stimulation, along with evidence for BMI systems' clinical efficacy suggest that BMI-related strategies will play an increasing role in neurorehabilitation of stroke.

Combining decoder design and neural adaptation in brain-machine interfaces

Brain-machine interfaces (BMIs) aim to help people with paralysis by decoding movement-related neural signals into control signals for guiding computer cursors, prosthetic arms, and other assistive devices. Despite compelling laboratory experiments and ongoing FDA pilot clinical trials, system performance, robustness, and generalization remain challenges. We provide a perspective on how two complementary lines of investigation, that have focused on decoder design and neural adaptation largely separately, could be brought together to advance BMIs. This BMI paradigm should also yield new scientific insights into the function and dysfunction of the nervous system.

The role of muscle synergies in myoelectric control: trends and challenges for simultaneous multifunction control

Myoelectric control is filled with potential to significantly change human-robot interaction due to the ability to non-invasively measure human motion intent. However, current control schemes have struggled to achieve the robust performance that is necessary for use in commercial applications. As demands in myoelectric control trend toward simultaneous multifunctional control, multi-muscle coordinations, or synergies, play larger roles in the success of the control scheme. Detecting and refining patterns in muscle activations robust to the high variance and transient changes associated with surface electromyography is essential for efficient, user-friendly control. This article reviews the role of muscle synergies in myoelectric control schemes by dissecting each component of the scheme with respect to associated challenges for achieving robust simultaneous control of myoelectric interfaces. Electromyography recording details, signal feature extraction, pattern recognition and motor learning based control schemes are considered, and future directions are proposed as steps toward fulfilling the potential of myoelectric control in clinically and commercially viable applications.

Closed-loop decoder adaptation shapes neural plasticity for skillful neuroprosthetic control

Neuroplasticity may play a critical role in developing robust, naturally controlled neuroprostheses. This learning, however, is sensitive to system changes such as the neural activity used for control. The ultimate utility of neuroplasticity in real-world neuroprostheses is thus unclear. Adaptive decoding methods hold promise for improving neuroprosthetic performance in nonstationary systems. Here, we explore the use of decoder adaptation to shape neuroplasticity in two scenarios relevant for real-world neuroprostheses: nonstationary recordings of neural activity and changes in control context. Nonhuman primates learned to control a cursor to perform a reaching task using semistationary neural activity in two contexts: with and without simultaneous arm movements. Decoder adaptation was used to improve initial performance and compensate for changes in neural recordings. We show that beneficial neuroplasticity can occur alongside decoder adaptation, yielding performance improvements, skill retention, and resistance to interference from native motor networks. These results highlight the utility of neuroplasticity for real-world neuroprostheses.

Acute human brain responses to intracortical microelectrode arrays: challenges and future prospects

The emerging field of neuroprosthetics is focused on the development of new therapeutic interventions that will be able to restore some lost neural function by selective electrical stimulation or by harnessing activity recorded from populations of neurons. As more and more patients benefit from these approaches, the interest in neural interfaces has grown significantly and a new generation of penetrating microelectrode arrays are providing unprecedented access to the neurons of the central nervous system (CNS). These microelectrodes have active tip dimensions that are similar in size to neurons and because they penetrate the nervous system, they provide selective access to these cells (within a few microns). However, the very long-term viability of chronically implanted microelectrodes and the capability of recording the same spiking activity over long time periods still remain to be established and confirmed in human studies. Here we review the main responses to acute implantation of microelectrode arrays, and emphasize that it will become essential to control the neural tissue damage induced by these intracortical microelectrodes in order to achieve the high clinical potentials accompanying this technology.

Decoding methods for neural prostheses: where have we reached?

This article reviews advances in decoding methods for brain-machine interfaces (BMIs). Recent work has focused on practical considerations for future clinical deployment of prosthetics. This review is organized by open questions in the field such as what variables to decode, how to design neural tuning models, which neurons to select, how to design models of desired actions, how to learn decoder parameters during prosthetic operation, and how to adapt to changes in neural signals and neural tuning. The concluding discussion highlights the need to design and test decoders within the context of their expected use and the need to answer the question of how much control accuracy is good enough for a prosthetic.

Personalized neuroprosthetics

Decades of technological developments have populated the field of neuroprosthetics with myriad replacement strategies, neuromodulation therapies, and rehabilitation procedures to improve the quality of life for individuals with neuromotor disorders. Despite the few but impressive clinical successes, and multiple breakthroughs in animal models, neuroprosthetic technologies remain mainly confined to sophisticated laboratory environments. We summarize the core principles and latest achievements in neuroprosthetics, but also address the challenges that lie along the path toward clinical fruition. We propose a pragmatic framework to personalize neurotechnologies and rehabilitation for patient-specific impairments to achieve the timely dissemination of neuroprosthetic medicine.

The receptive field is dead. Long live the receptive field?

Advances in experimental techniques, including behavioral paradigms using rich stimuli under closed loop conditions and the interfacing of neural systems with external inputs and outputs, reveal complex dynamics in the neural code and require a revisiting of standard concepts of representation. High-throughput recording and imaging methods along with the ability to observe and control neuronal subpopulations allow increasingly detailed access to the neural circuitry that subserves neural representations and the computations they support. How do we harness theory to build biologically grounded models of complex neural function?

Subject-oriented training for motor imagery brain-computer interfaces

Successful operation of motor imagery (MI)-based brain-computer interfaces (BCI) requires mutual adaptation between the human subject and the BCI. Traditional training methods, as well as more recent ones based on co-adaptation, have mainly focused on the machine-learning aspects of BCI training. This work presents a novel co-adaptive training protocol shifting the focus on subject-related performances and the optimal accommodation of the interactions between the two learning agents of the BCI loop. Preliminary results with 8 able-bodied individuals demonstrate that the proposed method has been able to bring 3 naive users into control of a MI BCI within a few runs and to improve the BCI performances of 3 experienced BCI users by an average of 0.36 bits/sec.