Behavioral and neural correlates of visuomotor adaptation observed through a brain-computer interface in primary motor cortex

In Vivo Mapping of Cortical Columnar Networks in the Monkey with Focal Electrical and Optical Stimulation

There are currently largescale efforts to understand the brain as a connection machine. However, there has been little emphasis on understanding connection patterns between functionally specific cortical columns. Here, we review development and application of focal electrical and optical stimulation methods combined with optical imaging and fMRI mapping in the non-human primate. These new approaches, when applied systematically on a large scale, will elucidate functionally specific intra-areal and inter-areal network connection patterns. Such functionally specific network data can provide accurate views of brain network topology.

Single-trial decoding of intended eye movement goals from lateral prefrontal cortex neural ensembles

Neurons in the lateral prefrontal cortex (LPFC) encode sensory and cognitive signals, as well as commands for goal-directed actions. Therefore, the LPFC might be a good signal source for a goal-selection brain-computer interface (BCI) that decodes the intended goal of a motor action previous to its execution. As a first step in the development of a goal-selection BCI, we set out to determine if we could decode simple behavioral intentions to direct gaze to eight different locations in space from single-trial LPFC neural activity. We recorded neuronal spiking activity from microelectrode arrays implanted in area 8A of the LPFC of two adult macaques while they made visually guided saccades to one of eight targets in a center-out task. Neuronal activity encoded target location immediately after target presentation, during a delay epoch, during the execution of the saccade, and every combination thereof. Many (40%) of the neurons that encoded target location during multiple epochs preferred different locations during different epochs. Despite heterogeneous and dynamic responses, the neuronal feature set that best predicted target location was the averaged firing rates from the entire trial and it was best classified using linear discriminant analysis (63.6-96.9% in 12 sessions, mean 80.3%; information transfer rate: 21-59, mean 32.8 bits/min). Our results demonstrate that it is possible to decode intended saccade target location from single-trial LPFC activity and suggest that the LPFC is a suitable signal source for a goal-selection cognitive BCI.

A stabilized dual Kalman filter for adaptive tracking of brain-computer interface decoding parameters

Neural prosthetics are a promising technology for alleviating paralysis by actuating devices directly from the intention to move. Typical implementations of these devices require a calibration session to define decoding parameters that map recorded neural activity into movement of the device. However, a major factor limiting the clinical deployment of this technology is stability: with fixed decoding parameters, control of the prosthetic device has been shown to degrade over time. Here we apply a dual estimation procedure to adaptively capture changes in decoding parameters. In simulation, we find that our stabilized dual Kalman filter can run autonomously for hundreds of thousands of trials with little change in performance. Further, when we apply our algorithm off-line to estimate arm trajectories from neural data recorded over five consecutive days, we find that it outperforms a static Kalman filter, even when it is re-calibrated at the beginning of each day.

The impact of command signal power distribution, processing delays, and speed scaling on neurally-controlled devices

OBJECTIVE

Decoding algorithms for brain-machine interfacing (BMI) are typically only optimized to reduce the magnitude of decoding errors. Our goal was to systematically quantify how four characteristics of BMI command signals impact closed-loop performance: (1) error magnitude, (2) distribution of different frequency components in the decoding errors, (3) processing delays, and (4) command gain.

APPROACH

To systematically evaluate these different command features and their interactions, we used a closed-loop BMI simulator where human subjects used their own wrist movements to command the motion of a cursor to targets on a computer screen. Random noise with three different power distributions and four different relative magnitudes was added to the ongoing cursor motion in real time to simulate imperfect decoding. These error characteristics were tested with four different visual feedback delays and two velocity gains.

MAIN RESULTS

Participants had significantly more trouble correcting for errors with a larger proportion of low-frequency, slow-time-varying components than they did with jittery, higher-frequency errors, even when the error magnitudes were equivalent. When errors were present, a movement delay often increased the time needed to complete the movement by an order of magnitude more than the delay itself. Scaling down the overall speed of the velocity command can actually speed up target acquisition time when low-frequency errors and delays are present.

SIGNIFICANCE

This study is the first to systematically evaluate how the combination of these four key command signal features (including the relatively-unexplored error power distribution) and their interactions impact closed-loop performance independent of any specific decoding method. The equations we derive relating closed-loop movement performance to these command characteristics can provide guidance on how best to balance these different factors when designing BMI systems. The equations reported here also provide an efficient way to compare a diverse range of decoding options offline.

Recasting brain-machine interface design from a physical control system perspective

With the goal of improving the quality of life for people suffering from various motor control disorders, brain-machine interfaces provide direct neural control of prosthetic devices by translating neural signals into control signals. These systems act by reading motor intent signals directly from the brain and using them to control, for example, the movement of a cursor on a computer screen. Over the past two decades, much attention has been devoted to the decoding problem: how should recorded neural activity be translated into the movement of the cursor? Most approaches have focused on this problem from an estimation standpoint, i.e., decoders are designed to return the best estimate of motor intent possible, under various sets of assumptions about how the recorded neural signals represent motor intent. Here we recast the decoder design problem from a physical control system perspective, and investigate how various classes of decoders lead to different types of physical systems for the subject to control. This framework leads to new interpretations of why certain types of decoders have been shown to perform better than others. These results have implications for understanding how motor neurons are recruited to perform various tasks, and may lend insight into the brain's ability to conceptualize artificial systems.

Brain computer interface learning for systems based on electrocorticography and intracortical microelectrode arrays

A brain-computer interface (BCI) system transforms neural activity into control signals for external devices in real time. A BCI user needs to learn to generate specific cortical activity patterns to control external devices effectively. We call this process BCI learning, and it often requires significant effort and time. Therefore, it is important to study this process and develop novel and efficient approaches to accelerate BCI learning. This article reviews major approaches that have been used for BCI learning, including computer-assisted learning, co-adaptive learning, operant conditioning, and sensory feedback. We focus on BCIs based on electrocorticography and intracortical microelectrode arrays for restoring motor function. This article also explores the possibility of brain modulation techniques in promoting BCI learning, such as electrical cortical stimulation, transcranial magnetic stimulation, and optogenetics. Furthermore, as proposed by recent BCI studies, we suggest that BCI learning is in many ways analogous to motor and cognitive skill learning, and therefore skill learning should be a useful metaphor to model BCI learning.

Encoder-decoder optimization for brain-computer interfaces

Neuroprosthetic brain-computer interfaces are systems that decode neural activity into useful control signals for effectors, such as a cursor on a computer screen. It has long been recognized that both the user and decoding system can adapt to increase the accuracy of the end effector. Co-adaptation is the process whereby a user learns to control the system in conjunction with the decoder adapting to learn the user's neural patterns. We provide a mathematical framework for co-adaptation and relate co-adaptation to the joint optimization of the user's control scheme ("encoding model") and the decoding algorithm's parameters. When the assumptions of that framework are respected, co-adaptation cannot yield better performance than that obtainable by an optimal initial choice of fixed decoder, coupled with optimal user learning. For a specific case, we provide numerical methods to obtain such an optimized decoder. We demonstrate our approach in a model brain-computer interface system using an online prosthesis simulator, a simple human-in-the-loop pyschophysics setup which provides a non-invasive simulation of the BCI setting. These experiments support two claims: that users can learn encoders matched to fixed, optimal decoders and that, once learned, our approach yields expected performance advantages.

Neuroplasticity subserving the operation of brain-machine interfaces

Neuroplasticity is key to the operation of brain machine interfaces (BMIs)-a direct communication pathway between the brain and a man-made computing device. Whereas exogenous BMIs that associate volitional control of brain activity with neurofeedback have been shown to induce long lasting plasticity, endogenous BMIs that use prolonged activity-dependent stimulation--and thus may curtail the time scale that governs natural sensorimotor integration loops--have been shown to induce short lasting plasticity. Here we summarize recent findings from studies using both categories of BMIs, and discuss the fundamental principles that may underlie their operation and the longevity of the plasticity they induce. We draw comparison to plasticity mechanisms known to mediate natural sensorimotor skill learning and discuss principles of homeostatic regulation that may constrain endogenous BMI effects in the adult mammalian brain. We propose that BMIs could be designed to facilitate structural and functional plasticity for the purpose of re-organization of target brain regions and directed augmentation of sensorimotor maps, and suggest possible avenues for future work to maximize their efficacy and viability in clinical applications.

Comprehensive chronic laminar single-unit, multi-unit, and local field potential recording performance with planar single shank electrode arrays

BACKGROUND

Intracortical electrode arrays that can record extracellular action potentials from small, targeted groups of neurons are critical for basic neuroscience research and emerging clinical applications. In general, these electrode devices suffer from reliability and variability issues, which have led to comparative studies of existing and emerging electrode designs to optimize performance. Comparisons of different chronic recording devices have been limited to single-unit (SU) activity and employed a bulk averaging approach treating brain architecture as homogeneous with respect to electrode distribution.

NEW METHOD

In this study, we optimize the methods and parameters to quantify evoked multi-unit (MU) and local field potential (LFP) recordings in eight mice visual cortices.

RESULTS

These findings quantify the large recording differences stemming from anatomical differences in depth and the layer dependent relative changes to SU and MU recording performance over 6-months. For example, performance metrics in Layer V and stratum pyramidale were initially higher than Layer II/III, but decrease more rapidly. On the other hand, Layer II/III maintained recording metrics longer. In addition, chronic changes at the level of layer IV are evaluated using visually evoked current source density.

COMPARISON WITH EXISTING METHOD(S)

The use of MU and LFP activity for evaluation and tracking biological depth provides a more comprehensive characterization of the electrophysiological performance landscape of microelectrodes.

CONCLUSIONS

A more extensive spatial and temporal insight into the chronic electrophysiological performance over time will help uncover the biological and mechanical failure mechanisms of the neural electrodes and direct future research toward the elucidation of design optimization for specific applications.

Ten-dimensional anthropomorphic arm control in a human brain-machine interface: difficulties, solutions, and limitations

OBJECTIVE

In a previous study we demonstrated continuous translation, orientation and one-dimensional grasping control of a prosthetic limb (seven degrees of freedom) by a human subject with tetraplegia using a brain-machine interface (BMI). The current study, in the same subject, immediately followed the previous work and expanded the scope of the control signal by also extracting hand-shape commands from the two 96-channel intracortical electrode arrays implanted in the subject's left motor cortex.

APPROACH

Four new control signals, dictating prosthetic hand shape, replaced the one-dimensional grasping in the previous study, allowing the subject to control the prosthetic limb with ten degrees of freedom (three-dimensional (3D) translation, 3D orientation, four-dimensional hand shaping) simultaneously.

MAIN RESULTS

Robust neural tuning to hand shaping was found, leading to ten-dimensional (10D) performance well above chance levels in all tests. Neural unit preferred directions were broadly distributed through the 10D space, with the majority of units significantly tuned to all ten dimensions, instead of being restricted to isolated domains (e.g. translation, orientation or hand shape). The addition of hand shaping emphasized object-interaction behavior. A fundamental component of BMIs is the calibration used to associate neural activity to intended movement. We found that the presence of an object during calibration enhanced successful shaping of the prosthetic hand as it closed around the object during grasping.

SIGNIFICANCE

Our results show that individual motor cortical neurons encode many parameters of movement, that object interaction is an important factor when extracting these signals, and that high-dimensional operation of prosthetic devices can be achieved with simple decoding algorithms. ClinicalTrials.gov Identifier: NCT01364480.

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.

Motor cortex single-neuron and population contributions to compensation for multiple dynamic force fields

To elucidate how primary motor cortex (M1) neurons contribute to the performance of a broad range of different and even incompatible motor skills, we trained two monkeys to perform single-degree-of-freedom elbow flexion/extension movements that could be perturbed by a variety of externally generated force fields. Fields were presented in a pseudorandom sequence of trial blocks. Different computer monitor background colors signaled the nature of the force field throughout each block. There were five different force fields: null field without perturbing torque, assistive and resistive viscous fields proportional to velocity, a resistive elastic force field proportional to position and a resistive viscoelastic field that was the linear combination of the resistive viscous and elastic force fields. After the monkeys were extensively trained in the five field conditions, neural recordings were subsequently made in M1 contralateral to the trained arm. Many caudal M1 neurons altered their activity systematically across most or all of the force fields in a manner that was appropriate to contribute to the compensation for each of the fields. The net activity of the entire sample population likewise provided a predictive signal about the differences in the time course of the external forces encountered during the movements across all force conditions. The neurons showed a broad range of sensitivities to the different fields, and there was little evidence of a modular structure by which subsets of M1 neurons were preferentially activated during movements in specific fields or combinations of fields.

Cortical magnification plus cortical plasticity equals vision?

Most approaches to visual prostheses have focused on the retina, and for good reasons. The earlier that one introduces signals into the visual system, the more one can take advantage of its prodigious computational abilities. For methods that make use of microelectrodes to introduce electrical signals, however, the limited density and volume occupying nature of the electrodes place severe limits on the image resolution that can be provided to the brain. In this regard, non-retinal areas in general, and the primary visual cortex in particular, possess one large advantage: "magnification factor" (MF)-a value that represents the distance across a sheet of neurons that represents a given angle of the visual field. In the foveal representation of primate primary visual cortex, the MF is enormous-on the order of 15-20 mm/deg in monkeys and humans, whereas on the retina, the MF is limited by the optical design of the eye to around 0.3m m/deg. This means that, for an electrode array of a given density, a much higher-resolution image can be introduced into V1 than onto the retina (or any other visual structure). In addition to this tremendous advantage in resolution, visual cortex is plastic at many different levels ranging from a very local ability to learn to better detect electrical stimulation to higher levels of learning that permit human observers to adapt to radical changes to their visual inputs. We argue that the combination of the large magnification factor and the impressive ability of the cerebral cortex to learn to recognize arbitrary patterns, might outweigh the disadvantages of bypassing earlier processing stages and makes V1 a viable option for the restoration of vision.

Developing a new experimental system for an undergraduate laboratory exercise to teach theories of visuomotor learning

Humans have a flexible motor ability to adapt their movements to changes in the internal/external environment. For example, using arm-reaching tasks, a number of studies experimentally showed that participants adapt to a novel visuomotor environment. These results helped develop computational models of motor learning implemented in the central nervous system. Despite the importance of such experimental paradigms for exploring the mechanisms of motor learning, because of the cost and preparation time, most students are unable to participate in such experiments. Therefore, in the current study, to help students better understand motor learning theories, we developed a simple finger-reaching experimental system using commonly used laptop PC components with an open-source programming language (Processing Motor Learning Toolkit: PMLT). We found that compared to a commercially available robotic arm-reaching device, our PMLT accomplished similar learning goals (difference in the error reduction between the devices, P = 0.10). In addition, consistent with previous reports from visuomotor learning studies, the participants showed after-effects indicating an adaptation of the motor learning system. The results suggest that PMLT can serve as a new experimental system for an undergraduate laboratory exercise of motor learning theories with minimal time and cost for instructors.

Toward more versatile and intuitive cortical brain-machine interfaces

Brain-machine interfaces have great potential for the development of neuroprosthetic applications to assist patients suffering from brain injury or neurodegenerative disease. One type of brain-machine interface is a cortical motor prosthetic, which is used to assist paralyzed subjects. Motor prosthetics to date have typically used the motor cortex as a source of neural signals for controlling external devices. The review will focus on several new topics in the arena of cortical prosthetics. These include using: recordings from cortical areas outside motor cortex; local field potentials as a source of recorded signals; somatosensory feedback for more dexterous control of robotics; and new decoding methods that work in concert to form an ecology of decode algorithms. These new advances promise to greatly accelerate the applicability and ease of operation of motor prosthetics.

Delineating the whole brain BOLD response to passive movement kinematics

The field of brain-machine interfaces (BMIs) has made great advances in recent years, converting thought to movement, with some of the most successful implementations measuring directly from the motor cortex. However, the ability to record from additional regions of the brain could potentially improve flexibility and robustness of use. In addition, BMIs of the future will benefit from integrating kinesthesia into the control loop. Here, we examine whether changes in passively induced forefinger movement amplitude are represented in different regions than forefinger velocity via a MR compatible robotic manipulandum. Using functional magnetic resonance imaging (fMRI), five healthy participants were exposed to combinations of forefinger movement amplitude and velocity in a factorial design followed by an epoch-based analysis. We found that primary and secondary somatosensory regions were activated, as well as cingulate motor area, putamen and cerebellum, with greater activity from changes in velocity compared to changes in amplitude. This represents the first investigation into whole brain response to parametric changes in passive movement kinematics. In addition to informing BMIs, these results have implications towards neural correlates of robotic rehabilitation.

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.

Intentional signal in prefrontal cortex generalizes across different sensory modalities

Biofeedback-EEG training to learn the mental control of an external device (e.g., a cursor on the screen) has been an important paradigm to attempt to understand the involvements of various areas of the brain in the volitional control and the modulation of intentional thought processes. Often the areas to adapt and to monitor progress are selected a priori. Less explored, however, has been the notion of automatically emerging activation in a particular area or subregions within that area recruited above and beyond the rest of the brain. Likewise, the notion of evoking such a signal as an amodal, abstract one remaining robust across different sensory modalities could afford some exploration. Here we develop a simple binary control task in the context of brain-computer interface (BCI) and use a Bayesian sparse probit classification algorithm to automatically uncover brain regional activity that maximizes task performance. We trained and tested 19 participants using the visual modality for instructions and feedback. Across training blocks we quantified coupling of the frontoparietal nodes and selective involvement of visual and auditory regions as a function of the real-time sensory feedback. The testing phase under both forms of sensory feedback revealed automatic recruitment of the prefrontal cortex with a parcellation of higher strength levels in Brodmann's areas 9, 10, and 11 significantly above those in other brain areas. We propose that the prefrontal signal may be a neural correlate of externally driven intended direction and discuss our results in the context of various aspects involved in the cognitive control of our thoughts.

Rapid acquisition of novel interface control by small ensembles of arbitrarily selected primary motor cortex neurons

Pioneering studies demonstrated that novel degrees of freedom could be controlled individually by directly encoding the firing rate of single motor cortex neurons, without regard to each neuron's role in controlling movement of the native limb. In contrast, recent brain-computer interface work has emphasized decoding outputs from large ensembles that include substantially more neurons than the number of degrees of freedom being controlled. To bridge the gap between direct encoding by single neurons and decoding output from large ensembles, we studied monkeys controlling one degree of freedom by comodulating up to four arbitrarily selected motor cortex neurons. Performance typically exceeded random quite early in single sessions and then continued to improve to different degrees in different sessions. We therefore examined factors that might affect performance. Performance improved with larger ensembles. In contrast, other factors that might have reflected preexisting synaptic architecture-such as the similarity of preferred directions-had little if any effect on performance. Patterns of comodulation among ensemble neurons became more consistent across trials as performance improved over single sessions. Compared with the ensemble neurons, other simultaneously recorded neurons showed less modulation. Patterns of voluntarily comodulated firing among small numbers of arbitrarily selected primary motor cortex (M1) neurons thus can be found and improved rapidly, with little constraint based on the normal relationships of the individual neurons to native limb movement. This rapid flexibility in relationships among M1 neurons may in part underlie our ability to learn new movements and improve motor skill.

Restoring sensorimotor function through intracortical interfaces: progress and looming challenges

The loss of a limb or paralysis resulting from spinal cord injury has devastating consequences on quality of life. One approach to restoring lost sensory and motor abilities in amputees and patients with tetraplegia is to supply them with implants that provide a direct interface with the CNS. Such brain-machine interfaces might enable a patient to exert voluntary control over a prosthetic or robotic limb or over the electrically induced contractions of paralysed muscles. A parallel interface could convey sensory information about the consequences of these movements back to the patient. Recent developments in the algorithms that decode motor intention from neuronal activity and in approaches to convey sensory feedback by electrically stimulating neurons, using biomimetic and adaptation-based approaches, have shown the promise of invasive interfaces with sensorimotor cortices, although substantial challenges remain.

Motor cortical control of movement speed with implications for brain-machine interface control

Motor cortex plays a substantial role in driving movement, yet the details underlying this control remain unresolved. We analyzed the extent to which movement-related information could be extracted from single-trial motor cortical activity recorded while monkeys performed center-out reaching. Using information theoretic techniques, we found that single units carry relatively little speed-related information compared with direction-related information. This result is not mitigated at the population level: simultaneously recorded population activity predicted speed with significantly lower accuracy relative to direction predictions. Furthermore, a unit-dropping analysis revealed that speed accuracy would likely remain lower than direction accuracy, even given larger populations. These results suggest that the instantaneous details of single-trial movement speed are difficult to extract using commonly assumed coding schemes. This apparent paucity of speed information takes particular importance in the context of brain-machine interfaces (BMIs), which rely on extracting kinematic information from motor cortex. Previous studies have highlighted subjects' difficulties in holding a BMI cursor stable at targets. These studies, along with our finding of relatively little speed information in motor cortex, inspired a speed-dampening Kalman filter (SDKF) that automatically slows the cursor upon detecting changes in decoded movement direction. Effectively, SDKF enhances speed control by using prevalent directional signals, rather than requiring speed to be directly decoded from neural activity. SDKF improved success rates by a factor of 1.7 relative to a standard Kalman filter in a closed-loop BMI task requiring stable stops at targets. BMI systems enabling stable stops will be more effective and user-friendly when translated into clinical applications.

Self-recalibrating classifiers for intracortical brain-computer interfaces

OBJECTIVE

Intracortical brain-computer interface (BCI) decoders are typically retrained daily to maintain stable performance. Self-recalibrating decoders aim to remove the burden this may present in the clinic by training themselves autonomously during normal use but have only been developed for continuous control. Here we address the problem for discrete decoding (classifiers).

APPROACH

We recorded threshold crossings from 96-electrode arrays implanted in the motor cortex of two rhesus macaques performing center-out reaches in 7 directions over 41 and 36 separate days spanning 48 and 58 days in total for offline analysis.

MAIN RESULTS

We show that for the purposes of developing a self-recalibrating classifier, tuning parameters can be considered as fixed within days and that parameters on the same electrode move up and down together between days. Further, drift is constrained across time, which is reflected in the performance of a standard classifier which does not progressively worsen if it is not retrained daily, though overall performance is reduced by more than 10% compared to a daily retrained classifier. Two novel self-recalibrating classifiers produce a ~15% increase in classification accuracy over that achieved by the non-retrained classifier to nearly recover the performance of the daily retrained classifier.

SIGNIFICANCE

We believe that the development of classifiers that require no daily retraining will accelerate the clinical translation of BCI systems. Future work should test these results in a closed-loop setting.

Using reinforcement learning to provide stable brain-machine interface control despite neural input reorganization

Brain-machine interface (BMI) systems give users direct neural control of robotic, communication, or functional electrical stimulation systems. As BMI systems begin transitioning from laboratory settings into activities of daily living, an important goal is to develop neural decoding algorithms that can be calibrated with a minimal burden on the user, provide stable control for long periods of time, and can be responsive to fluctuations in the decoder’s neural input space (e.g. neurons appearing or being lost amongst electrode recordings). These are significant challenges for static neural decoding algorithms that assume stationary input/output relationships. Here we use an actor-critic reinforcement learning architecture to provide an adaptive BMI controller that can successfully adapt to dramatic neural reorganizations, can maintain its performance over long time periods, and which does not require the user to produce specific kinetic or kinematic activities to calibrate the BMI. Two marmoset monkeys used the Reinforcement Learning BMI (RLBMI) to successfully control a robotic arm during a two-target reaching task. The RLBMI was initialized using random initial conditions, and it quickly learned to control the robot from brain states using only a binary evaluative feedback regarding whether previously chosen robot actions were good or bad. The RLBMI was able to maintain control over the system throughout sessions spanning multiple weeks. Furthermore, the RLBMI was able to quickly adapt and maintain control of the robot despite dramatic perturbations to the neural inputs, including a series of tests in which the neuron input space was deliberately halved or doubled.

Brain-machine interfaces: an overview

Brain-machine interfaces (BMIs) hold promise to treat neurological disabilities by linking intact brain circuitry to assistive devices, such as limb prostheses, wheelchairs, artificial sensors, and computers. BMIs have experienced very rapid development in recent years, facilitated by advances in neural recordings, computer technologies and robots. BMIs are commonly classified into three types: sensory, motor and bidirectional, which subserve motor, sensory and sensorimotor functions, respectively. Additionally, cognitive BMIs have emerged in the domain of higher brain functions. BMIs are also classified as noninvasive or invasive according to the degree of their interference with the biological tissue. Although noninvasive BMIs are safe and easy to implement, their information bandwidth is limited. Invasive BMIs hold promise to improve the bandwidth by utilizing multichannel recordings from ensembles of brain neurons. BMIs have a broad range of clinical goals, as well as the goal to enhance normal brain functions.

Activity in a premotor cortical nucleus of zebra finches is locally organized and exhibits auditory selectivity in neurons but not in glia

Motor functions are often guided by sensory experience, most convincingly illustrated by complex learned behaviors. Key to sensory guidance in motor areas may be the structural and functional organization of sensory inputs and their evoked responses. We study sensory responses in large populations of neurons and neuron-assistive cells in the songbird motor area HVC, an auditory-vocal brain area involved in sensory learning and in adult song production. HVC spike responses to auditory stimulation display remarkable preference for the bird's own song (BOS) compared to other stimuli. Using two-photon calcium imaging in anesthetized zebra finches we measure the spatio-temporal structure of baseline activity and of auditory evoked responses in identified populations of HVC cells. We find strong correlations between calcium signal fluctuations in nearby cells of a given type, both in identified neurons and in astroglia. In identified HVC neurons only, auditory stimulation decorrelates ongoing calcium signals, less for BOS than for other sound stimuli. Overall, calcium transients show strong preference for BOS in identified HVC neurons but not in astroglia, showing diversity in local functional organization among identified neuron and astroglia populations.

Studies in RF power communication, SAR, and temperature elevation in wireless implantable neural interfaces

Implantable neural interfaces are designed to provide a high spatial and temporal precision control signal implementing high degree of freedom real-time prosthetic systems. The development of a Radio Frequency (RF) wireless neural interface has the potential to expand the number of applications as well as extend the robustness and longevity compared to wired neural interfaces. However, it is well known that RF signal is absorbed by the body and can result in tissue heating. In this work, numerical studies with analytical validations are performed to provide an assessment of power, heating and specific absorption rate (SAR) associated with the wireless RF transmitting within the human head. The receiving antenna on the neural interface is designed with different geometries and modeled at a range of implanted depths within the brain in order to estimate the maximum receiving power without violating SAR and tissue temperature elevation safety regulations. Based on the size of the designed antenna, sets of frequencies between 1 GHz to 4 GHz have been investigated. As expected the simulations demonstrate that longer receiving antennas (dipole) and lower working frequencies result in greater power availability prior to violating SAR regulations. For a 15 mm dipole antenna operating at 1.24 GHz on the surface of the brain, 730 uW of power could be harvested at the Federal Communications Commission (FCC) SAR violation limit. At approximately 5 cm inside the head, this same antenna would receive 190 uW of power prior to violating SAR regulations. Finally, the 3-D bio-heat simulation results show that for all evaluated antennas and frequency combinations we reach FCC SAR limits well before 1 °C. It is clear that powering neural interfaces via RF is possible, but ultra-low power circuit designs combined with advanced simulation will be required to develop a functional antenna that meets all system requirements.

Creating new functional circuits for action via brain-machine interfaces

Brain-machine interfaces (BMIs) are an emerging technology with great promise for developing restorative therapies for those with disabilities. BMIs also create novel, well-defined functional circuits for action that are distinct from the natural sensorimotor apparatus. Closed-loop control of BMI systems can also actively engage learning and adaptation. These properties make BMIs uniquely suited to study learning of motor and non-physical, abstract skills. Recent work used motor BMIs to shed light on the neural representations of skill formation and motor adaptation. Emerging work in sensory BMIs, and other novel interface systems, also highlight the promise of using BMI systems to study fundamental questions in learning and sensorimotor control. This paper outlines the interpretation of BMIs as novel closed-loop systems and the benefits of these systems for studying learning. We review BMI learning studies, their relation to motor control, and propose future directions for this nascent field. Understanding learning in BMIs may both elucidate mechanisms of natural motor and abstract skill learning, and aid in developing the next generation of neuroprostheses.

Internal models engaged by brain-computer interface control

Internal models have been proposed to explain the brain's ability to compensate for sensory feedback delays by predicting the sensory consequences of movement commands. Single-neuron studies in the oculomotor and vestibulo-ocular systems have provided evidence of internal models, as have behavioral studies in the skeletomotor system. Here, we present evidence of internal models from simultaneously recorded population activity underlying closed-loop brain-computer interface (BCI) control. We studied cursor-based BCI control by a nonhuman primate implanted with a multi-electrode array in motor cortex. Using a novel BCI task, we measured the visual feedback processing delay to be about 130 milliseconds. By examining the task-based appropriateness of the population activity at different time lags, we found evidence that the subject compensates for the feedback delay by predicting upcoming cursor positions, suggesting the use of an internal forward model. Lastly, we examined the time course of internal model adaptation after altering the mapping between population activity and cursor movements. This study suggests that closed-loop BCI experiments combined with novel statistical analyses can provide insight into the neural substrates of feedback motor control and motor learning.

Design and validation of a real-time spiking-neural-network decoder for brain-machine interfaces

OBJECTIVE

Cortically-controlled motor prostheses aim to restore functions lost to neurological disease and injury. Several proof of concept demonstrations have shown encouraging results, but barriers to clinical translation still remain. In particular, intracortical prostheses must satisfy stringent power dissipation constraints so as not to damage cortex.

APPROACH

One possible solution is to use ultra-low power neuromorphic chips to decode neural signals for these intracortical implants. The first step is to explore in simulation the feasibility of translating decoding algorithms for brain-machine interface (BMI) applications into spiking neural networks (SNNs).

MAIN RESULTS

Here we demonstrate the validity of the approach by implementing an existing Kalman-filter-based decoder in a simulated SNN using the Neural Engineering Framework (NEF), a general method for mapping control algorithms onto SNNs. To measure this system's robustness and generalization, we tested it online in closed-loop BMI experiments with two rhesus monkeys. Across both monkeys, a Kalman filter implemented using a 2000-neuron SNN has comparable performance to that of a Kalman filter implemented using standard floating point techniques.

SIGNIFICANCE

These results demonstrate the tractability of SNN implementations of statistical signal processing algorithms on different monkeys and for several tasks, suggesting that a SNN decoder, implemented on a neuromorphic chip, may be a feasible computational platform for low-power fully-implanted prostheses. The validation of this closed-loop decoder system and the demonstration of its robustness and generalization hold promise for SNN implementations on an ultra-low power neuromorphic chip using the NEF.

An implantable wireless neural interface for recording cortical circuit dynamics in moving primates

OBJECTIVE

Neural interface technology suitable for clinical translation has the potential to significantly impact the lives of amputees, spinal cord injury victims and those living with severe neuromotor disease. Such systems must be chronically safe, durable and effective.

APPROACH

We have designed and implemented a neural interface microsystem, housed in a compact, subcutaneous and hermetically sealed titanium enclosure. The implanted device interfaces the brain with a 510k-approved, 100-element silicon-based microelectrode array via a custom hermetic feedthrough design. Full spectrum neural signals were amplified (0.1 Hz to 7.8 kHz, 200× gain) and multiplexed by a custom application specific integrated circuit, digitized and then packaged for transmission. The neural data (24 Mbps) were transmitted by a wireless data link carried on a frequency-shift-key-modulated signal at 3.2 and 3.8 GHz to a receiver 1 m away by design as a point-to-point communication link for human clinical use. The system was powered by an embedded medical grade rechargeable Li-ion battery for 7 h continuous operation between recharge via an inductive transcutaneous wireless power link at 2 MHz.

MAIN RESULTS

Device verification and early validation were performed in both swine and non-human primate freely-moving animal models and showed that the wireless implant was electrically stable, effective in capturing and delivering broadband neural data, and safe for over one year of testing. In addition, we have used the multichannel data from these mobile animal models to demonstrate the ability to decode neural population dynamics associated with motor activity.

SIGNIFICANCE

We have developed an implanted wireless broadband neural recording device evaluated in non-human primate and swine. The use of this new implantable neural interface technology can provide insight into how to advance human neuroprostheses beyond the present early clinical trials. Further, such tools enable mobile patient use, have the potential for wider diagnosis of neurological conditions and will advance brain research.

Learning an Internal Dynamics Model from Control Demonstration

Much work in optimal control and inverse control has assumed that the controller has perfect knowledge of plant dynamics. However, if the controller is a human or animal subject, the subject's internal dynamics model may differ from the true plant dynamics. Here, we consider the problem of learning the subject's internal model from demonstrations of control and knowledge of task goals. Due to sensory feedback delay, the subject uses an internal model to generate an internal prediction of the current plant state, which may differ from the actual plant state. We develop a probabilistic framework and exact EM algorithm to jointly estimate the internal model, internal state trajectories, and feedback delay. We applied this framework to demonstrations by a nonhuman primate of brain-machine interface (BMI) control. We discovered that the subject's internal model deviated from the true BMI plant dynamics and provided significantly better explanation of the recorded neural control signals than did the true plant dynamics.