Learning to control a brain-machine interface for reaching and grasping by primates

Reaching and grasping in primates depend on the coordination of neural activity in large frontoparietal ensembles. Here we demonstrate that primates can learn to reach and grasp virtual objects by controlling a robot arm through a closed-loop brain–machine interface (BMIc) that uses multiple mathematical models to extract several motor parameters (i.e., hand position, velocity, gripping force, and the EMGs of multiple arm muscles) from the electrical activity of frontoparietal neuronal ensembles. As single neurons typically contribute to the encoding of several motor parameters, we observed that high BMIc accuracy required recording from large neuronal ensembles. Continuous BMIc operation by monkeys led to significant improvements in both model predictions and behavioral performance. Using visual feedback, monkeys succeeded in producing robot reach-and-grasp movements even when their arms did not move. Learning to operate the BMIc was paralleled by functional reorganization in multiple cortical areas, suggesting that the dynamic properties of the BMIc were incorporated into motor and sensory cortical representations.     

A review on directional information in neural signals for brain-machine interfaces

Brain-machine interfaces (BMIs) can be characterized by the technique used to measure brain activity and by the way different brain signals are translated into commands that control an effector. We give an overview of different approaches and focus on a particular BMI approach: the movement of an artificial effector (e.g. arm prosthesis to the right) by those motor cortical signals that control the equivalent movement of a corresponding body part (e.g. arm movement to the right). This approach has been successfully applied in monkeys and humans by accurately extracting parameters of movements from the spiking activity of multiple single-units. Here, we review recent findings showing that analog neuronal population signals, ranging from intracortical local field potentials over epicortical ECoG to non-invasive EEG and MEG, can also be used to decode movement direction and continuous movement trajectories. Therefore, these signals might provide additional or alternative control for this BMI approach, with possible advantages due to reduced invasiveness.     

Activity in the lateral prefrontal cortex reflects multiple steps of future events in action plans

To achieve a behavioral goal in a complex environment, we must plan multiple steps of motor behavior. On planning a series of actions, we anticipate future events that will occur as a result of each action and mentally organize the temporal sequence of events. To investigate the involvement of the lateral prefrontal cortex (PFC) in such multistep planning, we examined neuronal activity in the PFC of monkeys performing a maze task that required the planning of stepwise cursor movements to reach a goal. During the preparatory period, PFC neurons reflected each of all forthcoming cursor movements, rather than arm movements. In contrast, in the primary motor cortex, most neuronal activity reflected arm movements but little of cursor movements during the preparatory period, as well as during movement execution. Our data suggest that the PFC is involved primarily in planning multiple future events that occur as a consequence of behavioral actions.     

Emergence of a Stable Cortical Map for Neuroprosthetic Control

Cortical control of neuroprosthetic devices is known to require neuronal adaptations. It remains unclear whether a stable cortical representation for prosthetic function can be stored and recalled in a manner that mimics our natural recall of motor skills. Especially in light of the mixed evidence for a stationary neuron-behavior relationship in cortical motor areas, understanding this relationship during long-term neuroprosthetic control can elucidate principles of neural plasticity as well as improve prosthetic function. Here, we paired stable recordings from ensembles of primary motor cortex neurons in macaque monkeys with a constant decoder that transforms neural activity to prosthetic movements. Proficient control was closely linked to the emergence of a surprisingly stable pattern of ensemble activity, indicating that the motor cortex can consolidate a neural representation for prosthetic control in the presence of a constant decoder. The importance of such a cortical map was evident in that small perturbations to either the size of the neural ensemble or to the decoder could reversibly disrupt function. Moreover, once a cortical map became consolidated, a second map could be learned and stored. Thus, long-term use of a neuroprosthetic device is associated with the formation of a cortical map for prosthetic function that is stable across time, readily recalled, resistant to interference, and resembles a putative memory engram.     

Top-down control of motor cortex ensembles by dorsomedial prefrontal cortex

Dorsomedial prefrontal cortex is critical for the temporal control of behavior. Dorsomedial prefrontal cortex might alter neuronal activity in areas such as motor cortex to inhibit temporally inappropriate responses. We tested this hypothesis by recording from neuronal ensembles in rodent dorsomedial prefrontal cortex during a delayed-response task. One-third of dorsomedial prefrontal neurons were significantly modulated during the delay period. The activity of many of these neurons was predictive of premature responding. We then reversibly inactivated dorsomedial prefrontal cortex while recording ensemble activity in motor cortex. Inactivation of dorsomedial prefrontal cortex reduced delay-related firing, but not response-related firing, in motor cortex. Finally, we made simultaneous recordings in dorsomedial prefrontal cortex and motor cortex and found strong delay-related temporal correlations between neurons in the two cortical areas. These data suggest that functional interactions between dorsomedial prefrontal cortex and motor cortex might serve as a top-down control signal that inhibits inappropriate responding.     

Recording Large-scale Neuronal Ensembles with Silicon Probes in the Anesthetized Rat

Large scale electrophysiological recordings from neuronal ensembles offer the opportunity to investigate how the brain orchestrates the wide variety of behaviors from the spiking activity of its neurons. One of the most effective methods to monitor spiking activity from a large number of neurons in multiple local neuronal circuits simultaneously is by using silicon electrode arrays^1-3.Action potentials produce large transmembrane voltage changes in the vicinity of cell somata. These output signals can be measured by placing a conductor in close proximity of a neuron. If there are many active (spiking) neurons in the vicinity of the tip, the electrode records combined signal from all of them, where contribution of a single neuron is weighted by its ''electrical distance''. Silicon probes are ideal recording electrodes to monitor multiple neurons because of a large number of recording sites (+64) and a small volume. Furthermore, multiple sites can be arranged over a distance of millimeters, thus allowing for the simultaneous recordings of neuronal activity in the various cortical layers or in multiple cortical columns (Fig. 1). Importantly, the geometrically precise distribution of the recording sites also allows for the determination of the spatial relationship of the isolated single neurons⁴. Here, we describe an acute, large-scale neuronal recording from the left and right forelimb somatosensory cortex simultaneously in an anesthetized rat with silicon probes (Fig. 2).     

Rapid alterations in corticostriatal ensemble coordination during acute dopamine-dependent motor dysfunction

Dopaminergic dysregulation can cause motor dysfunction, but the mechanisms underlying dopamine-related motor disorders remain under debate. We used an inducible and reversible pharmacogenetic approach in dopamine transporter knockout mice to investigate the simultaneous activity of neuronal ensembles in the dorsolateral striatum and primary motor cortex during hyperdopaminergia ( approximately 500% of controls) with hyperkinesia, and after rapid and profound dopamine depletion (<0.2%) with akinesia in the same animal. Surprisingly, although most cortical and striatal neurons ( approximately 70%) changed firing rate during the transition between dopamine-related hyperkinesia and akinesia, the overall cortical firing rate remained unchanged. Conversely, neuronal oscillations and ensemble activity coordination within and between cortex and striatum did change rapidly between these periods. During hyperkinesia, corticostriatal activity became largely asynchronous, while during dopamine-depletion the synchronicity increased. Thus, dopamine-related disorders like Parkinson's disease may not stem from changes in the overall levels of cortical activity, but from dysfunctional activity coordination in corticostriatal circuits.     

Redundant information encoding in primary motor cortex during natural and prosthetic motor control

Redundant encoding of information facilitates reliable distributed information processing. To explore this hypothesis in the motor system, we applied concepts from information theory to quantify the redundancy of movement-related information encoded in the macaque primary motor cortex (M1) during natural and neuroprosthetic control. Two macaque monkeys were trained to perform a delay center-out reaching task controlling a computer cursor under natural arm movement (manual control, ‘MC’), and using a brain-machine interface (BMI) via volitional control of neural ensemble activity (brain control, ‘BC’). During MC, we found neurons in contralateral M1 to contain higher and more redundant information about target direction than ipsilateral M1 neurons, consistent with the laterality of movement control. During BC, we found that the M1 neurons directly incorporated into the BMI (‘direct’ neurons) contained the highest and most redundant target information compared to neurons that were not incorporated into the BMI (‘indirect’ neurons). This effect was even more significant when comparing to M1 neurons of the opposite hemisphere. Interestingly, when we retrained the BMI to use ipsilateral M1 activity, we found that these neurons were more redundant and contained higher information than contralateral M1 neurons, even though ensembles from this hemisphere were previously less redundant during natural arm movement. These results indicate that ensembles most associated to movement contain highest redundancy and information encoding, which suggests a role for redundancy in proficient natural and prosthetic motor control.     

Arm movements in monkeys: behavior and neurophysiology

Reaching to objects of interest is very common in the behavioral repertoire of primates. Monkeys possess keen binocular vision and make graceful and accurate arm movements. This review focuses on behavioral and neurophysiological aspects of eye-hand coordination in behaving monkeys, including neural coding mechanisms at the single cell level and in neuronal populations. The results of these studies have converged to a common behavioral-neurophysiological ground and provided a springboard for studies of brain mechanisms underlying motor cognitive function.     

Choosing goals, not rules: deciding among rule-based action plans

In natural situations, movements are often directed toward locations different from that of the evoking sensory stimulus. Movement goals must then be inferred from the sensory cue based on rules. When there is uncertainty about the rule that applies for a given cue, planning a movement involves both choosing the relevant rule and computing the movement goal based on that rule. Under these conditions, it is not clear whether primates compute multiple movement goals based on all possible rules before choosing an action, or whether they first choose a rule and then only represent the movement goal associated with that rule. Supporting the former hypothesis, we show that neurons in the frontoparietal reach areas of monkeys simultaneously represent two different rule-based movement goals, which are biased by the monkeys' choice preferences. Apparently, primates choose between multiple behavioral options by weighing against each other the movement goals associated with each option.     

Task-Dependent Changes in Cross-Level Coupling between Single Neurons and Oscillatory Activity in Multiscale Networks

Understanding the principles governing the dynamic coordination of functional brain networks remains an important unmet goal within neuroscience. How do distributed ensembles of neurons transiently coordinate their activity across a variety of spatial and temporal scales? While a complete mechanistic account of this process remains elusive, evidence suggests that neuronal oscillations may play a key role in this process, with different rhythms influencing both local computation and long-range communication. To investigate this question, we recorded multiple single unit and local field potential (LFP) activity from microelectrode arrays implanted bilaterally in macaque motor areas. Monkeys performed a delayed center-out reach task either manually using their natural arm (Manual Control, MC) or under direct neural control through a brain-machine interface (Brain Control, BC). In accord with prior work, we found that the spiking activity of individual neurons is coupled to multiple aspects of the ongoing motor beta rhythm (10–45 Hz) during both MC and BC, with neurons exhibiting a diversity of coupling preferences. However, here we show that for identified single neurons, this beta-to-rate mapping can change in a reversible and task-dependent way. For example, as beta power increases, a given neuron may increase spiking during MC but decrease spiking during BC, or exhibit a reversible shift in the preferred phase of firing. The within-task stability of coupling, combined with the reversible cross-task changes in coupling, suggest that task-dependent changes in the beta-to-rate mapping play a role in the transient functional reorganization of neural ensembles. We characterize the range of task-dependent changes in the mapping from beta amplitude, phase, and inter-hemispheric phase differences to the spike rates of an ensemble of simultaneously-recorded neurons, and discuss the potential implications that dynamic remapping from oscillatory activity to spike rate and timing may hold for models of computation and communication in distributed functional brain networks.     

The impact of brain imaging technology on our understanding of motor function and dysfunction

Brain imaging techniques have demonstrated functional specialisation of multiple areas within the motor system. They have also defined the patterns of interactions between these regions during normal motor function and in motor disorders. Functional imaging makes visible the changes in cortical activity that take place over time during motor functions, from the activations a fraction of a second before voluntary action to cortical neuronal plasticity several weeks after injury. Recently, the functional abnormalities underlying various acquired and developmental motor disorders have been described, as well as the effects of therapeutic intervention.     

Neuronal activity in the primate supplementary motor area and the primary motor cortex in relation to spatio-temporal bimanual coordination

Single neuronal activity was recorded from the supplementary motor area (SMA-proper and pre-SMA) and primary motor cortex (M1) in two Macaca fascicularis trained to perform a delayed conditional sequence of coordinated bimanual pull and grasp movements. The behavioural paradigm was designed to distinguish neuronal activity associated with bimanual coordination from that related to a comparable motor sequence but executed unimanually (left or right arm only). The bimanual and unimanual trials were instructed in a random order by a visual cue. Following the cue, there was a waiting period until presentation of a "go-signal", signalling the monkey to perform the instructed movement. A total of 143 task-related neurons were recorded from the SMA (SMA-proper, 62; pre-SMA, 81). Most SMA units (87%) were active in both unimanual contralateral and unimanual ipsilateral trials (bilateral neurons), whereas 9% of units were active only in unimanual contralateral trials and 3% were active only in unimanual ipsilateral trials. Forty-eight per cent of SMA task-related units were classified as bimanual, defined as neurons in which the activity observed in bimanual trials could not be predicted from that associated with unimanual trials when comparing the same events related to the same arm. For direct comparison, 527 neurons were recorded from M1 in the same monkeys performing the same tasks. The comparison showed that M1 contains significantly less bilateral neurons (75%) than the SMA, whereas the reverse was observed for contralateral neurons (22% in M1). The proportion of M1 bimanual cells (53%) was not statistically different from that observed in the SMA. The results suggest that both the SMA and M1 may contribute to the control of sequential bimanual coordinated movements. Interlimb coordination may then take place in a distributed network including at least the SMA and M1, but the contribution of other cortical and subcortical areas such as cingulate motor cortex and basal ganglia remains to be investigated.     

Spatial coding of visually guided arm movements in primate motor cortex

Previous studies of the motor cortex in behaving animals were focused on the relations between the activity of single cells, usually pyramidal tract neurons, and parameters of isometric contraction (e.g., intensity of force) or parameters of movement along one axis (e.g., flexion-extension) of a single joint (e.g., elbow or wrist). However, the commonly meaningful behavioral parameter is the trajectory of the hand in extrapersonal space, which is realized by simultaneous motions about two or three joints (e.g., elbow, shoulder, wrist) and concurrent engagement of several muscles. The spatial parameters of a straight trajectory are its direction and extent. We hypothesized that a major function of the motor cortex, among other possible roles, is the specification and control of the direction of the movement trajectory in space. This reference of motor cortical function to the control of spatial aspects of the trajectory differentiated our approach from the other approaches outlined above. We investigated the directional selectivity cells in the arm area of the motor cortex by recording their activity while monkeys moved their hands in various directions in space towards visual targets. There were two salient findings of these studies. First, the intensity of the discharge of single cells varies in an orderly fashion with the direction of movement in space, so that the discharge rate is highest with movements in a preferred direction, and decreases progressively with movements made in directions more and more away from the preferred one. Thus single cells are broadly tuned around a preferred direction which differs among different cells.(ABSTRACT TRUNCATED AT 250 WORDS)     

Functional implications of temporal structure in primate cortical information processing

Processing of information in the cerebral cortex of primates is characterized by distributed representations and processing in neuronal assemblies rather than by detector neurons, cardinal cells or command neurons. Responses of individual neurons in sensory cortical areas contain limited and ambiguous information on common features of the natural environment which is disambiguated by comparison with the responses of other, related neurons. Distributed representations are also capable to represent the enormous complexity and variability of the natural environment by the large number of possible combinations of neurons that can engage in the representation of a stimulus or other content. A critical problem of distributed representation and processing is the superposition of several assemblies activated at the same time since interpretation and processing of a population code requires that the responses related to a single representation can be identified and distinguished from other, related activity. A possible mechanism which tags related responses is the synchronization of neuronal responses of the same assembly with a precision in the millisecond range. This mechanism also supports the separate processing of distributed activity and dynamic assembly formation. Experimental evidence from electrophysiological investigations of non-human primates and human subjects shows that synchronous activity can be found in visual, auditory and motor areas of the cortex. Simultaneous recordings of neurons in the visual cortex indicate that individual neurons synchronize their activity with each other, if they respond to the same stimulus but not if they are part of different assemblies representing different contents. Furthermore, evidence for synchronous activity related to perception, expectation, memory, and attention has been observed.     

Water metabolic parameter changes in rhesus monkeys during exposure to prolonged restriction of motor activity

The objective of this investigation was to determine the effect of prolonged restriction of motor activity (hypokinesia [HK]) on several parameters of water metabolism in primates. The studies were performed on 12 rhesus monkeys aged 4–5 yr (5.10–6.85 kg) during the hypokinetic period of 90 d and during the prehypokinetic period of 30 d. They were divided into two equal groups: the first group was placed under ordinary vivarium conditions (vivarium control animals) and the second group was subjected to 90 d of HK (hypokinetic animals). For the simulation of the hypokinetic effect, the primates were immobilized on their abdomens in special tables. The legs of the monkeys were immobilized with hip and knee joints extended. The primates retained freedom of movement at elbow, wrist, and ankle. During the preexperimental period of 30 d and during the experimental period of 90 d, the following variables were determined: body weight, total body fluid content, specific total body fluid, mean fluid consumed and eliminated in urine, specific plasma resistance, hematocrit level, and plasma concentrations of sodium (Na) and potassium (K). In the hypokinetic primates, body weight decreased significantly when compared to the controls. Mean fluid intake, total body fluid, and specific total body fluid decreased, whereas mean daily fluid loss and specific mean daily fluid elimination increased significantly. Specific plasma resistance, hematocrit level, and plasma electrolyte concentrations increased significantly when compared to the control primates. It was concluded that prolonged restriction of motor activity induces significant changes in water metabolic parameters of primates leading in decreased total water content of the body.     

Complex movements evoked by microstimulation of precentral cortex

Electrical microstimulation was used to study primary motor and premotor cortex in monkeys. Each stimulation train was 500 ms in duration, approximating the time scale of normal reaching and grasping movements and the time scale of the neuronal activity that normally accompanies movement. This stimulation on a behaviorally relevant time scale evoked coordinated, complex postures that involved many joints. For example, stimulation of one site caused the mouth to open and also caused the hand to shape into a grip posture and move to the mouth. Stimulation of this site always drove the joints toward this final posture, regardless of the direction of movement required to reach the posture. Stimulation of other cortical sites evoked different postures. Postures that involved the arm were arranged across cortex to form a map of hand positions around the body. This stimulation-evoked map encompassed both primary motor and the adjacent premotor cortex. We suggest that these regions fit together into a single map of the workspace around the body.     

Differential corticostriatal plasticity during fast and slow motor skill learning in mice

BACKGROUND: Motor skill learning usually comprises "fast" improvement in performance within the initial training session and "slow" improvement that develops across sessions. Previous studies have revealed changes in activity and connectivity in motor cortex and striatum during motor skill learning. However, the nature and dynamics of the plastic changes in each of these brain structures during the different phases of motor learning remain unclear. RESULTS: By using multielectrode arrays, we recorded the simultaneous activity of neuronal ensembles in motor cortex and dorsal striatum of mice during the different phases of skill learning on an accelerating rotarod. Mice exhibited fast improvement in the task during the initial session and also slow improvement across days. Throughout training, a high percentage of striatal (57%) and motor cortex (55%) neurons were task related; i.e., changed their firing rate while mice were running on the rotarod. Improvement in performance was accompanied by substantial plastic changes in both striatum and motor cortex. We observed parallel recruitment of task-related neurons in both structures specifically during the first session. Conversely, during slow learning across sessions we observed differential refinement of the firing patterns in each structure. At the neuronal ensemble level, we observed considerable changes in activity within the first session that became less evident during subsequent sessions. CONCLUSIONS: These data indicate that cortical and striatal circuits exhibit remarkable but dissociable plasticity during fast and slow motor skill learning and suggest that distinct neural processes mediate the different phases of motor skill learning.     

Bond Type and Discretization of Nonmuscle Myosin II Are Critical for Simulated Contractile Dynamics

Molecular motors drive cytoskeletal rearrangements to change cell shape. Myosins are the motors that move, cross-link, and modify the actin cytoskeleton. The primary force generator in contractile actomyosin networks is nonmuscle myosin II (NMMII), a molecular motor that assembles into ensembles that bind, slide, and cross-link actin filaments (F-actin). The multivalence of NMMII ensembles and their multiple roles have confounded the resolution of crucial questions, including how the number of NMMII subunits affects dynamics and what affects the relative contribution of ensembles’ cross-linking versus motoring activities. Because biophysical measurements of ensembles are sparse, modeling of actomyosin networks has aided in discovering the complex behaviors of NMMII ensembles. Myosin ensembles have been modeled via several strategies with variable discretization or coarse graining and unbinding dynamics, and although general assumptions that simplify motor ensembles result in global contractile behaviors, it remains unclear which strategies most accurately depict cellular activity. Here, we used an agent-based platform, Cytosim, to implement several models of NMMII ensembles. Comparing the effects of bond type, we found that ensembles of catch-slip and catch motors were the best force generators and binders of filaments. Slip motor ensembles were capable of generating force but unbound frequently, resulting in slower contractile rates of contractile networks. Coarse graining of these ensemble types from two sets of 16 motors on opposite ends of a stiff rod to two binders, each representing 16 motors, reduced force generation, contractility, and the total connectivity of filament networks for all ensemble types. A parallel cluster model, previously used to describe ensemble dynamics via statistical mechanics, allowed better contractility with coarse graining, though connectivity was still markedly reduced for this ensemble type with coarse graining. Together, our results reveal substantial tradeoffs associated with the process of coarse graining NMMII ensembles and highlight the robustness of discretized catch-slip ensembles in modeling actomyosin networks.     

New insights into motor cortex

An exciting new experiment on the motor cortex of monkeys, by Shenoy and colleagues, begins to elucidate how the neuronal ensemble travels in a systematic fashion through state space. This trajectory through state space may help to explain how the motor cortex sets up and then triggers arm movements.