Extrinsic and intrinsic dynamics in movement intermittency

Multimodal subspace identification for modeling discrete-continuous spiking and field potential population activity

Objective.Learning dynamical latent state models for multimodal spiking and field potential activity can reveal their collective low-dimensional dynamics and enable better decoding of behavior through multimodal fusion. Toward this goal, developing unsupervised learning methods that are computationally efficient is important, especially for real-time learning applications such as brain-machine interfaces (BMIs). However, efficient learning remains elusive for multimodal spike-field data due to their heterogeneous discrete-continuous distributions and different timescales.Approach.Here, we develop a multiscale subspace identification (multiscale SID) algorithm that enables computationally efficient learning for modeling and dimensionality reduction for multimodal discrete-continuous spike-field data. We describe the spike-field activity as combined Poisson and Gaussian observations, for which we derive a new analytical SID method. Importantly, we also introduce a novel constrained optimization approach to learn valid noise statistics, which is critical for multimodal statistical inference of the latent state, neural activity, and behavior. We validate the method using numerical simulations and with spiking and local field potential population activity recorded during a naturalistic reach and grasp behavior.Main results.We find that multiscale SID accurately learned dynamical models of spike-field signals and extracted low-dimensional dynamics from these multimodal signals. Further, it fused multimodal information, thus better identifying the dynamical modes and predicting behavior compared to using a single modality. Finally, compared to existing multiscale expectation-maximization learning for Poisson-Gaussian observations, multiscale SID had a much lower training time while being better in identifying the dynamical modes and having a better or similar accuracy in predicting neural activity and behavior.Significance.Overall, multiscale SID is an accurate learning method that is particularly beneficial when efficient learning is of interest, such as for online adaptive BMIs to track non-stationary dynamics or for reducing offline training time in neuroscience investigations.

Submovement interpersonal coupling is associated to audio-motor coordination performance

Acting in concert with others, a key aspect of our social life, requires behavioral coordination between persons on multiple timescales. When zooming in on the kinematic properties of movements, it appears that small speed fluctuations, called submovements, are embedded within otherwise smooth end-point trajectories. Submovements, by occurring at a faster timescale than that of movements, offer a novel window upon the functional relationship between distinct motor timescales. In this regard, it has previously been shown that when partners visually synchronize their movements, they also coordinate the timing of their submovement by following an alternated pattern. However, it remains unclear whether the mechanisms behind submovement coordination are domain-general or specific to the visual modality, and whether they have relevance for interpersonal coordination also at the scale of whole movements. In a series of solo and dyadic tasks, we show that submovements are also present and coordinated across partners when sensorimotor interactions are mediated by auditory feedback only. Importantly, the accuracy of task-instructed interpersonal coordination at the movement level correlates with the strength of submovement coordination. These results demonstrate that submovement coordination is a potentially fundamental mechanism that participates in interpersonal motor coordination regardless of the sensory domain mediating the interaction.

Modeling and dissociation of intrinsic and input-driven neural population dynamics underlying behavior

Neural dynamics can reflect intrinsic dynamics or dynamic inputs, such as sensory inputs or inputs from other brain regions. To avoid misinterpreting temporally structured inputs as intrinsic dynamics, dynamical models of neural activity should account for measured inputs. However, incorporating measured inputs remains elusive in joint dynamical modeling of neural-behavioral data, which is important for studying neural computations of behavior. We first show how training dynamical models of neural activity while considering behavior but not input or input but not behavior may lead to misinterpretations. We then develop an analytical learning method for linear dynamical models that simultaneously accounts for neural activity, behavior, and measured inputs. The method provides the capability to prioritize the learning of intrinsic behaviorally relevant neural dynamics and dissociate them from both other intrinsic dynamics and measured input dynamics. In data from a simulated brain with fixed intrinsic dynamics that performs different tasks, the method correctly finds the same intrinsic dynamics regardless of the task while other methods can be influenced by the task. In neural datasets from three subjects performing two different motor tasks with task instruction sensory inputs, the method reveals low-dimensional intrinsic neural dynamics that are missed by other methods and are more predictive of behavior and/or neural activity. The method also uniquely finds that the intrinsic behaviorally relevant neural dynamics are largely similar across the different subjects and tasks, whereas the overall neural dynamics are not. These input-driven dynamical models of neural-behavioral data can uncover intrinsic dynamics that may otherwise be missed.

Scale-invariant changes in corticospinal excitability reflect multiplexed oscillations in the motor output

In the absence of disease, humans produce smooth and accurate movement trajectories. Despite such 'macroscopic' aspect, the 'microscopic' structure of movements reveals recurrent (quasi-rhythmic) discontinuities. To date, it is unclear how the sensorimotor system contributes to the macroscopic and microscopic architecture of movement. Here, we investigated how corticospinal excitability changes in relation to microscopic fluctuations that are naturally embedded within larger macroscopic variations in motor output. Participants performed a visuomotor tracking task. In addition to the 0.25 Hz modulation that is required for task fulfilment (macroscopic scale), the motor output shows tiny but systematic fluctuations at ∼2 and 8 Hz (microscopic scales). We show that motor-evoked potentials (MEPs) elicited by transcranial magnetic stimulation (TMS) during task performance are consistently modulated at all (time) scales. Surprisingly, MEP modulation covers a similar range at both micro- and macroscopic scales, even though the motor output differs by several orders of magnitude. Thus, corticospinal excitability finely maps the multiscale temporal patterning of the motor output, but it does so according to a principle of scale invariance. These results suggest that corticospinal excitability indexes a relatively abstract level of movement encoding that may reflect the hierarchical organisation of sensorimotor processes. KEY POINTS: Motor behaviour is organised on multiple (time)scales. Small but systematic ('microscopic') fluctuations are engrained in larger and slower ('macroscopic') variations in motor output, which are instrumental in deploying the desired motor plan. Corticospinal excitability is modulated in relation to motor fluctuations on both macroscopic and microscopic (time)scales. Corticospinal excitability obeys a principle of scale invariance, that is, it is modulated similarly at all (time)scales, possibly reflecting hierarchical mechanisms that optimise motor encoding.

Dynamical flexible inference of nonlinear latent factors and structures in neural population activity

Modelling the spatiotemporal dynamics in the activity of neural populations while also enabling their flexible inference is hindered by the complexity and noisiness of neural observations. Here we show that the lower-dimensional nonlinear latent factors and latent structures can be computationally modelled in a manner that allows for flexible inference causally, non-causally and in the presence of missing neural observations. To enable flexible inference, we developed a neural network that separates the model into jointly trained manifold and dynamic latent factors such that nonlinearity is captured through the manifold factors and the dynamics can be modelled in tractable linear form on this nonlinear manifold. We show that the model, which we named 'DFINE' (for 'dynamical flexible inference for nonlinear embeddings') achieves flexible inference in simulations of nonlinear dynamics and across neural datasets representing a diversity of brain regions and behaviours. Compared with earlier neural-network models, DFINE enables flexible inference, better predicts neural activity and behaviour, and better captures the latent neural manifold structure. DFINE may advance the development of neurotechnology and investigations in neuroscience.

The microstructure of intra- and interpersonal coordination

Movements are naturally composed of submovements, i.e. recurrent speed pulses (2-3 Hz), possibly reflecting intermittent feedback-based motor adjustments. In visuomotor (unimanual) synchronization tasks, partners alternate submovements over time, indicating mutual coregulation. However, it is unclear whether submovement coordination is organized differently between and within individuals. Indeed, different types of information may be variably exploited for intrapersonal and interpersonal coordination. Participants performed a series of bimanual tasks alone or in pairs, with or without visual feedback (solo task only). We analysed the relative timing of submovements between their own hands or between their own hands and those of their partner. Distinct coordinative structures emerged at the submovement level depending on the relevance of visual feedback. Specifically, the relative timing of submovements (between partners/effectors) shifts from alternation to simultaneity and a mixture of both when coordination is achieved using vision (interpersonal), proprioception/efference-copy only (intrapersonal, without vision) or all information sources (intrapersonal, with vision), respectively. These results suggest that submovement coordination represents a behavioural proxy for the adaptive weighting of different sources of information within action-perception loops. In sum, the microstructure of movement reveals common principles governing the dynamics of sensorimotor control to achieve both intra- and interpersonal coordination.

Multimodal subspace identification for modeling discrete-continuous spiking and field potential population activity

Learning dynamical latent state models for multimodal spiking and field potential activity can reveal their collective low-dimensional dynamics and enable better decoding of behavior through multimodal fusion. Toward this goal, developing unsupervised learning methods that are computationally efficient is important, especially for real-time learning applications such as brain-machine interfaces (BMIs). However, efficient learning remains elusive for multimodal spike-field data due to their heterogeneous discrete-continuous distributions and different timescales. Here, we develop a multiscale subspace identification (multiscale SID) algorithm that enables computationally efficient modeling and dimensionality reduction for multimodal discrete-continuous spike-field data. We describe the spike-field activity as combined Poisson and Gaussian observations, for which we derive a new analytical subspace identification method. Importantly, we also introduce a novel constrained optimization approach to learn valid noise statistics, which is critical for multimodal statistical inference of the latent state, neural activity, and behavior. We validate the method using numerical simulations and spike-LFP population activity recorded during a naturalistic reach and grasp behavior. We find that multiscale SID accurately learned dynamical models of spike-field signals and extracted low-dimensional dynamics from these multimodal signals. Further, it fused multimodal information, thus better identifying the dynamical modes and predicting behavior compared to using a single modality. Finally, compared to existing multiscale expectation-maximization learning for Poisson-Gaussian observations, multiscale SID had a much lower computational cost while being better in identifying the dynamical modes and having a better or similar accuracy in predicting neural activity. Overall, multiscale SID is an accurate learning method that is particularly beneficial when efficient learning is of interest.

Dynamical flexible inference of nonlinear latent structures in neural population activity

Inferring complex spatiotemporal dynamics in neural population activity is critical for investigating neural mechanisms and developing neurotechnology. These activity patterns are noisy observations of lower-dimensional latent factors and their nonlinear dynamical structure. A major unaddressed challenge is to model this nonlinear structure, but in a manner that allows for flexible inference, whether causally, non-causally, or in the presence of missing neural observations. We address this challenge by developing DFINE, a new neural network that separates the model into dynamic and manifold latent factors, such that the dynamics can be modeled in tractable form. We show that DFINE achieves flexible nonlinear inference across diverse behaviors and brain regions. Further, despite enabling flexible inference unlike prior neural network models of population activity, DFINE also better predicts the behavior and neural activity, and better captures the latent neural manifold structure. DFINE can both enhance future neurotechnology and facilitate investigations across diverse domains of neuroscience.

Modeling and dissociation of intrinsic and input-driven neural population dynamics underlying behavior

Neural dynamics can reflect intrinsic dynamics or dynamic inputs, such as sensory inputs or inputs from other regions. To avoid misinterpreting temporally-structured inputs as intrinsic dynamics, dynamical models of neural activity should account for measured inputs. However, incorporating measured inputs remains elusive in joint dynamical modeling of neural-behavioral data, which is important for studying neural computations of a specific behavior. We first show how training dynamical models of neural activity while considering behavior but not input, or input but not behavior may lead to misinterpretations. We then develop a novel analytical learning method that simultaneously accounts for neural activity, behavior, and measured inputs. The method provides the new capability to prioritize the learning of intrinsic behaviorally relevant neural dynamics and dissociate them from both other intrinsic dynamics and measured input dynamics. In data from a simulated brain with fixed intrinsic dynamics that performs different tasks, the method correctly finds the same intrinsic dynamics regardless of task while other methods can be influenced by the change in task. In neural datasets from three subjects performing two different motor tasks with task instruction sensory inputs, the method reveals low-dimensional intrinsic neural dynamics that are missed by other methods and are more predictive of behavior and/or neural activity. The method also uniquely finds that the intrinsic behaviorally relevant neural dynamics are largely similar across the three subjects and two tasks whereas the overall neural dynamics are not. These input-driven dynamical models of neural-behavioral data can uncover intrinsic dynamics that may otherwise be missed.

Motor invariants in action execution and perception

The nervous system is sensitive to statistical regularities of the external world and forms internal models of these regularities to predict environmental dynamics. Given the inherently social nature of human behavior, being capable of building reliable predictive models of others' actions may be essential for successful interaction. While social prediction might seem to be a daunting task, the study of human motor control has accumulated ample evidence that our movements follow a series of kinematic invariants, which can be used by observers to reduce their uncertainty during social exchanges. Here, we provide an overview of the most salient regularities that shape biological motion, examine the role of these invariants in recognizing others' actions, and speculate that anchoring socially-relevant perceptual decisions to such kinematic invariants provides a key computational advantage for inferring conspecifics' goals and intentions.

Conserved Population Dynamics in the Cerebro-Cerebellar System between Waking and Sleep

Despite the importance of the cerebellum for motor learning, and the recognized role of sleep in motor memory consolidation, surprisingly little is known about neural activity in the sleeping cerebro-cerebellar system. Here, we used wireless recording from primary motor cortex (M1) and the cerebellum in three female monkeys to examine the relationship between patterns of single-unit spiking activity observed during waking behavior and in natural sleep. Across the population of recorded units, we observed similarities in the timing of firing relative to local field potential features associated with both movements during waking and up state during sleep. We also observed a consistent pattern of asymmetry in pairwise cross-correlograms, indicative of preserved sequential firing in both wake and sleep at low frequencies. Despite the overall similarity in population dynamics between wake and sleep, there was a global change in the timing of cerebellar activity relative to motor cortex, from contemporaneous in the awake state to motor cortex preceding the cerebellum in sleep. We speculate that similar population dynamics in waking and sleep may imply that cerebellar internal models are activated in both states, despite the absence of movement when asleep. Moreover, spindle frequency coherence between the cerebellum and motor cortex may provide a mechanism for cerebellar computations to influence sleep-dependent learning processes in the motor cortex.SIGNIFICANCE STATEMENT It is well known that sleep can lead to improved motor performance. One possibility is that off-line learning results from neural activity during sleep in brain areas responsible for the control of movement. In this study we show for the first time that neuronal patterns in the cerebro-cerebellar system are conserved during both movements and sleep up-states, albeit with a shift in the relative timing between areas. Additionally, we show the presence of simultaneous M1-cerebellar spike coherence at spindle frequencies associated with up-state replay and postulate that this is a mechanism whereby a cerebellar internal model can shape plasticity in neocortical circuits during sleep.

Response Preparation of a Secondary Reaction Time Task is Influenced by Movement Phase within a Continuous Visuomotor Tracking Task

The simultaneous performance of two motor tasks is challenging. Currently, it is unclear how response preparation of a secondary task is impacted by the performance of a continuous primary task. The purpose of the present experiment was to investigate whether the position of the limb performing the primary cyclical tracking task impacts response preparation of a secondary reaction time task. Participants (n=20) performed a continuous tracking task with their left hand that involved cyclical and targeted wrist flexion and extension. Occasionally, a probe reaction time task requiring isometric wrist extension was performed with the right hand in response to an auditory stimulus (80 dB or 120 dB) that was triggered when the left hand passed through one of ten locations identified within the movement cycle. On separate trials, transcranial magnetic stimulation was applied over the left primary motor cortex and triggered at the same 10 stimulus locations to assess corticospinal excitability associated with the probe reaction time task. Results revealed that probe reaction times were significantly longer and motor evoked potential amplitudes were significantly larger when the left hand was in the middle of a movement cycle compared to an endpoint, suggesting that response preparation of a secondary probe reaction time task was modulated by the phase of movement within the continuous primary task. These results indicate that primary motor task requirements can impact preparation of a secondary task, reinforcing the importance of considering primary task characteristics in dual-task experimental design.

Interpersonal synchronization of movement intermittency

Most animal species group together and coordinate their behavior in quite sophisticated manners for mating, hunting, or defense purposes. In humans, coordination at a macroscopic level (the pacing of movements) is evident both in daily life (e.g., walking) and skilled (e.g., music and dance) behaviors. By examining the fine structure of movement, we here show that interpersonal coordination is established also at a microscopic – submovement – level. Natural movements appear as marked by recurrent (2–3 Hz) speed breaks, i.e., submovements, that are traditionally considered the result of intermittency in (visuo)motor feedback-based control. In a series of interpersonal coordination tasks, we show that submovements produced by interacting partners are not independent but alternate tightly over time, reflecting online mutual adaptation. These findings unveil a potential core mechanism for behavioral coordination that is based on between-persons synchronization of the intrinsic dynamics of action-perception cycles.

•

Movements show intermittent speed pulses occurring at 2–3 Hz, called submovements

•

Submovements are actively coordinated in counter-phase by interacting partners

•

Submovements coordination depends on spatial alignment but not movement congruency

•

Behavioral coordination occurs both at macro- and microscopic movement scales

Cyclic, Condition-Independent Activity in Primary Motor Cortex Predicts Corrective Movement Behavior

Reaching movements are known to have large condition-independent (CI) neural activity and cyclic neural dynamics. A new precision center-out task was performed by rhesus macaques to test the hypothesis that cyclic, CI neural activity in the primary motor cortex (M1) occurs not only during initial reaching movements but also during subsequent corrective movements. Corrective movements were observed to be discrete with time courses and bell-shaped speed profiles similar to the initial movements. CI cyclic neural trajectories were similar and repeated for initial and each additional corrective submovement. The phase of the cyclic CI neural activity predicted the time of peak movement speed more accurately than regression of instantaneous firing rate, even when the subject made multiple corrective movements. Rather than being controlled as continuations of the initial reach, a discrete cycle of motor cortex activity encodes each corrective submovement.

Modeling multiscale causal interactions between spiking and field potential signals during behavior

Objective. Brain recordings exhibit dynamics at multiple spatiotemporal scales, which are measured with spike trains and larger-scale field potential signals. To study neural processes, it is important to identify and model causal interactions not only at a single scale of activity, but also across multiple scales, i.e. between spike trains and field potential signals. Standard causality measures are not directly applicable here because spike trains are binary-valued but field potentials are continuous-valued. It is thus important to develop computational tools to recover multiscale neural causality during behavior, assess their performance on neural datasets, and study whether modeling multiscale causalities can improve the prediction of neural signals beyond what is possible with single-scale causality. Approach. We design a multiscale model-based Granger-like causality method based on directed information and evaluate its success both in realistic biophysical spike-field simulations and in motor cortical datasets from two non-human primates (NHP) performing a motor behavior. To compute multiscale causality, we learn point-process generalized linear models that predict the spike events at a given time based on the history of both spike trains and field potential signals. We also learn linear Gaussian models that predict the field potential signals at a given time based on their own history as well as either the history of binary spike events or that of latent firing rates. Main results. We find that our method reveals the true multiscale causality network structure in biophysical simulations despite the presence of model mismatch. Further, models with the identified multiscale causalities in the NHP neural datasets lead to better prediction of both spike trains and field potential signals compared to just modeling single-scale causalities. Finally, we find that latent firing rates are better predictors of field potential signals compared with the binary spike events in the NHP datasets. Significance. This multiscale causality method can reveal the directed functional interactions across spatiotemporal scales of brain activity to inform basic science investigations and neurotechnologies.

Sensory feedback can give rise to neural rotations

Investigating how an artificial network of neurons controls a simulated arm suggests that rotational patterns of activity in the motor cortex may rely on sensory feedback from the moving limb.

Rotational dynamics in motor cortex are consistent with a feedback controller

Recent studies have identified rotational dynamics in motor cortex (MC), which many assume arise from intrinsic connections in MC. However, behavioral and neurophysiological studies suggest that MC behaves like a feedback controller where continuous sensory feedback and interactions with other brain areas contribute substantially to MC processing. We investigated these apparently conflicting theories by building recurrent neural networks that controlled a model arm and received sensory feedback from the limb. Networks were trained to counteract perturbations to the limb and to reach toward spatial targets. Network activities and sensory feedback signals to the network exhibited rotational structure even when the recurrent connections were removed. Furthermore, neural recordings in monkeys performing similar tasks also exhibited rotational structure not only in MC but also in somatosensory cortex. Our results argue that rotational structure may also reflect dynamics throughout the voluntary motor system involved in online control of motor actions.

De novo learning versus adaptation of continuous control in a manual tracking task

How do people learn to perform tasks that require continuous adjustments of motor output, like riding a bicycle? People rely heavily on cognitive strategies when learning discrete movement tasks, but such time-consuming strategies are infeasible in continuous control tasks that demand rapid responses to ongoing sensory feedback. To understand how people can learn to perform such tasks without the benefit of cognitive strategies, we imposed a rotation/mirror reversal of visual feedback while participants performed a continuous tracking task. We analyzed behavior using a system identification approach, which revealed two qualitatively different components of learning: adaptation of a baseline controller and formation of a new, task-specific continuous controller. These components exhibited different signatures in the frequency domain and were differentially engaged under the rotation/mirror reversal. Our results demonstrate that people can rapidly build a new continuous controller de novo and can simultaneously deploy this process with adaptation of an existing controller.

Spatial and Temporal Arrangement of Recurrent Inhibition in the Primate Upper Limb

Renshaw cells mediate recurrent inhibition between motoneurons within the spinal cord. The function of this circuit is not clear; we previously suggested based on computational modeling that it may cancel oscillations in muscle activity around 10 Hz, thereby reducing physiological tremor. Such tremor is especially problematic for dexterous hand movements, yet knowledge of recurrent inhibitory function is sparse for the control of the primate upper limb, where no direct measurements have been made to date. In this study, we made intracellular penetrations into 89 motoneurons in the cervical enlargement of four terminally anesthetized female macaque monkeys, and recorded recurrent IPSPs in response to antidromic stimulation of motor axons. Recurrent inhibition was strongest to motoneurons innervating shoulder muscles and elbow extensors, weak to wrist and digit extensors, and almost absent to the intrinsic muscles of the hand. Recurrent inhibitory connections often spanned joints, for example from motoneurons innervating wrist and digit muscles to those controlling the shoulder and elbow. Wrist and digit flexor motoneurons sometimes inhibited the corresponding extensors, and vice versa. This complex connectivity presumably reflects the flexible usage of the primate upper limb. Using trains of stimuli to motor nerves timed as a Poisson process and coherence analysis, we also examined the temporal properties of recurrent inhibition. The recurrent feedback loop effectively carried frequencies up to 100 Hz, with a coherence peak around 20 Hz. The coherence phase validated predictions from our previous computational model, supporting the idea that recurrent inhibition may function to reduce tremor.

SIGNIFICANCE STATEMENT We present the first direct measurements of recurrent inhibition in primate upper limb motoneurons, revealing that it is more flexibly organized than previous observations in cat. Recurrent inhibitory connections were relatively common between motoneurons controlling muscles that act at different joints, and between flexors and extensors. As in the cat, connections were minimal for motoneurons innervating the most distal intrinsic hand muscles. Empirical data are consistent with previous modeling: temporal properties of the recurrent inhibitory feedback loop are compatible with a role in reducing physiological tremor by suppressing oscillations around 10 Hz.

Theta, but Not Gamma Oscillations in Area V4 Depend on Input from Primary Visual Cortex

Theta (3-9 Hz) and gamma (30-100 Hz) oscillations have been observed at different levels along the hierarchy of cortical areas and across a wide set of cognitive tasks. In the visual system, the emergence of both rhythms in primary visual cortex (V1) and mid-level cortical areas V4 has been linked with variations in perceptual reaction times.1-5 Based on analytical methods to infer causality in neural activation patterns, it was concluded that gamma and theta oscillations might both reflect feedforward sensory processing from V1 to V4.6-10 Here, we report on experiments in macaque monkeys in which we experimentally assessed the presence of both oscillations in the neural activity recorded from multi-electrode arrays in V1 and V4 before and after a permanent V1 lesion. With intact cortex, theta and gamma oscillations could be reliably elicited in V1 and V4 when monkeys viewed a visual contour illusion and showed phase-to-amplitude coupling. Laminar analysis in V1 revealed that both theta and gamma oscillations occurred primarily in the supragranular layers, the cortical output compartment of V1. However, there was a clear dissociation between the two rhythms in V4 that became apparent when the major feedforward input to V4 was removed by lesioning V1: although V1 lesioning eliminated V4 theta, it had little effect on V4 gamma power except for delaying its emergence by >100 ms. These findings suggest that theta is more tightly associated with feedforward processing than gamma and pose limits on the proposed role of gamma as a feedforward mechanism.

Modelling and prediction of the dynamic responses of large-scale brain networks during direct electrical stimulation

Direct electrical stimulation can modulate the activity of brain networks for the treatment of several neurological and neuropsychiatric disorders and for restoring lost function. However, precise neuromodulation in an individual requires the accurate modelling and prediction of the effects of stimulation on the activity of their large-scale brain networks. Here, we report the development of dynamic input-output models that predict multiregional dynamics of brain networks in response to temporally varying patterns of ongoing microstimulation. In experiments with two awake rhesus macaques, we show that the activities of brain networks are modulated by changes in both stimulation amplitude and frequency, that they exhibit damping and oscillatory response dynamics, and that variabilities in prediction accuracy and in estimated response strength across brain regions can be explained by an at-rest functional connectivity measure computed without stimulation. Input-output models of brain dynamics may enable precise neuromodulation for the treatment of disease and facilitate the investigation of the functional organization of large-scale brain networks.

Multiscale low-dimensional motor cortical state dynamics predict naturalistic reach-and-grasp behavior

Motor function depends on neural dynamics spanning multiple spatiotemporal scales of population activity, from spiking of neurons to larger-scale local field potentials (LFP). How multiple scales of low-dimensional population dynamics are related in control of movements remains unknown. Multiscale neural dynamics are especially important to study in naturalistic reach-and-grasp movements, which are relatively under-explored. We learn novel multiscale dynamical models for spike-LFP network activity in monkeys performing naturalistic reach-and-grasps. We show low-dimensional dynamics of spiking and LFP activity exhibited several principal modes, each with a unique decay-frequency characteristic. One principal mode dominantly predicted movements. Despite distinct principal modes existing at the two scales, this predictive mode was multiscale and shared between scales, and was shared across sessions and monkeys, yet did not simply replicate behavioral modes. Further, this multiscale mode's decay-frequency explained behavior. We propose that multiscale, low-dimensional motor cortical state dynamics reflect the neural control of naturalistic reach-and-grasp behaviors.

Modeling behaviorally relevant neural dynamics enabled by preferential subspace identification

Neural activity exhibits complex dynamics related to various brain functions, internal states and behaviors. Understanding how neural dynamics explain specific measured behaviors requires dissociating behaviorally relevant and irrelevant dynamics, which is not achieved with current neural dynamic models as they are learned without considering behavior. We develop preferential subspace identification (PSID), which is an algorithm that models neural activity while dissociating and prioritizing its behaviorally relevant dynamics. Modeling data in two monkeys performing three-dimensional reach and grasp tasks, PSID revealed that the behaviorally relevant dynamics are significantly lower-dimensional than otherwise implied. Moreover, PSID discovered distinct rotational dynamics that were more predictive of behavior. Furthermore, PSID more accurately learned behaviorally relevant dynamics for each joint and recording channel. Finally, modeling data in two monkeys performing saccades demonstrated the generalization of PSID across behaviors, brain regions and neural signal types. PSID provides a general new tool to reveal behaviorally relevant neural dynamics that can otherwise go unnoticed.

Analysis of Visuo Motor Control between Dominant Hand and Non-Dominant Hand for Effective Human-Robot Collaboration

The human-in-the-loop technology requires studies on sensory-motor characteristics of each hand for an effective human-robot collaboration. This study aims to investigate the differences in visuomotor control between the dominant (DH) and non-dominant hands in tracking a target in the three-dimensional space. We compared the circular tracking performances of the hands on the frontal plane of the virtual reality space in terms of radial position error (ΔR), phase error (Δθ), acceleration error (Δa), and dimensionless squared jerk (DSJ) at four different speeds for 30 subjects. ΔR and Δθ significantly differed at relatively high speeds (ΔR: 0.5 Hz; Δθ: 0.5, 0.75 Hz), with maximum values of ≤1% compared to the target trajectory radius. DSJ significantly differed only at low speeds (0.125, 0.25 Hz), whereas Δa significantly differed at all speeds. In summary, the feedback-control mechanism of the DH has a wider range of speed control capability and is efficient according to an energy saving model. The central nervous system (CNS) uses different models for the two hands, which react dissimilarly. Despite the precise control of the DH, both hands exhibited dependences on limb kinematic properties at high speeds (0.75 Hz). Thus, the CNS uses a different strategy according to the model for optimal results.