Using Computational Neuroscience to Define Common Input to Spinal Motor Neurons

Fatigue-related changes in intermuscular electromyographic coherence across rotator cuff and deltoid muscles in individuals with and without subacromial pain

Neuromuscular fatigue induces superior migration of the humeral head in individuals with subacromial pain. This has been attributed to weakness of rotator cuff muscles and overactive deltoid muscles. Investigation of common inputs to motoneuron pools of the rotator cuff and deltoid muscles offers valuable insight into the underlying mechanisms of neuromuscular control deficits associated with subacromial pain. This study aims to investigate intermuscular coherence across the rotator cuff and deltoid muscles during a sustained submaximal isometric fatiguing contraction in individuals with and without subacromial pain. Twenty symptomatic and 18 asymptomatic young adults participated in this study. Surface electromyogram (EMG) was recorded from the middle deltoid (MD) and infraspinatus (IS). Intramuscular EMG was recorded with fine-wire electrodes in the supraspinatus (SS). Participants performed an isometric fatiguing contraction of 30° scaption at 25% maximum voluntary contraction (MVC) until endurance limit. Pooled coherence of muscle pairs (SS-IS, SS-MD, IS-MD) in the 2-5 Hz (delta), 5-15 Hz (alpha), and 15-35 Hz (beta) frequency bands during the initial and final 30 s of the fatigue task were compared. SS-IS and SS-MD delta-band coherence increased with fatigue in the asymptomatic group but not the symptomatic group. In the alpha and beta bands, SS-IS and SS-MD coherence increased with fatigue in both groups. IS-MD beta-band coherence was greater in the symptomatic than the asymptomatic group. Individuals with subacromial pain failed to increase common drive across rotator cuff and deltoid muscles and have altered control strategies during neuromuscular fatigue. This may contribute to glenohumeral joint instability and subacromial pain experienced by these individuals.NEW & NOTEWORTHY Through the computation of shared neural drive across glenohumeral muscles, this study reveals that individuals with subacromial pain were unable to increase shared neural drive within the rotator cuff and across the supraspinatus and deltoid muscles during neuromuscular fatigue induced by sustained isometric contraction. These deficits in common drive across the shoulder muscles likely contribute to the joint instability and pain experienced by these individuals.

The Effects of Conscious Movement Processing on the Neuromuscular Control of Posture

Maintaining balance is thought to primarily occur sub-consciously. Occasionally, however, individuals will direct conscious attention towards balance, e.g., in response to a threat to balance. Such conscious movement processing (CMP) increases the reliance on attentional resources and may disrupt balance performance. However, the underlying changes in neuromuscular control remain poorly understood. We investigated the effects of CMP (manipulated using verbal instructions) on neural control of posture in twenty-five adults (11 females, mean age = 23.9, range = 18-33). Participants performed 90-s, bipedal stance balance trials in high- and low-CMP conditions, during both stable (solid surface) and unstable (foam) task conditions. Postural sway amplitude, frequency and complexity were used to assess postural control. Surface EMG was recorded bilaterally from lower leg muscles (Soleus, Tibialis Anterior, Gastrocnemius Medialis, Peroneus Longus) and intermuscular coherence (IMC) was assessed for 12 muscle pairs across four frequency bands. We observed significantly increased sway amplitude, and decreased sway frequency and complexity in the high- compared to the low-CMP conditions. All sway variables increased in the unstable compared to the stable conditions. We observed reduced beta band IMC between several muscle pairs during high- compared to low-CMP, but these findings did not remain significant after controlling for multiple comparisons. Finally, IMC significantly increased in the unstable conditions for most muscle combinations and frequency bands. In all, results tentatively suggest that CMP-induced changes in sway outcomes may be facilitated by reduced beta-band IMC, but these findings need to be replicated before they can be interpreted more conclusively.

Estimating muscle activation from EMG using deep learning-based dynamical systems models

Objective. To study the neural control of movement, it is often necessary to estimate how muscles are activated across a variety of behavioral conditions. One approach is to try extracting the underlying neural command signal to muscles by applying latent variable modeling methods to electromyographic (EMG) recordings. However, estimating the latent command signal that underlies muscle activation is challenging due to its complex relation with recorded EMG signals. Common approaches estimate each muscle's activation independently or require manual tuning of model hyperparameters to preserve behaviorally-relevant features.Approach. Here, we adapted AutoLFADS, a large-scale, unsupervised deep learning approach originally designed to de-noise cortical spiking data, to estimate muscle activation from multi-muscle EMG signals. AutoLFADS uses recurrent neural networks to model the spatial and temporal regularities that underlie multi-muscle activation.Main results. We first tested AutoLFADS on muscle activity from the rat hindlimb during locomotion and found that it dynamically adjusts its frequency response characteristics across different phases of behavior. The model produced single-trial estimates of muscle activation that improved prediction of joint kinematics as compared to low-pass or Bayesian filtering. We also applied AutoLFADS to monkey forearm muscle activity recorded during an isometric wrist force task. AutoLFADS uncovered previously uncharacterized high-frequency oscillations in the EMG that enhanced the correlation with measured force. The AutoLFADS-inferred estimates of muscle activation were also more closely correlated with simultaneously-recorded motor cortical activity than were other tested approaches.Significance.This method leverages dynamical systems modeling and artificial neural networks to provide estimates of muscle activation for multiple muscles. Ultimately, the approach can be used for further studies of multi-muscle coordination and its control by upstream brain areas, and for improving brain-machine interfaces that rely on myoelectric control signals.

Individual differences in the neural strategies to control the lateral and medial head of the quadriceps during a mechanically constrained task

We observed that the distribution of the strength of neural drive between the vastus lateralis and vastus medialis during a single-joint isometric task varied across participants. Also, we observed that the proportion of neural drive that was shared within and between these muscles also varied across participants. These results provide evidence that the neural strategies to control the vastus lateralis and vastus medialis muscles widely vary across individuals, even during a mechanically constrained task.

Muscle Synergies and Coherence Networks Reflect Different Modes of Coordination During Walking

When walking speed is increased, the frequency ratio between the arm and leg swing switches spontaneously from 2:1 to 1:1. We examined whether these switches are accompanied by changes in functional connectivity between multiple muscles. Subjects walked on a treadmill with their arms swinging along their body while kinematics and surface electromyography (EMG) of 26 bilateral muscles across the body were recorded. Walking speed was varied from very slow to normal. We decomposed EMG envelopes and intermuscular coherence spectra using non-negative matrix factorization (NMF), and the resulting modes were combined into multiplex networks and analyzed for their community structure. We found five relevant muscle synergies that significantly differed in activation patterns between 1:1 and 2:1 arm-leg coordination and the transition period between them. The corresponding multiplex network contained a single module indicating pronounced muscle co-activation patterns across the whole body during a gait cycle. NMF of the coherence spectra distinguished three EMG frequency bands: 4-8, 8-22, and 22-60 Hz. The community structure of the multiplex network revealed four modules, which clustered functional and anatomical linked muscles across modes of coordination. Intermuscular coherence at 4-22 Hz between upper and lower body and within the legs was particularly pronounced for 1:1 arm-leg coordination and was diminished when switching between modes of coordination. These findings suggest that the stability of arm-leg coordination is associated with modulations in long-distant neuromuscular connectivity.

The human central nervous system transmits common synaptic inputs to distinct motor neuron pools during non-synergistic digit actions

KEY POINTS

Neural connectivity between distinct motor neuronal modules in the spinal cord is classically studied through electrical stimulation or multi-muscle EMG recordings. We quantified the strength of correlation in the activity of two distinct populations of motor neurons innervating the thenar and first dorsal interosseous muscles during tasks that required the two hand muscles to exert matched or un-matched forces in different directions. We show that when the two hand muscles are concurrently activated, synaptic input to the two motor neuron pools is shared across all frequency bandwidths (representing cortical and spinal input) associated with force control. The observed connectivity indicates that motor neuron pools receive common input even when digit actions do not belong to a common behavioural repertoire.

ABSTRACT

Neural connectivity between distinct motor neuronal modules in the spinal cord is classically studied through electrical stimulation or multi-muscle EMG recordings. Here we quantify the strength of correlation in the activity of two distinct populations of motor neurons innervating the thenar and first dorsal interosseous muscles in humans during voluntary contractions. To remove confounds associated with previous studies, we used a task that required the two hand muscles to exert matched or un-matched forces in different directions. Despite the force production task consisting of uncommon digit force coordination patterns, we found that synaptic input to motor neurons is shared across all frequency bands, reflecting cortical and spinal inputs associated with force control. The coherence between discharge timings of the two pools of motor neurons was significant at the delta (0-5 Hz), alpha (5-15 Hz) and beta (15-35 Hz) bands (P < 0.05). These results suggest that correlated input to motor neurons of two hand muscles can occur even during tasks not belonging to a common behavioural repertoire and despite lack of common innervation. Moreover, we show that the extraction of activity from motor neurons during voluntary force control removes cross-talk associated with global EMG recordings, thus allowing direct in vivo interrogation of spinal motor neuron activity.

Functional connectivity analysis of multiplex muscle network across frequencies

Physiological networks reveal information about the interaction between subsystems of the human body. Here we investigated the interaction between the central nervous system and the musculoskeletal system by mapping functional muscle networks. Muscle networks were extracted using coherence analysis of muscle activity assessed using surface electromyography (EMG). Surface EMG was acquired from 36 muscles distributed throughout the body while participants were standing upright and performing a bimanual pointing task. Non-negative matrix factorization revealed functional connectivity in four frequency bands. The spatial arrangement differed considerably across frequencies supporting a multiplex network organisation. Graph-theory analysis of layer-specific network revealed a consistent fat-tail distribution of the edges weights, distinct efficiency values, and core-periphery properties. These frequency bands may be spectral fingerprints of different neural pathways that innervate the spinal motor neurons to control the musculoskeletal system.

Variability in common synaptic input to motor neurons modulates both force steadiness and pegboard time in young and older adults

KEY POINTS

The fluctuations in force during a steady isometric contraction (force steadiness) are associated with oscillations in common synaptic input to the involved motor neurons. Decreases in force steadiness are associated with increases in pegboard times in older adults, although a mechanism for this link has not been established. We used a state-space model to estimate the variability in common synaptic input to motor neurons during steady, isometric contractions. The estimate of common synaptic input was derived from the discharge times of motor units as recorded with high-density surface electrodes. We found that the variability in common synaptic input to motor neurons modulates force steadiness for young and older adults, as well as pegboard time for older adults.

ABSTRACT

We investigated the associations between grooved pegboard times, force steadiness (coefficient of variation for force) and variability in an estimate of the common synaptic input to motor neurons innervating the wrist extensor muscles during steady contractions performed by young and older adults. The discharge times of motor units were derived from recordings obtained with high-density surface electrodes when participants performed steady isometric contractions at 10% and 20% of maximal voluntary contraction force. The steady contractions were performed with a pinch grip and wrist extension, both independently (single action) and concurrently (double action). The variance in common synaptic input to motor neurons was estimated with a state-space model of the latent common input dynamics. There was a statistically significant association between the coefficient of variation for force during the steady contractions and the estimated variance in common synaptic input in young (r2  = 0.31) and older (r2  = 0.39) adults, although not between either the mean or the coefficient of variation for interspike interval of single motor units with the coefficient of variation for force. Moreover, the estimated variance in common synaptic input during the double-action task with the wrist extensors at the 20% target was significantly associated with grooved pegboard time (r2  = 0.47) for older adults but not young adults. These findings indicate that longer pegboard times of older adults were associated with worse force steadiness and greater fluctuations in the estimated common synaptic input to motor neurons during steady contractions.

Decrease in force steadiness with aging is associated with increased power of the common but not independent input to motor neurons

Declines in motor function with advancing age have been attributed to changes occurring at all levels of the neuromuscular system. However, the impact of aging on the control of muscle force by spinal motor neurons is not yet understood. In this study on 20 individuals aged between 24 and 75 yr (13 men, 7 women), we investigated the common synaptic input to motor neurons of the tibialis anterior muscle and its impact on force control. Motor unit discharge times were identified from high-density surface EMG recordings during isometric contractions at forces of 20% of maximal voluntary effort. Coherence analysis between motor unit spike trains was used to characterize the input to motor neurons. The decrease in force steadiness with age ( R² = 0.6, P < 0.01) was associated with an increase in the amplitude of low-frequency oscillations of functional common synaptic input to motor neurons ( R² = 0.59; P < 0.01). The relative proportion of common input to independent noise at low frequencies increased with variability (power) in common synaptic input. Moreover, variability in interspike interval did not change and strength of the common input in the gamma band decreased with age ( R² = 0.22; P < 0.01). The findings indicate that age-related reduction in the accuracy of force control is associated with increased common fluctuations to motor neurons at low frequencies and not with an increase in independent synaptic input. NEW & NOTEWORTHY The influence of aging on the role of spinal motor neurons in accurate force control is not yet understood. We demonstrate that aging is associated with increased oscillations in common input to motor neurons at low frequencies and with a decrease in the relative strength of gamma oscillations. These results demonstrate that the synaptic inputs to motor neurons change across the life span and contribute to a decline in force control.

Network structure of the human musculoskeletal system shapes neural interactions on multiple time scales

Human motor control requires the coordination of muscle activity under the anatomical constraints imposed by the musculoskeletal system. Interactions within the central nervous system are fundamental to motor coordination, but the principles governing functional integration remain poorly understood. We used network analysis to investigate the relationship between anatomical and functional connectivity among 36 muscles. Anatomical networks were defined by the physical connections between muscles, and functional networks were based on intermuscular coherence assessed during postural tasks. We found a modular structure of functional networks that was strongly shaped by the anatomical constraints of the musculoskeletal system. Changes in postural tasks were associated with a frequency-dependent reconfiguration of the coupling between functional modules. These findings reveal distinct patterns of functional interactions between muscles involved in flexibly organizing muscle activity during postural control. Our network approach to the motor system offers a unique window into the neural circuitry driving the musculoskeletal system.

A latent low-dimensional common input drives a pool of motor neurons: a probabilistic latent state-space model

Motor neurons appear to be activated with a common input signal that modulates the discharge activity of all neurons in the motor nucleus. It has proven difficult for neurophysiologists to quantify the variability in a common input signal, but characterization of such a signal may improve our understanding of how the activation signal varies across motor tasks. Contemporary methods of quantifying the common input to motor neurons rely on compiling discrete action potentials into continuous time series, assuming the motor pool acts as a linear filter, and requiring signals to be of sufficient duration for frequency analysis. We introduce a space-state model in which the discharge activity of motor neurons is modeled as inhomogeneous Poisson processes and propose a method to quantify an abstract latent trajectory that represents the common input received by motor neurons. The approach also approximates the variation in synaptic noise in the common input signal. The model is validated with four data sets: a simulation of 120 motor units, a pair of integrate-and-fire neurons with a Renshaw cell providing inhibitory feedback, the discharge activity of 10 integrate-and-fire neurons, and the discharge times of concurrently active motor units during an isometric voluntary contraction. The simulations revealed that a latent state-space model is able to quantify the trajectory and variability of the common input signal across all four conditions. When compared with the cumulative spike train method of characterizing common input, the state-space approach was more sensitive to the details of the common input current and was less influenced by the duration of the signal. The state-space approach appears to be capable of detecting rather modest changes in common input signals across conditions.NEW & NOTEWORTHY We propose a state-space model that explicitly delineates a common input signal sent to motor neurons and the physiological noise inherent in synaptic signal transmission. This is the first application of a deterministic state-space model to represent the discharge characteristics of motor units during voluntary contractions.

Correlation between discharge timings of pairs of motor units reveals the presence but not the proportion of common synaptic input to motor neurons

We investigated whether correlation measures derived from pairs of motor unit (MU) spike trains are reliable indicators of the degree of common synaptic input to motor neurons. Several 50-s isometric contractions of the biceps brachii muscle were performed at different target forces ranging from 10 to 30% of the maximal voluntary contraction relying on force feedback. Forty-eight pairs of MUs were examined at various force levels. Motor unit synchrony was assessed by cross-correlation analysis using three indexes: the output correlation as the peak of the cross-histogram (ρ) and the number of synchronous spikes per second (CIS) and per trigger (E). Individual analysis of MU pairs revealed that ρ, CIS, and E were most often positively associated with discharge rate (87, 85, and 76% of the MU pairs, respectively) and negatively with interspike interval variability (69, 65, and 62% of the MU pairs, respectively). Moreover, the behavior of synchronization indexes with discharge rate (and interspike interval variability) varied greatly among the MU pairs. These results were consistent with theoretical predictions, which showed that the output correlation between pairs of spike trains depends on the statistics of the input current and motor neuron intrinsic properties that differ for different motor neuron pairs. In conclusion, the synchronization between MU firing trains is necessarily caused by the (functional) common input to motor neurons, but it is not possible to infer the degree of shared common input to a pair of motor neurons on the basis of correlation measures of their output spike trains.NEW & NOTEWORTHY The strength of correlation between output spike trains is only poorly associated with the degree of common input to the population of motor neurons. The synchronization between motor unit firing trains is necessarily caused by the (functional) common input to motor neurons, but it is not possible to infer the degree of shared common input to a pair of motor neurons on the basis of correlation measures of their output spike trains.