Representations of uncertainty in sensorimotor control

Uncertainty is ubiquitous in our sensorimotor interactions, arising from factors such as sensory and motor noise and ambiguity about the environment. Setting it apart from previous theories, a quintessential property of the Bayesian framework for making inference about the state of world so as to select actions, is the requirement to represent the uncertainty associated with inferences in the form of probability distributions. In the context of sensorimotor control and learning, the Bayesian framework suggests that to respond optimally to environmental stimuli the central nervous system needs to construct estimates of the sensorimotor transformations, in the form of internal models, as well as represent the structure of the uncertainty in the inputs, outputs and in the transformations themselves. Here we review Bayesian inference and learning models that have been successful in demonstrating the sensitivity of the sensorimotor system to different forms of uncertainty as well as recent studies aimed at characterizing the representation of the uncertainty at different computational levels.     

Computational mechanisms of sensorimotor control

In order to generate skilled and efficient actions, the motor system must find solutions to several problems inherent in sensorimotor control, including nonlinearity, nonstationarity, delays, redundancy, uncertainty, and noise. We review these problems and five computational mechanisms that the brain may use to limit their deleterious effects: optimal feedback control, impedance control, predictive control, Bayesian decision theory, and sensorimotor learning. Together, these computational mechanisms allow skilled and fluent sensorimotor behavior.     

Scaling of sensorimotor control in terrestrial mammals

Sensorimotor control is greatly affected by two factors—the time it takes for an animal to sense and respond to stimuli (responsiveness), and the ability of an animal to distinguish between sensory stimuli and generate graded muscle forces (resolution). Here, we demonstrate that anatomical limitations force a necessary trade-off between responsiveness and resolution with increases in animal size. To determine whether responsiveness is prioritized over resolution, or resolution over responsiveness, we studied how size influences the physiological mechanisms underlying sensorimotor control. Using both new electrophysiological experiments and existing data, we determined the maximum axonal conduction velocity (CV) in animals ranging in size from shrews to elephants. Over the 100-fold increase in leg length, CV was nearly constant, increasing proportionally with mass to the 0.04 power. As a consequence, larger animals are burdened with relatively long physiological delays, which may have broad implications for their behaviour, ecology and evolution, including constraining agility and requiring prediction to help control movements.     

Role of presynaptic inputs to proprioceptive afferents in tuning sensorimotor pathways of an insect joint control network

The femur-tibia (FT) joint of insects is governed by a neuronal network that controls activity in tibial motoneurons by processing sensory information about tibial position and movement provided by afferents of the femoral chordotonal organ (fCO). We show that central arborizations of fCO afferents receive presynaptic depolarizing synaptic inputs. With an average resting potential of −71.9 ± 3.72 mV (n = 10), the reversal potential of these potentials is on average −62.8 ± 2.3 mV (n = 5). These synaptic potentials occur either spontaneously or are related to movements at the fCO. They are thus induced by signals from other fCO afferents. Therefore, the synaptic inputs to fCO afferents are specific and depend on the sensitivity of the individual afferent affected. These potentials reduce the amplitude of concurrent afferent action potentials. Bath application of picrotoxin, a noncompetitive blocker of chloride ion channels, blocks these potentials, which indicates that they are mediated by chloride ions. From these results, it is concluded that these are inhibitory synaptic potentials generated in the central terminals of fCO afferents. Pharmacologic removal of these potentials affects the tuning of the complete FT control system. Following removal, the dependence of the FT control loop on the tibia position increases relative to the dependency on the velocity of tibia movements. This is due to changes in the relative weighting of the position and velocity signals in the parallel interneuronal pathways from the fCO onto tibial motoneurons. Consequently, the FT joint is no longer able to perform twig mimesis (i.e., catalepsy), which is known to rely on a low position compared to the high-velocity dependency of the FT control system. © 1997 John Wiley & Sons, Inc. J Neurobiol 32: 359–376, 1997.     

A model for sensorimotor control and learning

A model for motor learning, generalization, and adaptation is presented. It is shown that the equations of motion of a limb can be expressed in a parametric form that facilitates transformation of desired trajectories into plans. These parametric equations are used in conjunction with a quantized multidimensional memory organized by state variables. The memory is supplied with data derived from the analysis of practice movements. A small computer and mechanical arm are used to implement the model and study its properties. Results verify the ability to acquire new movements, adapt to mechanical loads, and generalize between similar movements.This research was done while the author was a graduate student at the Massachusetts Institute of Technology in the Artificial Intelligence Laboratory and Department of Psychology. It was supported in part by training grant NGMS 5-T01-GM01064-15     

Upper extremity sensorimotor control among collegiate football players

Injuries stemming from shoulder instability are very common among athletes participating in contact sports, such as football. Previous research has shown that increased laxity negatively affects the function of the sensorimotor system potentially leading to a pathological cycle of shoulder dysfunction. Currently, there are no data detailing such effects among football players. Therefore, the purpose of this study was to examine the differences in upper extremity sensorimotor control among football players compared with that of a control group. Forty-five collegiate football players and 70 male control subjects with no previous experience in contact sports participated. All the subjects had no recent history of upper extremity injury. Each subject performed three 30-second upper extremity balance trials on each arm. The balance trials were conducted in a single-arm push-up position with the test arm in the center of a force platform and the subjects' feet on a labile device. The trials were averaged, and the differences in radial area deviation between groups were analyzed using separate 1-way analyses of variance (p < 0.05). The football players showed significantly more radial area deviation of the dominant (0.41 ± 1.23 cm2, p = 0.02) and nondominant arms (0.47 ± 1.63 cm2, p = 0.03) when compared with the control group. These results suggest that football players may have decreased sensorimotor control of the upper extremity compared with individuals with no contact sport experience. The decreased upper extremity sensorimotor control among the football players may be because of the frequent impacts accumulated during football participation. Football players may benefit from exercises that target the sensorimotor system. These findings may also be beneficial in the evaluation and treatment of various upper extremity injuries among football players.     

Resistance training alters the sensorimotor control of vasti muscles

The present study examined and compared two modes of weight training (bodybuilding and power-lifting) on the surface EMG of vasti muscles, knee joint position sense and isometric knee extension force in 48 able-bodied subjects. Subjects were randomly allocated into either a moderate loading and repetitions (bodybuilding) training or a high loading and low repetitions (power-lifting) training, or a no training control group. Training was conducted on alternate days with individual supervision. After 8 weeks of training, subjects from both training groups showed significantly earlier EMG onset timing and higher amplitude of vastus medialis obliquus relative to vastus lateralis (p = 0.005 or <0.001), and improved knee joint position sense (p < 0.001), but no such changes were found in the control group. However, the changes were not significantly different (p > 0.05) between the two training groups. The findings suggested that the neuromotor control of the vasti muscles could be altered by regular weight training.     

A neuroeconomics approach to inferring utility functions in sensorimotor control

Making choices is a fundamental aspect of human life. For over a century experimental economists have characterized the decisions people make based on the concept of a utility function. This function increases with increasing desirability of the outcome, and people are assumed to make decisions so as to maximize utility. When utility depends on several variables, indifference curves arise that represent outcomes with identical utility that are therefore equally desirable. Whereas in economics utility is studied in terms of goods and services, the sensorimotor system may also have utility functions defining the desirability of various outcomes. Here, we investigate the indifference curves when subjects experience forces of varying magnitude and duration. Using a two-alternative forced-choice paradigm, in which subjects chose between different magnitude–duration profiles, we inferred the indifference curves and the utility function. Such a utility function defines, for example, whether subjects prefer to lift a 4-kg weight for 30 s or a 1-kg weight for a minute. The measured utility function depends nonlinearly on the force magnitude and duration and was remarkably conserved across subjects. This suggests that the utility function, a central concept in economics, may be applicable to the study of sensorimotor control.     

Adaptive dynamic programming as a theory of sensorimotor control

Many characteristics of sensorimotor control can be explained by models based on optimization and optimal control theories. However, most of the previous models assume that the central nervous system has access to the precise knowledge of the sensorimotor system and its interacting environment. This viewpoint is difficult to be justified theoretically and has not been convincingly validated by experiments. To address this problem, this paper presents a new computational mechanism for sensorimotor control from a perspective of adaptive dynamic programming (ADP), which shares some features of reinforcement learning. The ADP-based model for sensorimotor control suggests that a command signal for the human movement is derived directly from the real-time sensory data, without the need to identify the system dynamics. An iterative learning scheme based on the proposed ADP theory is developed, along with rigorous convergence analysis. Interestingly, the computational model as advocated here is able to reproduce the motor learning behavior observed in experiments where a divergent force field or velocity-dependent force field was present. In addition, this modeling strategy provides a clear way to perform stability analysis of the overall system. Hence, we conjecture that human sensorimotor systems use an ADP-type mechanism to control movements and to achieve successful adaptation to uncertainties present in the environment.     

Similar sensorimotor transformations control balance during standing and walking

Standing and walking balance control in humans relies on the transformation of sensory information to motor commands that drive muscles. Here, we evaluated whether sensorimotor transformations underlying walking balance control can be described by task-level center of mass kinematics feedback similar to standing balance control. We found that delayed linear feedback of center of mass position and velocity, but not delayed linear feedback from ankle angles and angular velocities, can explain reactive ankle muscle activity and joint moments in response to perturbations of walking across protocols (discrete and continuous platform translations and discrete pelvis pushes). Feedback gains were modulated during the gait cycle and decreased with walking speed. Our results thus suggest that similar task-level variables, i.e. center of mass position and velocity, are controlled across standing and walking but that feedback gains are modulated during gait to accommodate changes in body configuration during the gait cycle and in stability with walking speed. These findings have important implications for modelling the neuromechanics of human balance control and for biomimetic control of wearable robotic devices. The feedback mechanisms we identified can be used to extend the current neuromechanical models that lack balance control mechanisms for the ankle joint. When using these models in the control of wearable robotic devices, we believe that this will facilitate shared control of balance between the user and the robotic device.     

Risk-sensitivity and the mean-variance trade-off: decision making in sensorimotor control

Numerous psychophysical studies suggest that the sensorimotor system chooses actions that optimize the average cost associated with a movement. Recently, however, violations of this hypothesis have been reported in line with economic theories of decision-making that not only consider the mean payoff, but are also sensitive to risk, that is the variability of the payoff. Here, we examine the hypothesis that risk-sensitivity in sensorimotor control arises as a mean-variance trade-off in movement costs. We designed a motor task in which participants could choose between a sure motor action that resulted in a fixed amount of effort and a risky motor action that resulted in a variable amount of effort that could be either lower or higher than the fixed effort. By changing the mean effort of the risky action while experimentally fixing its variance, we determined indifference points at which participants chose equiprobably between the sure, fixed amount of effort option and the risky, variable effort option. Depending on whether participants accepted a variable effort with a mean that was higher, lower or equal to the fixed effort, they could be classified as risk-seeking, risk-averse or risk-neutral. Most subjects were risk-sensitive in our task consistent with a mean-variance trade-off in effort, thereby, underlining the importance of risk-sensitivity in computational models of sensorimotor control.     

Role of cortical inhibition in heterosensory interaction of neurons of the cat sensorimotor cortex

Background and evoked neuronal activity in the cat sensorimotor cortex was recorded under a-chloralose anesthesia. Pairs of heterogeneous stimuli were applied, spaced at intervals of 0, 100, 200, 300, and 400 msec. A clicking sound, flashing light, and electroshock to the contralateral forepaw were used as stimuli. Partial or complete blockade of response to test stimuli presentations spaced 100–200 msec apart were observed when using stimulation of varying modality. The greatest test response was recorded at interstimulus intervals of 200–300 msec. Intracellular mechanisms of heterosensory interaction were investigated by applying the inhibitory transmitter antagonist picrotoxin microiontophoretically to the test cell to produce local attenuation of inhibitory effects. This substance also reduced the duration of blockage following the conditioning stimulus and the occurrence of peak level test response at a lower interstimulus interval than in the controls. Either a consistent increase in the number of spikes per response at one of the interstimulus intervals or a uniform reinforcement in unit response to several different interstimulus intervals were observed in a proportion of the cells. The contribution of intracortical inhibitory influences to the mechanisms of heterosensory interaction on neurons of the cat sensorimotor cortex is discussed in the light of our findings.A. A. Ukhtomskii Institute of Physiology, A. A. Zhdanov State University, Leningrad. Translated from Neirofiziologiya, Vol. 19, No. 2, pp. 147–156, March–April, 1987.     

Role of intracortical inhibition in forming homo- and heterosensory interaction in sensorimotor cortex neurons in kittens

Homo- and heterosensory interaction were investigated in sensorimotor cortex neurons before and after picrotoxin application to anesthetized and immobilized kittens belonging to three age groups (12–30 days, 31–47 days, and 2–4 months old). Only slight inhibition of response to presentation of a second stimulus was observed in a small proportion of cells in the youngest age group at test intervals of 100, 200, and 300 msec. Picrotoxin application only produced the effect of raised background activity. Numbers of neurons with partially or fully inhibited response to test stimuli (especially spaced at 100 msec intervals) rose in the middle and older age groups. The dynamics of heterosensory interaction and how this is affected by picrotoxin application gradually approximated to that observed with adult animals. The subject of the development of inhibitory mechanisms and how they contribute to the organization of homo- and heterosensory interaction during early postnatal ontogenesis is considered in the light of the results obtained.A. A. Ukhtomsii Institute of Physiology, Leningrad. A. A. Zhdanov State University, Leningrad. Translated from Neirofiziologiya, Vol. 20, No. 2, pp. 234–243, March–April, 1988.     

The role of spinal manipulation in addressing disordered sensorimotor integration and altered motor control

This review provides an overview of some of the growing body of research on the effects of spinal manipulation on sensory processing, motor output, functional performance and sensorimotor integration. It describes a body of work using somatosensory evoked potentials (SEPs), transcranial magnetic nerve stimulation, and electromyographic techniques to demonstrate neurophysiological changes following spinal manipulation. This work contributes to the understanding of how an initial episode(s) of back or neck pain may lead to ongoing changes in input from the spine which over time lead to altered sensorimotor integration of input from the spine and limbs.     

Role of proprioceptors in neural control

The conventional notion that peripheral muscle-related signals provide the basis for resistance to external perturbations is no longer sufficient. Proprioceptive information seems to be required for spatial steering of multi-joint movements, and also for temporal coordination among the joints in certain tasks. In rhythmic movements, peripheral and centrally generated signals appear to interact in a complementary manner. The complex effects of proprioceptive afferents on motor output continue to be delineated vigorously. Global effects of local perturbation in multi-joint contexts are emerging as being particularly significant.     

Learning in sensorimotor circuits

The study of plasticity in the central nervous system is a major and very dynamic neuroscience research field with enormous clinical potential. Considerable advances in this field have been made during the past 10 years. It now appears that most circuits in the brain and spinal cord show plasticity and that they can be modified by experience. Knowledge of the mechanisms of plasticity in the nervous system is therefore essential for the understanding of how the nervous system is wired during development and how it adapts in response to changes in the body and environment. Recent findings indicate that functional sensorimotor modules probe the sensory signals from the body that are generated as a consequence of module specific activity and use this sensory feedback to calibrate the strength in its input-output connections. This experience-dependent signal adapts the circuitry in the sensorimotor module to the body anatomy and biomechanics.     

Neural ensemble dynamics in trunk and hindlimb sensorimotor cortex encode for the control of postural stability

1. Download : Download high-res image (191KB)
2. Download : Download full-size image