Proprioceptive and Visual Feedback Responses in Macaques Exploit Goal Redundancy.

A self paced linear positioning task was used to study the effects of visual and proprioceptive feedback on learning and performance. Subjects were trained with knowledge of results (KR) and tested without it. The analysis of the absolute error scores of the no-KR trials is discussed in this paper. Visual feedback was the more effective source of sensory feedback, but proprioceptive feedback was also effective. An observation that the response did not become independent of sensory feedback as a result of learning, was interpreted as supporting Adams closed loop theory of motor learning in preference to the motor program hypothesis. Other data showed that the presence of visual feedback during learning could inhibit the later effectiveness of proprioceptive feedback.

In most applications of virtual environments (VEs), like training and design evaluation, a good sense of orientation is needed in the VE. Orientation performance when moving around in the real world relies on visual as well as proprioceptive feedback. However, the navigation metaphors which are used to move around the VE often lack proprioceptive feedback. Furthermore, the visual feedback in a VE is often relatively poor compared to the visual feedback available in the real world. Therefore, we have quantified the influence of visual and proprioceptive feedback on orientation performance in VEs. Subjects were immersed in a virtual forest and were asked to turn specific angles using three navigation metaphors, differing in the kind of proprioceptive feedback which is provided (no proprioceptive feedback, vestibular feedback, and vestibular and kinesthetic feedback). The results indicate that the most accurate turn performance is found when kinesthetic feedback is present, in a condition where subjects use their legs to turn around. This indicates that incorporating this kind of feedback in navigation metaphors is quite beneficial. Orientation on only the visual component is most inaccurate, leading to progressively larger undershoots for larger angles.

We routinely have physical interactions with others, whether it be handing someone a glass of water or jointly moving a heavy object together. These sensorimotor interactions between humans typically rely on visual feedback and haptic feedback. Recent single-participant studies have highlighted that the unique noise and time delays of each sense must be considered to estimate the state, such as the position and velocity, of one's own movement. However, we know little about how visual feedback and haptic feedback are used to estimate the state of another person. Here, we tested how humans utilize visual feedback and haptic feedback to estimate the state of their partner during a collaborative sensorimotor task. Across two experiments, we show that visual feedback dominated haptic feedback during collaboration. Specifically, we found that visual feedback led to comparatively lower task-relevant movement variability, smoother collaborative movements, and faster trial completion times. We also developed an optimal feedback controller that considered the noise and time delays of both visual feedback and haptic feedback to estimate the state of a partner. This model was able to capture both lower task-relevant movement variability and smoother collaborative movements. Taken together, our empirical and modeling results support the idea that visual accuracy is more important than haptic speed to perform state estimation of a partner during collaboration.NEW & NOTEWORTHY Physical collaboration between two or more individuals involves both visual and haptic feedback. Here, we investigated how visual and haptic feedback is used to estimate the movements of a partner during a collaboration task. Our experimental and computational modeling results parsimoniously support the notion that greater visual accuracy is more important than faster yet noisier haptic feedback when estimating the state of a partner.

Humans have the ability to identify external physical properties because skilled object manipulation requires knowledge, or internal models, of object dynamics relating to applied force or movement. The accuracy of identification depends on those of the input and output signal to the object : motor command from brain as the input and information from multiple sensory modalities as the output. In this paper, control strategies of isometric contraction by hand are investigated. Subjects are asked to generate constant levels of force under two conditions : force production tasks a) by self scaling without visual feedback and b) with visual feed back of reference force levels. The results show subjective assessments of force are larger than objective one, and it is indicated that the variability is composed of signal dependent noise and error of human intention. It is also demonstrated that visual feedback information has an important role of generating accurate reference force level.

Previous research has questioned whether motor adaptation is shaped by an optimal combination of multisensory error signals. Here, we expanded on this work by investigating how the use of visual and somatosensory error signals during online correction influences single-trial adaptation. To this end, we exposed participants to a random sequence of force-field perturbations and recorded their corrective responses as well as the after-effects exhibited during the subsequent unperturbed movement. In addition to the force perturbation we artificially decreased or increased visual errors by multiplying hand deviations by a gain smaller or larger than one. Corrective responses to the force perturbation clearly scaled with the size of the visual error, but this scaling did not transfer one-to-one to motor adaptation and we observed no consistent interaction between limb and visual errors on adaptation. However, reducing visual errors during perturbation led to a small reduction of after-effects and this residual influence of visual feedback was eliminated when we instructed participants to control their hidden hand instead of the visual hand cursor. Taken together, our results demonstrate that task instructions and the need to correct for errors during perturbation are important factors to consider if we want to understand how the sensorimotor system uses and combines multimodal error signals to adapt movements.Significance StatementWe investigated the factors influencing visual and proprioceptive feedback contributions to movement control and adaptation. While online corrections increased with the size of visual errors, this scaling did not transfer one-to-one to adaptation. Instead, we observed a consistent relationship between limb displacement during perturbation and subsequent after-effects which was independent of the visual error. However, adaptation was slightly reduced when visual errors were artificially decreased and this influence of vision was modulated by the task-instruction given to participants. Our results demonstrate that task-related factors, such as the need to correct movement errors and the instructions, have to be considered to advance our understanding of how the sensorimotor system uses multisensory feedback to adapt motor commands.

Kinematic and EMG characteristics of simple shoulder movements with proprioception and visual feedback

Development of virtual reality proprioceptive rehabilitation system for stroke patients

Neural Correlates of State Estimation in Visually Guided Movements: an Event-Related FMRI Study

This paper addresses a fundamental question of how motor commands specify target torque levels. Human subjects produced fast and accurate changes in torque with the isometric elbow joint. Visual stimuli were used to indicate target torque levels as well as to cue subjects to initiate their responses. During rapid changes in torque from one steady-state level to another, target torque was achieved through a sequence of approximations. During the first 200-250 msec of responses produced in the presence of visual feedback, 3 distinct control mechanisms were recruited to guide torque to the target level. The timing and accuracy of each control mechanism were evaluated. The first control mechanism was triggered by the visual stimulus and produced the initial rise in torque. Target torque predictability was found to strongly influence the accuracy of this control mechanism. The second control mechanism produced a corrective adjustment in torque within roughly the first 100 msec of responses. This mechanism incorporated target torque information provided by the stimulus into the response. The third control mechanism began 200-250 msec after response onset and produced corrective adjustments based on visual feedback of torque errors. The stability of the visual feedback mechanism was evaluated because of a long loop delay. Two strategies were used to control stability: low gain and information transfer between the visual feedback mechanism and the preceding (second) control mechanism.

Effectiveness of instruction and video feedback on staff's trainer behavior during one-to-one training with children with severe intellectual disabilities

Responses elicited after the shortest latency spinal reflexes but prior to the onset of voluntary activity can display sophistication beyond a stereotypical reflex. Two distinct behaviors have been identified for these rapid motor responses, often called long-latency reflexes. The first is to maintain limb stability by opposing external perturbations. The second is to quickly release motor actions planned prior to the disturbance, often called a triggered reaction. This study investigated their interaction when motor tasks involve both limb stabilization and motor planning. We used a robotic manipulator to change the stability of the haptic environment during 2D arm reaching tasks, and to apply perturbations that could elicit rapid motor responses. Stabilizing reflexes were modulated by the orientation of the haptic environment (field effect) whereas triggered reactions were modulated by the target to which subjects were instructed to reach (target effect). We observed that there were no significant interactions between the target and field effects in the early (50–75 ms) portion of the long-latency reflex, indicating that these components of the rapid motor response are initially controlled independently. There were small but significant interactions for two of the six relevant muscles in the later portion (75–100 ms) of the reflex response. In addition, the target effect was influenced by the direction of the perturbation used to elicit the motor response, indicating a later feedback correction in addition to the early component of the triggered reaction. Together, these results demonstrate how distinct components of the long-latency reflex can work independently and together to generate sophisticated rapid motor responses that integrate planning with reaction to uncertain conditions.

Recent advances in neuroscience and engineering have led to the development of technologies that permit the control of external devices through real-time decoding of brain activity (brain-machine interfaces; BMI). Though the feeling of controlling bodily movements (sense of agency; SOA) has been well studied and a number of well-defined sensorimotor and cognitive mechanisms have been put forth, very little is known about the SOA for BMI-actions. Using an on-line BMI, and verifying that our subjects achieved a reasonable level of control, we sought to describe the SOA for BMI-mediated actions. Our results demonstrate that discrepancies between decoded neural activity and its resultant real-time sensory feedback are associated with a decrease in the SOA, similar to SOA mechanisms proposed for bodily actions. However, if the feedback discrepancy serves to correct a poorly controlled BMI-action, then the SOA can be high and can increase with increasing discrepancy, demonstrating the dominance of visual feedback on the SOA. Taken together, our results suggest that bodily and BMI-actions rely on common mechanisms of sensorimotor integration for agency judgments, but that visual feedback dominates the SOA in the absence of overt bodily movements or proprioceptive feedback, however erroneous the visual feedback may be.

Skilled motor performance depends critically on rapid corrective responses that act to preserve the goal of the movement in the face of perturbations. Although it is well established that the gain of corrective responses elicited while reaching toward objects adapts to different contexts, little is known about the adaptability of corrective responses supporting the manipulation of objects after they are grasped. Here, we investigated the adaptability of the corrective response elicited when an object being lifted is heavier than expected and fails to lift off when predicted. This response involves a monotonic increase in vertical load force triggered, within ∼90 ms, by the absence of expected sensory feedback signaling lift off and terminated when actual lift off occurs. Critically, because the actual weight of the object cannot be directly sensed at the moment the object fails to lift off, any adaptation of the corrective response would have to be based on memory from previous lifts. We show that when humans, including men and women, repeatedly lift an object that on occasional catch trials increases from a baseline weight to a fixed heavier weight, they scale the gain of the response (i.e., the rate of force increase) to the heavier weight within two to three catch trials. We also show that the gain of the response scales, on the first catch trial, with the baseline weight of the object. Thus, the gain of the lifting response can be adapted by both short- and long-term experience. Finally, we demonstrate that this adaptation preserves the efficacy of the response across contexts.SIGNIFICANCE STATEMENT Here, we present the first investigation of the adaptability of the corrective lifting response elicited when an object is heavier than expected and fails to lift off when predicted. A striking feature of the response, which is driven by a sensory prediction error arising from the absence of expected sensory feedback, is that the magnitude of the error is unknown. That is, the motor system only receives a categorical error indicating that the object is heavier than expected but not its actual weight. Although the error magnitude is not known at the moment the response is elicited, we show that the response can be scaled to predictions of error magnitude based on both recent and long-term memories.

Oscillatory synchronization in the beta (approximately 20 Hz) band is a common feature of human motor control, manifest at cortical and muscular levels during tonic contraction. Here we test the hypothesis that the influence of visual feedback on performance in a positional hold task is increased during bursts of beta-band synchrony in the corticospinal motor system. Healthy subjects were instructed to extend their forefinger while receiving high-gain visual feedback of finger position on a PC screen. Small step displacements of the feedback signal were triggered either by bursts of beta oscillations in scalp electroencephalogram or randomly with respect to cortical beta activity, and the resulting positional corrections expressed as a percentage of the step displacement. Corrective responses to beta and randomly triggered step changes in visual feedback were 41.7+/-4.9 and 31.5+/-6.8%, respectively (P<0.05). A marked increase in the coherence in the beta band was also found between muscle activity and cortical activity during the beta-triggered condition. The results suggest that phasic elevations of beta activity in the corticospinal motor system are associated with an increase in the gain of the motor response to visual feedback during a tonic hold task. Beta activity may index a motor state in which processing relevant to the control of positional hold tasks is promoted, with behavioural consequences.

Adapting movements in relation to visual feedback is a ubiquitous characteristic of sensorimotor control and involves the integration of multiple sources of sensory information. We recorded sensory feedback from muscle spindle afferents during visuomotor adaptation while subjects performed an aiming task to investigate whether the activity of the muscle spindles was modulated by the fusimotor system under these learning conditions. None of the muscle spindles showed an increase in activity, rather in 83% of the trials the firing rates were decreased. These preliminary results suggest that the CNS reduces the sensory signals arising from muscle spindles perhaps as a means of resolving the conflict between visual and proprioceptive feedback during the task.