Visual monitoring of goal-directed aiming movements.

Goal-directed movements are subject to intrinsic planning and execution variability, which requires that the central nervous system closely monitor our movements to ensure endpoint accuracy. In the present study, we sought to determine how closely the visual system monitored goal-directed aiming movements. We used a cursor-jump paradigm in which a cursor was unexpectedly translated soon after movement initiation. Some of the trials included a second cursor jump, and the cursor remained visible for different durations. The results indicate that seeing the cursor for only 16 ms after the second cursor jump was sufficient to influence the movement endpoint, which suggests that the visual system continuously monitored goal-directed movements. The results also suggest that the perceived position/trajectory of the effector was likely to have been averaged over a period of approximately 70 ms.

The specificity of practice hypothesis in goal-directed movements: visual dominance or proprioception neglect?

The study aimed to examine whether modifying the proprioceptive feedback usually associated with a specific movement would decrease the dominance of visual feedback and/or decrease, which appears to be the neglect of proprioceptive feedback in ensuring the accuracy of goal-directed movements. We used a leg positioning recall task and measured the recall error after 15 and 165 acquisition trials performed with both vision and proprioception or proprioception only, under either a normal or a modified proprioception condition (i.e., with a 1-kg load attached to the participants' ankle). Participant learning was evaluated in transfer with proprioception only. In support of the specificity of practice hypothesis, the recall errors in acquisition were significantly smaller when practice occurred with both vision and proprioception, in either the loaded or the unloaded leg condition, and they increased significantly in transfer when vision was withdrawn. An important finding of the study highlighted that withdrawing vision after 165 acquisition trials had less deleterious effects on the recall errors when practice occurred under the loaded leg condition. Under that modified condition, recall errors in transfer were similar when practice occurred with and without vision, whereas larger errors were observed following practice with vision under the normal proprioceptive condition. Overall, these results highlighted the dominance of vision in ensuring accurate leg positioning recall and revealed that the dominance of vision is such that the processing of proprioceptive feedback may be neglected. Importantly, modifying the proprioceptive feedback has the advantage of reducing what appears to be the neglect of proprioceptive information when movement execution occurs in a visuo-proprioceptive context. Practical considerations for rehabilitation are discussed at the end of the manuscript.

Visual Online Control of Goal-Directed Aiming Movements in Children.

The present study investigated whether the initial impulse of goal-directed movements was visually monitored by 5- to 12-years-old children (n = 36) in a manner similar to adults (n = 12). The participants moved a cursor toward a fixed target. In some trials, the cursor was unpredictably translated by 20 mm following movement initiation. The results showed that even the youngest children visually monitor the initial impulse of goal-directed movements. This monitoring and the error correction process that it triggers seem automatic because it occurs even when the cursor jump is not consciously detected. Finally, it appears that this process does not fully mature before late childhood, which suggests that a putative dedicated channel for processing visual hand information develops during childhood.

Observational Learning: Tell Beginners What They Are about to Watch and They Will Learn Better.

Observation aids motor skill learning. When multiple models or different levels of performance are observed, does learning improve when the observer is informed of the performance quality prior to each observation trial or after each trial? We used a knock-down barrier task and asked participants to learn a new relative timing pattern that differed from that naturally emerging from the task constraints (Blandin et al., 1999). Following a physical execution pre-test, the participants observed two models demonstrating different levels of performance and were either informed of this performance prior to or after each observation trial. The results of the physical execution retention tests of the two experiments reported in the present study indicated that informing the observers of the demonstration quality they were about to see aided learning more than when this information was provided after each observation trial. Our results suggest that providing advanced information concerning the quality of the observation may help participants detect errors in the model's performance, which is something that novice participants have difficulty doing, and then learn from these observations.

Prism adaptation in virtual and natural contexts: Evidence for a flexible adaptive process.

Prism exposure when aiming at a visual target in a virtual condition (e.g., when the hand is represented by a video representation) produces no or only small adaptations (after-effects), whereas prism exposure in a natural condition produces large after-effects. Some researchers suggested that this difference may arise from distinct adaptive processes, but other studies suggested a unique process. The present study reconciled these conflicting interpretations. Forty participants were divided into two groups: One group used visual feedback of their hand (natural context), and the other group used computer-generated representational feedback (virtual context). Visual feedback during adaptation was concurrent or terminal. All participants underwent laterally displacing prism perturbation. The results showed that the after-effects were twice as large in the "natural context" than in the "virtual context". No significant differences were observed between the concurrent and terminal feedback conditions. The after-effects generalized to untested targets and workspace. These results suggest that prism adaptation in virtual and natural contexts involves the same process. The smaller after-effects in the virtual context suggest that the depth of adaptation is a function of the degree of convergence between the proprioceptive and visual information that arises from the hand.

Is visual-based, online control of manual-aiming movements disturbed when adapting to new movement dynamics?

Previous research has shown that for goal-directed movements, online visual feedback is not necessary for the adaptation of movement planning to novel movement dynamics. In the present study, we wanted to put this proposition to a stringent test and determine whether the usually dominant role of online visual feedback in movement control is diminished when goal-directed movements are performed in a condition that modifies limb dynamics. Participants performed a video-aiming task while the center of mass of their forearm was experimentally displaced by a 1.5-kg mass attached laterally to its longitudinal axis. A cursor representing the position of the participant's hand was either visible or not visible during the acquisition phase. Then, in a transfer test, the participants performed the task without online visual feedback and either with or without the lateral mass. During the acquisition phase, the participants adapted to the new movement dynamics imposed by the added mass regardless of whether online visual feedback was available. An important new finding of the present study was the observation that the role usually played by online visual feedback in refining movement planning and ensuring control of the initial portion of goal-directed movements was suppressed during adaptation to novel movement dynamics. This resulted in an increase in the role played by visual feedback late in the movement to ensure endpoint accuracy.

Mixed observation favors motor learning through better estimation of the model's performance.

Observation contributes to motor learning. It was recently demonstrated that the observation of both a novice and an expert model (mixed observation) resulted in better learning of a complex spatio-temporal task than the observation of either a novice or an expert model alone. In the present study, we sought to determine whether the advantage of mixed observation resulted from the development of a better error detection mechanism. The results revealed that mixed observation resulted in a better estimation of the model's performance than that with other regimens of observation. The results also suggest that observational learning is improved when observation with knowledge of the results (KR) is followed by an observation phase without KR.

Observation learning of a motor task: who and when?

Observation contributes to motor learning. It was recently demonstrated that the observation of both a novice and an expert model (mixed observation) resulted in better learning of a complex spatiotemporal task than the observation of either a novice or an expert model. In experiment 1, we aimed to determine whether mixed observation better promotes learning due to the information that can be gained from two models who exhibit different skill levels or simply because multiple models, regardless of their level of expertise, better promote learning than would a single model. The results revealed that the observation of both an expert and a novice model resulted in better short-term retention than the observation of either two novice or two expert models. In experiment 2, we wanted to determine whether these benefits would last longer if physical practice trials were interspersed with observation. Mixed and (to some extent) expert observations resulted in better long-term retention than observation of a novice model. We suggest that alternating mixed/expert observation with physical practice trials makes one's error more salient than when all observation trials are completed before one first starts performing the experimental task, which increases activation of the action observation network.

Success modulates consolidation of a visuomotor adaptation task.

Consolidation is a time-dependent process that is responsible for the storage of information in long-term memory. As such, it plays a crucial role in motor learning. Prior research suggests that some consolidation processes are triggered only when the learner experiences some success during practice. In the present study, we tested whether consolidation processes depend on the objective performance of the learner or on the learner's subjective evaluation of his or her own performance (i.e., how successful the learner believes he or she is). Four groups of participants performed 2 sessions of a visuomotor adaptation task for which they had to learn a new internal model of limb kinematics; these sessions were either 5 min or 24 hr apart. The task was identical for all participants, but each group was given a difficult or an easy objective that affected the participants' evaluation of their own performance during the initial practice session. All groups adapted their movements similarly to the rotation of the visual feedback during the first session. However, when retested the following day, participants who had a 24-hr rest interval and had initially experienced success performed significantly better than those who did not feel successful or who were given a 5-min rest interval. Our results indicate that a certain level of subjective success must be experienced to trigger certain consolidation processes.

Learning through observation: a combination of expert and novice models favors learning.

Observation of an expert or novice model promotes the learning of a motor skill. In two experiments, we determined the effects of a mixed observation schedule (a combination of expert and novice models) on the learning of a sequential timing task. In Experiment 1, participants observed a novice, expert, or both novice and expert models. The results of retention/transfer tests revealed that all observation groups and a physical practice group learned the task and outperformed a control group. However, observing a novice model was not as effective as observing expert and mixed models. Importantly, a mixed schedule of novice and expert observation resulted in a more stable movement time and better generalization of the imposed relative timing pattern than observation of either a novice or expert model alone. In Experiment 2, we aimed to determine whether a certain type of novice performance (highly variable, with or without error reduction with practice) in a mixed observation schedule would improved motor learning. The observation groups performed as well as a physical practice group and significantly better than a control group. No significant difference was observed with the type of novice model used in a mixed schedule of observation. The results suggest that mixed observation provides an accurate template of the movement (expert observation) that is enhanced when contrasted with the performance of less successful models.

Effects of the model's handedness and observer's viewpoint on observational learning.

Observation promotes motor skill learning. However, little is known about the type of model and conditions of observation that can optimize learning. In this study, we investigated the effects of the model's handedness and the observer's viewpoint on the learning of a complex spatiotemporal task. Four groups of right-handed participants observed, from either a first- or third-person viewpoint, right- or left-handed models performing the task. Observation resulted in significant learning. More importantly, observation of same-handed models resulted in improved learning as compared with observation of opposite-handed models, regardless of the observer's viewpoint. This suggests that the action observation network (AON) is more sensitive to the model's handedness than to the observer's viewpoint. Our results are consistent with recent studies that suggest that the AON is linked to or involves sensorimotor regions of the brain that simulate motor programming as if the observed movement was performed with one's own dominant hand.

Congruent visual and proprioceptive information results in a better encoding of initial hand position.

Goal-directed movements performed in a virtual environment pose serious challenges to the central nervous system because the visual and proprioceptive representations of one's hand position are not perfectly congruent. The aim of the present study was to determine whether the vision of one's hand or upper arm, compared with that of a cursor representing the tips of one's index finger and thumb, optimizes the planning and modulation of one's movement as the cursor nears the target. The participants performed manual aiming movements that differed by the source of static visual information available during movement planning and the source of dynamic information available during movement execution. The results revealed that the vision of one's hand during the movement planning phase results in more efficient online control processes than when the movement planning was based on a virtual representation of one's initial hand location. This observation was seen regardless of the availability of online visual feedback during movement execution. These results suggest that a more reliable estimation of the initial hand position results in more accurate estimation of the position of the cursor/hand at any one time resulting in more accurate online control.

Observation learning versus physical practice leads to different consolidation outcomes in a movement timing task.

Motor learning is a process that extends beyond training sessions. Specifically, physical practice triggers a series of physiological changes in the CNS that are regrouped under the term "consolidation" (Stickgold and Walker 2007). These changes can result in between-session improvement or performance stabilization (Walker 2005). In a series of three experiments, we tested whether consolidation also occurs following observation. In Experiment 1, participants observed an expert model perform a sequence of arm movements. Although we found evidence of observation learning, no significant difference was revealed between participants asked to reproduce the observed sequence either 5 min or 24 h later (no between-session improvement). In Experiment 2, two groups of participants observed an expert model perform two distinct movement sequences (A and B) either 10 min or 8 h apart; participants then physically performed both sequences after a 24-h break. Participants in the 8-h group performed Sequence B less accurately compared to participants in the 5-min group, suggesting that the memory representation of the first sequence had been stabilized and that it interfered with the learning of the second sequence. Finally, in Experiment 3, the initial observation phase was replaced by a physical practice phase. In contrast with the results of Experiment 2, participants in the 8-h group performed Sequence B significantly more accurately compared to participants in the 5-min group. Together, our results suggest that the memory representation of a skill learned through observation undergoes consolidation. However, consolidation of an observed motor skill leads to distinct behavioural outcomes in comparison with physical practice.

Suboptimal online control of aiming movements in virtual contexts.

We determined whether uncertainty about the location of one's hand in virtual environments limits the efficacy of online control processes. In the Non-aligned and Aligned conditions, the participant's hand was represented by a cursor on a vertical or horizontal display, respectively. In the Natural condition, participants saw their hand. During an acquisition phase, visual feedback was either permitted or not during movement execution. To test the hypothesis (Norris et al. 2001) that reliance on visual feedback increases as the task becomes less natural (Natural < Aligned < Non-aligned), following acquisition, participants performed a transfer phase without visual feedback. During acquisition in both visual feedback conditions, movement endpoint variability increased as the task became less natural. This suggests that the orientation of the display and the representation of one's hand by a cursor introduced uncertainty about its location, which limits the efficacy of online control processes. In contradiction with the hypothesis of Norris et al. (2001), withdrawing visual feedback in transfer had a larger deleterious effect on movement accuracy as the task became less natural. This suggests that the CNS increases the weight attributed to the input that can be processed without first having to be transformed.

Automatic movement error detection and correction processes in reaching movements.

Manual aiming movements can be amended during their execution. Recent evidence suggests that error detection and correction are based on automatic and even reflexive processing of afferent information. In this study, we wanted to determine whether these processes are affected by the occurrence of successive events requiring adjustments of the originally planned movement. To reach our goal, we used a video-aiming task. For a small proportion of the trials, the cursor moved by the participant was translated laterally by 15 mm (cursor jump) soon after movement initiation. For some of the cursor-jump trials, a second cursor jump occurred 100 ms after the first one and canceled or doubled the initial cursor translation. Results showed that participants were able to cancel or double the size of the correction in response to the second cursor jump. More importantly, in double-jump trials, the correction latency for the first and second cursor jumps did not differ from that of single-jump trials. Moreover, the correction for the second cursor jump blended seamlessly with the correction for the first cursor jump. These observations suggest that the processes leading of a correction for a cursor jump do not interfere with incoming visual information.

Automaticity of online control processes in manual aiming.

Experiments that manipulated the visual feedback of the moving limb have suggested the existence of efficient and automatic online correction processes. We wanted to determine whether the latency/gain of the correction for a cursor jump are only influenced by the size of the cursor jump or whether they are also influenced by the need of a correction for the target to be reached. In Experiment 1, we used two target sizes (5 and 30 mm) and three cursor-jump amplitudes (5, 15, and 25 mm), so that for some target size/cursor-jump combinations, no correction would be needed to reach the target. Participants were not aware of the cursor jump, but we observed a 65% correction regardless of target size. In Experiment 2, participants pointed at a large target for which a 15-mm cursor jump never impeded target attainment. Participants modified the trajectory of their movement in the direction opposite to the cursor jump (42% of the cursor jump). Our results indicate that the latency of the correction for a cursor jump was not influenced by the size of the cursor jump or that of the target. However, the correction tailored the movement's initial impulse according to the target's characteristics.

Developmental aspects of pluriarticular movement control.

Precise pluriarticular movement control is required to perform straight and smooth out-and-back movements. Our goal was to determine whether children perform out-and-back movements as accurately as adults do in the presence and absence of visual feedback. To reach our goal, 36 children aged between 6 and 12 years, and 12 young adults, performed an out-and-back movement in a normal-vision condition and in a target-only condition. Reversal angle and overlapping error were taken to represent the ability of children to control pluriarticular movement. The results showed that adults exhibited sharper movement reversal than the three children groups did, but only for eccentric targets relative to their midline. This suggests that pluriarticular movement control improved across the course of development for eccentric regions of the workspace. Visual feedback did not result in sharper movement reversal even when relatively large errors were noted (eccentric targets in children). This underlines the relatively minor role of visual feedback for interjoint coordination when proprioception is intact. Finally, we observed that directional variability was smaller at the 100-ms mark for the back than for the out portion of the movement, suggesting that movement-planning processes appear less variable when based on dynamic rather than static afferent information.

Distinct consolidation outcomes in a visuomotor adaptation task: Off-line leaning and persistent after-effect.

Consolidation is a time-dependent process responsible for the storage of information in long-term memory. As such, it plays a crucial role in motor learning. In two experiments, we sought to determine whether one's performance influences the outcome of the consolidation process. We used a visuomotor adaptation task in which the cursor moved by the participants was rotated 30 degrees clockwise. Thus, participants had to learn a new internal model to compensate for the rotation of the visual feedback. The results indicated that when participants showed good adaptation in the first session, consolidation resulted in a persistent after-effect in a no-rotation transfer test; they had difficulty returning to their normal no-rotation internal model. However, when participants showed poor adaptation in the first session, consolidation led to significant off-line learning (between sessions improvement) but labile after-effects. These observations suggest that distinct consolidation outcomes (off-line learning and persistent after-effect) may occur depending on the learner's initial performance.

Factors influencing online control of video-aiming movements performed without vision of the cursor.

A modulation of the primary impulse of manual/video-aiming movements performed without visual feedback has been reported. In the present study, we show that this modulation is modified (a) with increased practice, (b) the use of an aligned visual display, and (c) the availability of visual feedback on alternated trials. However, this modulation was not as efficient as that observed in a normal vision condition, which underlines the primary role of vision to ensure endpoint accuracy. Moreover, this modulation was observed only on the extent component of the task. This last observation indicates that proprioception can be used to modulate the extent component of goal-directed movements but that vision is necessary to modulate their direction.

Evidence for continuous processing of visual information in a manual video-aiming task.

Research shows that individuals are able to correct for an experimentally-induced and unexpected aiming error (i.e., a cursor jump), even when they do not detect it consciously. Researchers have interpreted these results to be evidence of continuous processing of visual afferent information for movement control. The authors conducted 2 experiments to determine whether they would gain additional support for this proposition by showing that correction for a cursor jump can be initiated outside the central visual field. In addition, the authors wanted to determine whether the normally occurring modulation of the ongoing movement is affected by detection and correction of the cursor jump. Participants performed video-aiming movements in which a 30-mm cursor jump occurred in a small proportion of the trials. The results indicate that correction for the cursor jump was initiated when the cursor was as far as 15 degrees of visual angle from the target. In addition, the authors observed accurate corrections when vision of the cursor was withdrawn soon after the cursor jump. Last, online control processes reducing initial movement variability were not significantly affected by the detection and correction for the cursor jump. The results suggest near continuous monitoring of visual afferent information but a more discrete movement-correction process.