Influence of Advance Information About Target Location and Visual Feedback on Movement Planning and Execution, The
This study was designed to determine if movement planning strategies incorporating the use of visual feedback during manual aiming are specific to individual movements. Advance information about target location and visual context was manipulated using precues. Participants exhibited a shorter reaction time and a longer movement time when they were certain of the target location and that vision would be available. The longer movement time was associated with greater time after peak velocity. Under conditions of uncertainty, participants prepared for the worstcase scenario. That is, they spent more time organizing their movements and produced trajectories that would be expected from greater open-loop control. Our results are consistent with hierarchical movement planning in which knowledge of the movement goal is an essential ingredient of visual feedback utilization.
Résumé Cette étude a été conçue pour déterminer si les stratégies de planification du mouvement intégrant l’utilisation de la rétroaction visuelle au cours d’un ciblage manuel sont particulières aux mouvements individuels. L’information anticipée au sujet de l’emplacement de la cible et du contexte visuel a été manipulée à l’aide de préindices. Les participants ont manifesté un temps de réaction plus court et un temps de mouvement plus long lorsqu’ils étaient certains de l’emplacement d’une cible et que la vision était possible. Le délai de mouvement le plus long a été associé à un délai plus grand après la vélocité de pointe. Dans des conditions d’incertitude, les participants se préparaient au scénario du pire cas, c’est-à-dire, qu’ils passaient plus de temps à organiser leurs mouvements et à produire des trajectoires propres à vu plus grand contrôle en boucle ouverte. Nos résultats concordent avec la planification du mouvement hiérarchique où la connaissance du but du mouvement est un élément essentiel de l’utilisation de la rétroaction visuelle.
More than half a century ago, Hick (1952) and Hyman (1953) independently demonstrated a linear increase in the time to initiate a movement that depended on the number of stimulus-response alternatives. The increase in reaction time (RT) was reliably predicted using a binary decision-making equation incorporating the logarithm to base-2 (log^sub 2^) of the number of stimulus-response alternatives. This relation between RT and the amount of information to be processed has held up well for simple key presses or releases involving one or more digits. The relation between RT and number of stimulus-response alternatives has come to be known as Hick’s Law or the HickHyman Law.
However, the situation becomes more complicated when an actual movement response includes moving to the location of the stimulus because it takes more time to organize complex movements than simple movements (Henry & Rogers, 1960; Sternberg, Monsell, Knoll, & Wright, 1978). Although movement complexity has a reliable impact on RT under simple RT conditions, it has an even greater impact on the time to initiate a response when the specific stimulus-response requirements are not known until the imperative stimulus occurs (i.e., choice RT conditions; Klapp & Erwin, 1976). Presumably, this occurs because under choice conditions the specific movement must be completely organized during the RT interval (Klapp, 1977) while in a simple RT situation, participants have an opportunity to “preprogram” at least some of the action prior to the movement imperative.
Although advance information that reduces the uncertainty about the required response generally leads to a reduction in RT, Rosenbaum (1980) suggested that the degree of benefit also depends on the specific movement parameters known prior to the imperative stimulus. This proposal is based on the notion that movements are hierarchically organized. Specifically, Rosenbaum hypothesized that shorter RT occurred when the effector (e.g., hand) was specified before movement direction (e.g., forward or backward) and when the programming of direction preceded movement amplitude (e.g., distance – short or long). Rosenbaum (1980) found support for this position in a series of studies using a “precue technique” in which, on a given trial, advance information (a precue) was provided about specific parameters of the upcoming movement. In his experiments, he found that knowing only movement amplitude was superior to having no advance information at all. However, greater reductions in RT were seen when the precue specified the arm required for the movement compared to precuing movement direction. Knowing movement direction favoured motor preparation compared to knowing movement extent. Consistent with the notion of hierarchical movement organization, Anson, Hyland, Kotter, and Wickens (2000) used a single limb task requiring pronation or supination to a near or far target and demonstrated that RT was significantly reduced when direction (pronation or supination) was specified compared to extent (near or far), and RT for both were shorter compared to the ambiguous precues involving direction and extent and the same number (n = 2) of alternatives (Mieschke, Elliott, Helsen, Carson, & Coull, et al., 2001; cf. Goodman & Kelso, 1980).
Although the majority of studies using the precue technique have concentrated on how advance information influences RT, prior knowledge about an upcoming movement can also influence the manner in which it is executed. For example, Mieschke et al. (2001) found that peak velocity was higher and the time to reach peak velocity was reduced when both direction and extent were specified in advance of executing rapid aiming movements. The time to reach peak velocity results can be taken as evidence that these aiming movements underwent greater preprogramming relative to movements involving a greater degree of advance uncertainty (see also Olivier & Bard, 2000 for similar results with children). Overall the findings from precue studies indicate that reducing the uncertainty about an upcoming movement enables the motor system to engage in more rapid and effective advance planning and execution.
Traditionally, the knowledge provided by the precue has consisted of information about the parameters of the movement to be prepared (e.g., movement direction, movement extent, and effector). There are also well-established differences in how movements are planned and executed that depend on the expected sensory context under which the movement will be performed. Not surprisingly, movements are performed more consistently and accurately when visual feedback of the limb and target are available during planning (Elliott & Madalena, 1987) and execution (Keele & Posner, 1968; Woodworth, 1899). For rapid movements, in which the limits of visual processing time (i.e., approximately 120 ms) are approached, the accuracy benefits associated with visual feedback depend on whether or not the performer knows that feedback will be available (Elliott & Allard, 1985; Zelaznik, Hawkins, & Kisselburgh, 1983). Also, preparation time and movement kinematics also differ depending on advance knowledge of the visual condition. Specifically, participants spend less time preparing movements (i.e., shorter RT) if they know vision will be available throughout the movement compared to conditions in which vision is removed at movement initiation (Khan, Elliott, Coull, Chua, & Lyons, 2002). Although RT is typically shorter, movement time (MT) is often longer when vision is available (e.g., Chua & Elliott, 1993). Moreover, participants spend a greater proportion of the overall MT after peak deceleration (Khan et al., 2002) where visual information is generally considered to be important for “homing-in” on the target (Elliott, Helsen, & Chua, 2001; Woodworth, 1899). Thus, participants use different strategies to control movements depending on their prior knowledge about the availability of visual feedback (Ricker et al., 1999). Elliott, Hansen, Mendoza, and Tremblay (2004), and Zelaznik and colleagues (1983) have suggested that when participants are unsure about the availability of vision for the upcoming trial, they prepare for the “worst-case” scenario. Specifically, they assume that vision will not be available and engage in a more feedforward mode of control. This type of control is characterized by symmetric velocity and acceleration profiles with fewer corrective submovements between peak velocity and the end of the movement (see Elliott, Binsted, & Heath, 1999 for a review). The idea of preparing for the “worst-case” scenario is similar to the “default” preparation described by Ghez, Hening, and Favilla (1990) when participants are forced to prepare to initiate a response to a stimulus that occurs just 100 ms after a precue.
For the present study, we independently manipulated advance information about the direction of a single hand (right) aiming movement (to a right or left target) and the availability of vision while executing the movement. We were interested in how information about the visual context for the upcoming trial influenced preparation and execution when participants did or did not know the direction the movement. Under conditions in which the target is known in advance of the imperative stimulus, we expected to find typical vision condition effects. That is, participants were predicted to adopt an advance planning strategy when they were cued that vision would not be available, or that vision would be equally likely to be present throughout, or removed during movement execution. It was predicted that this strategy would be characterized by longer RT, shorter MT, and more symmetric kinematic profiles than in conditions in which participants were informed prior to movement initiation about the availability of vision for movement execution. Our RT predictions are consistent with findings that indicate that participants take longer to prepare movements under no-vision conditions and under conditions of visual feedback uncertainty even when the target location is known in advance (e.g., Khan et al., 2002). More precise advance planning would also be expected to result in shorter MT because participants would spend less time engaged in online control.
Of special interest were trials in which the target was unknown (i.e., choice RT situation). In these conditions, optimal advance planning was not possible and the specification of movement direction could not take place until the illumination of the target (the imperative stimulus and the direction to move) signalled the participant to move (i.e., specification of direction and movement initiation had to occur during the RT interval). Moreover, movement planning is hierarchical (i.e., from what to how), preplanning associated with the use of vision feedback may not be helpful, because participants would not know the required direction in advance of the movement imperative. One possibility is that on no-vision trials and on trials of visual feedback uncertainty, RT associated with movement planning would be longer than under target-known conditions. This prediction is consistent with the idea that preparation associated with sensory feedback utilization and/or feedforward control is only possible once the movement goal has been specified. That is, goal selection should precede movement specification in a hierarchical movement planning process. Alternatively, goal selection and the associated muscle specification may proceed in parallel with planning for the impending mode of control (feedforward or feedback-based). In this case, target knowledge and feedback knowledge would have independent effects on RT and would not interact.
Twelve right-handed students from Simon Fraser University (eight females) with normal or corrected-to-normal vision (n = 7) completed the experiment. The data from one participant were removed due to equipment malfunction during collection. Ages ranged from 20 to 26 years (M = 22.8, SD = 1.85). The procedures were approved by the university ethics board and were conducted in accordance with the Declaration of Helsinki (1964). All participants gave informed consent prior to participation and received financial compensation for their time.
Participants performed 36.5-cm manual aiming movemets to two targets that were rear projected onto a Plexiglas screen with a Sanyo PROxtra projector. The projector was angled 38° down and the images were reflected by a mirror onto a screen that was angled at 54° from the table with a total projection distance of 112 cm. Participants sat in front of the screen holding a 12-cm stylus with an infrared emitting diode attached to the end. The outline of two targets was projected onto the screen for the duration of each trial. The targets were 36.5 cm from a home switch, 25 cm apart, and had a diameter of 2.5 cm. Once the participant was ready for the next trial, the outline of the two targets was presented for 500 ms, followed by a 500-ms precue that consisted of one or both targets changing colour. The colour of the precue denoted whether vision would be occluded during an aiming movement with a pair of liquid crystal goggles. Specifically, if the precue was green, participants knew vision would be available throughout their movement from the home switch to the centre of the target. However, if the precue was red, they knew vision would be occluded during the movement, and if the precue was yellow, they knew the probability that vision would be available/not available during the movement was equal. The number of targets that changed colour during the precue denoted potential targets. The precue was always predictive, therefore if one target changed colour, participants knew that would be the target (SRT situation), whereas if both targets changed colour, there was equal probability of either one being the target (CRT situation). Participants were informed about what the precue specified and were instructed to use this information to perform their movements as rapidly and accurately as possible. After the precue, one of the two targets turned white until participants began their movement (i.e., released the home switch). One target turning white was the imperative signal for participants to perform an aiming movement with the stylus to the centre of the target and for the PhoeniX Visualeyez(TM) to record the three-dimensional position at 200Hz for 2 s. When participants initiated their movement, they released the home switch (diameter = 1.5 cm). On no-vision trials, release of the home switch triggered the liquid crystal goggles to close. Custom software was designed with E-Prime® (Version 1.1a: Psychology Software Tools Inc.) to control the presentation of the stimuli, to synchronize the presentation of the imperative stimulus with the initiation of movement recording, and to record RT through a switch connected to the parallel port.
Each trial began when the participant depressed the home switch with the stylus. The outline of the two targets appeared for 500 ms, followed by the precue for 500 ms. The precue was followed by a random foreperiod (1,000 to 1,400 ms) after which one of the targets turned white until the participant released the home switch. The illumination of the target was the imperative signal for participants to execute an aiming movement with their dominant hand to the centre of the target. During no-vision trials, the liquid crystal goggles closed once the participant released the home switch for the remaining duration of the 2 s PhoeniX recording (i.e., movement execution only). Participants were also instructed to hold the stylus on the target until the outline of targets disappeared, or in the case of a no vision trial, the goggles reopened. The 16 trial types (2 Vision Knowledge by 2 Vision Conditions by 2 Target Knowledge by 2 Target Location) were randomized in 10 blocks for a total of 160 trials.
Reaction time (RT), movement time (MT), and kinematic variables were derived from the three-dimensional displacement data collected with the PhoeniX system. Specifically, resultant displacement data were filtered using a five-point running average. The filtered data were then differentiated to obtain velocity. Custom-programs were used to identify peak velocity as well as the beginning and end of the movement. A minimum velocity of 10 mm/s lasting for a minimum of 70 ms was used to identify the beginning and the end of each aiming movement.
All dependent measures were analyzed using separate 2 (Vision Knowledge: certain, uncertain) × 2 (Vision Condition: full vision, no vision) × 2 (Target Knowledge: simple, choice) ? 2 (Target Location: left, right) repeated measures analysis of variance. It should be noted that, for the sake of consistency, we chose to use this factorial arrangement for all dependent variables. We did this in spite of the fact that under vision – uncertain conditions, the reaction time interval was over before the participant discovered the vision condition for that trial.
Tukey’s HSD (p
The RT analysis yielded a main effect for Target Knowledge, F(1,10) = 6.53, p
The MT analysis revealed a main effect for Target Location, F(1,10) = 36.40, p
The radial error analysis revealed only a main effect for Vision Condition, F(1,10) = 29.77, p
In order to determine the locus of our MT effects, we partitioned MT into the time required to achieve peak velocity and the time between peak velocity and the termination of the movement. The former portion of the movement trajectory is typically associated with movement planning while the time after peak velocity is more associated with the time necessary for online control (e.g., Elliott et al., 2001). The time to peak velocity analysis revealed a main effect for Target Knowledge, F(1,10) = 7.55, p
In this study we used a variation of the precue technique (Rosenbaum, 1980) to explore how manual aiming movements are prepared and executed when advance information about the availability of vision and target location is manipulated. The RT, MT, and kinematic findings are all consistent with a fixed order model of movement preparation in which the spatial goal of the movement must be specified before specific planning for visual feedback to be used to the full extent. In the case where target location was known in advance (i.e., simple situation) with the knowledge that vision would be available, RT was shorter and MT was longer with no apparent difference in the overall time required to initiate and execute the movement. Longer MT under conditions in which participants knew they would receive visual feedback when aiming to a specific target was due primarily to the extra time spent after peak velocity while decelerating the movement. Our vision-no-vision results under simple RT conditions are similar to those of Khan et al. (2002) who also found that participants took less time to prepare their movements but more time to execute their movements when they knew that vision would be available for online regulation of the limb. The current work extends this finding by demonstrating that knowing about the availability of visual feedback has a limited impact on performance when the target location is unknown (i.e., choice RT conditions). That is, in the vision-known, target-unknown (choice) conditions, whether full vision or no vision was available had no significant effect on RT or MT or time to peak velocity or time after peak velocity. These findings suggest that participants prepared their responses in a similar manner and there appeared to be no RT advantage to knowing that full vision was available. Participants required knowledge of the target location before stimulus onset in order for the knowledge of availability of vision to be useful for movement preparation. Thus the relationship between preparatory processes and visual feedback utilization appear to benefit movement preparation only when the entire nature of the response is precued prior to stimulus onset.
Other studies concerned with use of advance information about the availability of visual information have found that participants are also more accurate (Elliott & Allard, 1985) and more consistent (Zelaznik et al., 1983) in their manual aiming if they know in advance that vision will be available. In this study, however, the radial error analysis revealed only an overall effect for the availability of vision. This suggests that even when participants prepared themselves for the worse case scenario, they were still able to use visual information when it was available to reduce aiming error. There are two relevant differences between the work reported here and previous research (i.e., Elliott & Allard, 1985; Zelaznik et al., 1983). Firstly, MT was quite long in this study compared to much of the earlier research on vision and upper limb control (see Elliott et al., 2001 for a review). Presumably, this extra time allowed participants to benefit from vision even under conditions in which they were not depending on its availability. As well, we treated movement time as a dependent variable rather than an independent variable (cf. Elliott & Allard, 1985; Keele & Posner, 1968; Zelaznik et al., 1983). By allowing movement time and error to covary, we were able to see how participants chose to trade off speed and accuracy under different feedback and advance information conditions. In this respect, our protocol was more like Khan et al. (2002). In the Khan et al. experiments and the work reported here, participants exhibited movement times that were well above current estimates of visual and proprioceptive processing time (Carlton, 1992). Unlike some of the previous studies that used a blocked versus random protocol for ordering vision and no-vision trials (e.g., Elliott & Allard, 1985; Zelaznik et al., 1983), the feedback schedule in the present work was always random with advance information being provided about the availability of visual feedback on only half the trials. This scheduling protocol ensures that any benefits associated with vision are not due to the use of visual information from the preceding trial, but rather on the processing of visual feedback while the limb is in flight (Khan et al., 2002). In other words, when a series of full vision trials are completed in a blocked manner, it is possible for the performer to use the visual information from the previous trial to improve his/her performance on the next trial. In contrast, during a random paradigm, visual information from the previous trial will not necessarily be advantageous to the performer if the next trial is completed under a different visual condition.
Overall, our results are consistent with a hierarchical model of movement organization in which knowing the constraints on performance prior to an imperative stimulus facilitates specific planning about the best way to achieve that goal and results in shorter motor preparation time. Although it is clear that both feedforward and online processes are important for the regulation of goal-directed aiming (Elliott et al., 2001; Woodworth, 1899), our findings indicate that the relative importance of these two complementary processes depends on prior knowledge about the potential quality of the feedback for online regulation. When the performer is certain that vision of the hand and target will be available during movement execution, a movement is directed toward the general target area and then visual feedback is used in later portions of the movement to bring the limb onto the precise target location. This is reflected in the time after peak velocity. Although endpoint accuracy was still quite good when the performer was provided with visual feedback under vision unknown conditions, RT and kinematic data indicate that the movement was prepared and executed quite differently. Under these conditions, it would appear that the preferred strategy was to engage in more precise advance planning. Presumably under these conditions using vision or depending on vision for successful performance would not be part of the initial plan. However, as the radial error results indicate, accuracy benefits are derived from the availability of vision regardless of whether or not participants planned in advance to use it. Certainly, other work indicates that this degree of flexibility would not be possible under more temporally constrained conditions (see Carlton, 1992 for a review).
Our findings that goal selection precedes the specification of the specific online control processes in movement planning should not be surprising. Indeed, the posterior parietal areas of the cortex responsible for the online visual regulation of goal-directed movements (Desmurget et al., 1999) can only be prepared for the expected sensory consequences of the upcoming movement once the more frontal movement planning areas of the brain have developed a general blueprint for the upcoming movement (Deiber, Ibanez, Sadato, & Hallet, 1996). This feedforward information about the movement plan has been generally referred to as efference copy and is thought to serve as reference against which the sensory consequences of a movement can be evaluated for online error reduction (von Holst, 1954). In the context of the work reported here, it would appear that dependent upon the expected sensory (i.e., visual) consequences of the goal-directed movement, a performer structures a movement plan that relies more or less on this reafferent process. The mere knowledge that vision will be available is not sufficient to provide a motor preparation advantage unless the nature (target location, response direction, and movement amplitude) of the response is precued before the onset of the imperative stimulus.
Anson, J. G., Hyland, B. L, Kotier, R., & Wickens, J. R. (2000). Parameter precuing and motor preparation. Motor Control, 4, 221-231.
Chua, R., & Elliott, D. (1993). Visual regulation of manual aiming. Human Movement Science, 12, 365-401.
Deiber, M-P., Ibanez, V., Sadato, N., & Hallet, M. (1996). Cerebral structures participating in motor preparation in humans: A positron emission tomography study. Journal of Neuropbysiology, 75, 233-247.
Desmurget, M., Epstein, C. M., Turner, R. S., Prablanc, C., Alexander, G. E., & Grafton, S. T. (1999). Role of the posterior parietal cortex in updating reaching movements to visual targets. Nature Neuroscience, 2, 563-567.
Elliott, D., & Allard, F. (1985). The utilization of visual feedback information during rapid pointing movements. The Quarterly Journal of Experimental Psychology, 37, 407-425.
Elliott, D., Binsted, G., & Heath, M. (1999). The control of goal directed aiming movements: Correcting errors in the trajectory. Human Movement Science, 18, 121-136.
Elliott, D., Hansen, S., Mendoza, J., & Tremblay, L. (2004). Learning to optimize speed, accuracy, and energy expenditure: A framework for understanding speedaccuracy relations in goal-directed aiming. Journal of Motor Behavior, 36, 339-351.
Elliott, D., Helsen, W. F., & Chua, R. (2001). A century later: Woodworth’s (1899) two-component model of goal directed aiming. Psychological Bulletin, 3, 342-357.
Elliott, D., & Madalena, J. (1987). The influence of premovement visual information on manual aiming. The Quarterly Journal of Experimental Psychology, 39, 541-559.
Erlhagen, W., & Schoner, G. (2002). Dynamic field theory of movement preparation. Psychological Review, 109, 545-572.
Ghez, C., Hening, W., & Favilla, M. (1990). Parallel interacting channels in the initiation and specification of motor response features. In M. Jeannerod (Ed.), Attention and performance XIII: Motor representation and control (pp. 265-293). Hillsdale, NJ: Lawrence Erlbaum Associates.
Goodman, D., & Kelso, J.A.S. (1980). Are movements prepared in parts? Not under compatible (naturalized) conditions. Journal of Experimental Psychology: General, 109, 475-495.
Henry, F. M., & Rogers, D. E. (1960). Increased response latency for complicated movements a “memory drum” theory of neuromotor reaction. Research Quarterly for Exercise and Sport, 31, 448-458.
Hick, W. E. (1952). On the rate of gain of information. Quarterly Journal of Experimental Psychology, 4, 11-26.
Hyman, R. (1953). Stimulus information as a determinant of reaction time. Journal of Experimental Psychology, 45, 188-196.
Keele, S. W., & Posner, M. I. (1968). Processing of visual feedback in rapid movements. Journal of Experimental Psychology, 77, 155-158.
Khan, M. A., Elliott, D., Coull, J., Chua, R., & Lyons, J. (2002). Optimal control strategies under different feedback schedules: Kinematic evidence. Journal of Motor Behavior, 34, 45-57.
Klapp, S. T. (1977). Reaction time analysis of programmed control. Exercise and Sport Science Reviews, 5, 231-253.
Klapp, S. T., & Erwin, C. I. (1976). Relation between programming time and duration of the response being programmed. Journal of Experimental Psychology: Human Perception and Performance, 2, 591-598.
Mieschke, P. E., Elliott, D., Helsen, W. F., Carson, R. G., & Coull, J. A. (2001). Manual asymmetries in the preparation and control of goal-directed movements. Brain & Cognition, 45, 129-140.
Olivier, I., & Bard, C. (2000). The effects of spatial movement component precues on the execution of rapid aiming in children aged 7, 9 and 11. Journal of Experimental Child Psychology, 77, 155-168.
Ricker, K. L., Elliott, D., Lyons, J., Gaulbie, D., Chua, R., & Byblow, W. (1999). The utilization of visual information in the control of sequential aiming movements. Acta Psychologica, 103, 103-123.
Rosenbaum, D. A. (1980). Human movement initiation: Specification of arm, direction, and extent. Journal of Experimental Psychology: General, 109, 444-474.
Sternberg, S., Monsell, S., Knoll, R. L, & Wright, C. E. (1978). The latency and duration of rapid sequences: Comparisons of speech and typewriting. In G. E. Stelmach (Ed.), Information processing in motor control and learning. New York: Academic Press.
von Hoist, E. (1954). Relations between the central nervous system and the peripheral organs. British Journal of Animal Behavior, 2, 89-94.
Woodworth, R. S. (1899). The accuracy of voluntary movement. Psychological Review, 3, (monograph supplement), 1-119.
Zelaznik, H. N., Hawkins, B, & Kisselburgh, L. (1983). Rapid visual feedback processing in single-aiming movements. Journal of Motor Behavior, 15, 217-236.
Steve Hansen and Cheryl M. Glazebrook, McMaster University
J. Greg Anson, University of Otago
Daniel J. Weeks, Simon Fraser University
Digby Elliott, McMaster University
The Natural Science and Engineering Research Council of Canada funded our research.
All correspondence should be addressed to Steve Hansen, Department of Kinesiology, McMaster University, Hamilton, Ontario L8S 4Kl (Phone: 905-525-9140, Ext. 27390; Email: email@example.com).
Copyright Canadian Psychological Association Sep 2006
Provided by ProQuest Information and Learning Company. All rights Reserved