Elsevier

Acta Psychologica

Volume 108, Issue 3, December 2001, Pages 219-245
Acta Psychologica

Peripheral constraint versus on-line programming in rapid aimed sequential movements

https://doi.org/10.1016/S0001-6918(01)00038-5Get rights and content

Abstract

The purpose of this investigation was to examine how the programming and control of a rapid aiming sequence shifts with increased complexity. One objective was to determine if a preprogramming/peripheral constraint explanation is adequate to characterize control of an increasingly complex rapid aiming sequence, and if not, at what point on-line programming better accounts for the data. A second objective was to examine when on-line programming occurs. Three experiments were conducted in which complexity was manipulated by increasing the number of targets from 1 to 11. Initiation- and execution-timing patterns, probe reaction time (RT), and movement kinematics were measured. Results supported the peripheral constraint/pre-programming explanation for sequences up to 7 targets if they were executed in a blocked fashion. For sequences executed in a random fashion (one length followed by a different length), preprogramming did not increase with complexity, and on-line programming occurred without time cost. Across all sequences there was evidence that the later targets created a peripheral constraint on movements to previous targets. We suggest that programming is influenced by two factors: the overall spatial trajectory, which is consistent with Sidaway's subtended angle (SA) hypothesis (1991), and average velocity, with the latter established based on the number of targets in the sequence. As the number of targets increases, average velocity decreases, which controls variability of error in the extent of each movement segment. Overall the data support a continuous model of processing, one in which programming and execution co-occur, and can do so without time cost.

Introduction

Sequential movements, such as tying shoelaces, talking, retrieving and placing objects, buttoning clothes, writing and typing are common motor behaviors. Work by Henry and Rogers (1960) and many others indicate that as sequential movements increase in complexity, the time needed to organize and initiate the movement also increases (see Christina, 1992 for a review; Garcia-Colera & Semjen, 1987; Gordon & Meyer, 1987; Lajoie & Franks, 1997). Most of these results are discussed in terms of a discrete model, in which all programming for the sequence is assumed to occur before movement is initiated. Three examples of discrete models are Henry's Memory Drum Theory (Henry & Rogers, 1960; Henry, 1980), Sternberg's sequence-preparation model (Sternberg, Monsell, & Wright, 1978), and Rosenbaum's Hierarchical Editor Model (Rosenbaum, Inhoff, & Gordon, 1984).

Other research, however, indicates that an on-line programming model, one in which programming occurs before and during execution, is necessary to describe the data from certain types of movement sequences. Rosenbaum altered his Hierarchical Editor Model to account for programming that occurs after movement initiation when longer sequences are required (Rosenbaum, Hindorff, & Munro, 1987). An on-line model was also supported in executing a well-learned handwriting sequence (Portier & van Galen, 1992; Portier, van Galen, & Meulenbroek, 1990). A unique element within a sequence can also cause programming and execution to co-occur (Garcia-Colera & Semjen, 1988; Smiley-Oyen & Worringham, 1996), as can a required pause longer than 200 ms (Ketelaars, Garry, & Franks, 1997).

The purpose of this investigation was to examine how the programming and control of a rapid aiming sequence shifts with increased complexity. One objective was to determine if a preprogramming/peripheral constraint explanation is adequate to characterize control of an increasingly complex rapid aiming sequence, and if not, at what point on-line programming better accounts for the data. A second objective was to examine when on-line programming occurs. Previous data suggested that programming “spills over” into the first segment (as evidenced by a longer movement time to the first target), although that interpretation is controversial. The alternative explanation is that movement time to the first target is longer because of the peripheral constraint placed on the movement by the remaining target(s).

The questions addressed in this investigation are part of a larger discussion in which the nature of information processing is debated – whether information processing is best characterized as discrete (consistent with a preprogramming model) or as continuous stages of processing (consistent with an on-line programming model) (Massaro & Cowan, 1993). Evidence from shorter sequences supports a preprogramming model, but no one has systematically addressed the shift from preprogramming to on-line programming as targets are added to a sequential aiming movement. Rapid aiming sequences have less of a memory component (if targets are visible to the subject), a greater spatial constraint, and a greater interdependence between elements (the end of one segment is the beginning of the next) than other types of sequences such as speaking syllables or typing. Thus, different factors are likely to govern their programming and execution.

The perspective underlying the present investigation is that central processing, specifically motor programming, directs the execution of a movement sequence.1 Motor programming refers to the selection, translation and activation of various features of the movement, such as order of sequential elements, distance, direction and velocity. With practice these features become selected more readily for a given movement, which indicates that some type of central representation of the movement exists, and is modified by practice (for example, see Fischman & Lim, 1991). When the movement to be executed is known before the go signal, as was true in the present study, selection of the correct parameters can occur before the go signal (Klapp, Wyatt, & Lingo, 1974; Osman, Kornblum, & Meyer, 1990). Therefore, the primary programming processes studied in the present investigation are translation and activation of the motor commands, and updating these commands as needed during movement.

Fischman (1984) addressed the effect of increasing movement complexity (increasing the number of targets up to five) specifically on preprogramming. He found that increasing the number of targets arranged in a straight line primarily affected premotor time (motor time varied little) and that simple reaction time (RT) increased in a linear fashion as targets were added to the sequence. He interpreted these results to support a preprogramming model in which increased complexity of the movement increased preprogramming time. He also found that movement time to the first target increased as up to three targets were added to the movement. No explanation was offered for this increase.

Chamberlin and Magill (1989) also found that as they added a second and third target to an aiming movement, RT increased, and so did movement time to the first target. They suggested that this greater time was the result of programming “spilling over” into that movement. Other evidence to support this position is in handwriting research in which programming shifts to the early part of the movement with practice (Portier et al., 1990).

Fischman and Reeve (1992) directly addressed whether the increased time to the first target was the result of increased programming. Subjects were asked to complete a one- and two-target sequence. In one of the two-target conditions they were to contact the second target, and in the other they were simply to move above the second target. They assumed that the programming demands for moving above the target were less than the demands for accurately striking the target.They found that movement time to the first target increased whether the stylus contacted the second target or was merely moved above the target. Based on these results they suggested that a second movement functions to restrain the limb as it approaches the first target, and therefore increases time to the first target.

This peripheral constraint perspective described above is generally consistent with the subtended angle (SA) hypothesis (Sidaway, 1991; Sidaway, Christina, & Shea, 1988). A SA is the angle created by joining the starting point of the sequence with the lateral edges of a circular target (See Fig. 1 for a depiction of a SA). Sidaway (1991) found that RT varied more based on the size of the smallest SA rather than the number of targets (up to 3). The explanation he offered is that a smaller target creates a smaller SA and therefore requires more directional accuracy perpendicular to the line of movement. A smaller target (or the same size target moved further from the starting position) creates a smaller SA. Greater directional accuracy requires the movement to be more precise, and therefore requires more time for preprogramming.

In addition to affecting RT, the precision of the movement itself should affect movement times and dispersion of contacts on a target. A more constrained pathway should result in less dispersion of target contacts and a longer movement time to that target. Functionally, the target size is reduced, or in Welford's terms (Welford, 1968), the “virtual target” is smaller. (Note the size of the “I” bars in Fig. 1.) A speed-accuracy trade-off dictates that as the target size decreases, time to the target increases. Thus, as the SA decreases, the virtual target decreases and execution times should increase. So, it would be predicted that movement time to target one would be longer as targets are added to the sequence because the SA is decreasing, making the virtual target-one smaller (less spatial dispersion perpendicular to the line of movement).

This extension of Sidaway's hypothesis has been supported in two studies in which two-target sequences were used. Sidaway, Sekiya, and Fairweather (1995) reported that variability of contacts on the first target was less when the second target was smaller, i.e., a smaller overall SA resulted in less dispersion of contacts on target one. In a subsequent study in which the actual pathway was measured, a small second target resulted in a more constrained pathway compared to a larger second target (Short, Fischman, & Wang, 1996).

This “peripheral constraint” explanation offered by Fischman and Reeve (1992) and the SA hypothesis (Sidaway, 1991) are consistent with a preprogramming model, which counters the on-line interpretation of the increased movement time to the first target. However, while a peripheral constraint explanation may be adequate for a two- or three-segment movement, it is possible that a longer sequence may cause the capacity of preprogramming to be reached, with on-line programming occurring during execution to the first (or subsequent) targets.

Thus, a primary question addressed in this series of experiments was whether the increased time to target one is indicative of increased central processing, or is more consistent with a peripheral constraint explanation. This was examined by comparing RTs, flight times (FTs) and contact times (CTs), target contact dispersions and kinematics across the sequences varying in length from one to seven targets. Data to support a peripheral constraint explanation would be that movement to all targets should increase, as the number of targets in the sequence increases, not just movement to the first target. If subsequent targets are causing previous movements to be constrained, then this effect should occur on all targets. A second prediction was that RT should increase with the number of targets, thus indicating increased preprogramming (not on-line programming) with increased complexity or accuracy of the movement. A third prediction is that the size of the virtual target should decrease as targets are added to the movement; and, based on the SA hypothesis, this decrease should be in the direction perpendicular to the line of movement.

Data to support an early on-line programming explanation (programming during movement to the first target) would be that, specifically, movement time to the first target should increase, but that this pattern should not be evident in movement times to the other targets. Secondly, RT should plateau at some point as targets are added to the sequence, indicating that on-line programming, not preprogramming, is occurring to accommodate the increased complexity.

Section snippets

Subjects

Twelve right-handed university students (six males, six females) between the ages of 20 and 36 were recruited for this study (M=22.4 years, S.D.=4.68). Handedness was confirmed using the Crovitz–Zener Handedness Inventory (Crovitz & Zener, 1962). Subjects received class credit or were paid $6 an hour for participation. All recruitment and testing of subjects were performed in accordance with institutional procedures.

Apparatus and task

Data were collected using an electro-mechanical digitizer that provided

Experiment two

The purpose of the second experiment was to test whether the amount of central processing occurring during movement to the first target increases as targets are added to the sequence. If the peripheral constraint explanation is sufficient to explain the increased time to the first target, then central processing should not differ as targets are added to the sequence. If, however, central processing were greater during the first movement, this would be interpreted as evidence for on-line

Experiment 3

To examine the effect of random versus blocked execution of trials on the distribution of programming (without the presence of a probe) a third experiment was conducted. It was expected that if random order of execution increased on-line programming then RT would not increase with targets, and flight and/or CTs would be longer. In addition, 9- and 11-target sequences were included to determine the limit to which a rapid aiming sequence can be preprogrammed, and to examine the resulting on-line

General discussion

The overall purpose of these three experiments was to examine how the programming and control of a rapid aiming sequence shift with increased complexity. One objective was to determine if a preprogramming/peripheral constraint explanation is adequate to characterize control of an increasingly complex rapid aiming sequence, and if not, at what point on-line programming better accounts for the data. A second objective was to examine when on-line programming occurred. Previous data suggested that

Acknowledgements

The authors would like to thank Dr. H. Zelaznik, Dr. J. Gallagher, Dr. J. Thomas and an autonomous referee for critically reviewing this manuscript. We also thank Michele Subbot for her help in data collection and analysis. Portions of this data were presented at the conference for the North American Society for the Psychology of Sport and Physical Activity, 1993, 1998 and 2000.

References (38)

  • D Elliot et al.

    The control of goal-directed limb movements: correcting errors in the trajectory

    Human Movement Science

    (1999)
  • Favilla, M., Gordon, J., Ghilardi, M. F., & Ghez, C. (1990). Discrete and continuous processes in the programming of...
  • M.G Fischman

    Programming time as a function of number of movement parts and changes in movement direction

    Journal of Motor Behavior

    (1984)
  • M.G Fischman et al.

    Influence of extended practice in programming time, movement time, and transfer in simple target-striking responses

    Journal of Motor Behavior

    (1991)
  • M.G Fischman et al.

    Slower movement times may not necessarily imply on-line programming

    Journal of Human Movement Studies

    (1992)
  • A Garcia-Colera et al.

    Distributed planning of movement sequences

    Journal of Motor Behavior

    (1988)
  • F.M Henry

    Use of simple reaction time in motor programming studies: a reply to Klapp, Wyatt and Lingo

    Journal of Motor Behavior

    (1980)
  • F.M Henry et al.

    Increased response latency for complicated movements and a “memory drum” theory of neuromotor reaction

    Research Quarterly

    (1960)
  • Klapp, S. T. (1996). Reaction time analysis of central motor control. In H. N. Zelaznik (Ed.), Advances in motor...
  • Cited by (0)

    View full text