Elsevier

Acta Psychologica

Volume 99, Issue 2, July 1998, Pages 201-215
Acta Psychologica

Planning of action sequences

https://doi.org/10.1016/S0001-6918(98)00011-0Get rights and content

Abstract

This paper reports an experiment on the advance planning of actions. Twelve subjects grasped an 8-sided object, and performed a sequence of movements involving juxtaposing a designated face of the object with a designated target. The location of each finger of the right hand on the faces of the 8-sided object was recorded. Results showed that the grip pattern differed significantly from a case where subjects were asked simply to grasp the 8-sided object, without moving it. Further, comparison of sequences of moves involving different target positions showed that subjects did chose grip patterns based on the specific moves they were asked to make. These adjustments were generally more common when the critical item occurred earlier in a sequence than when it occurred later, suggesting a gradient of advance planning comparable to the serial position effect in memory tasks. Further, the grip adjustments observed consisted primarily of reconfigurations of the hand, involving changing the point of contact on the object of a subset of fingers, rather than rotations of the whole hand with respect to the object.
PsycINFO classification: 2330

Introduction

Intuition tells us that we plan our actions in advance, but the human performance literature offers relatively little data on what we plan in advance, how far in advance we are capable of planning, or how good our plans are. The best-known evidence for planning in the human performance literature is the proportional relation between reaction time and the complexity of an action (e.g., Sternberg et al., 1978). A detailed review of this class of experiments may be found in Chapter 3 of Rosenbaum (1991). In these paradigms, the increased RT of more complex and extended actions is taken as evidence that the whole action sequence is planned in advance, and stored internally, before any movement occurs. Though influential, these studies have several shortcomings. First, they provide no direct evidence that the increased reaction time is occupied by preparation rather than by some other psychological process. Second, they provide no direct evidence about what is planned, but only that whatever is planned covaries with factors that influence reaction time. Third, the index of planning is not a feature of the movement itself, but of the period preceding the movement. Therefore, these studies give limited insight into what features of a movement can be used as indicators that it has or has not been planned.

Many of these shortcomings have been overcome in a method recently used by Rosenbaum et al. (1992). This method decides whether the second of two movements has already been planned when the first movement is made by measuring whether features of the first movement are influenced by specific demands of the second task. Rosenbaum et al. asked subjects to pick up a pointer and then use it to point in one of several possible directions designated in advance by the experimenter. The subjects' choice of overhand or underhand grip was related to ratings judging the awkwardness that they would experience with each grip type when performing the impending pointing movement in the designated direction. Rosenbaum et al. therefore concluded that subjects already represented the forthcoming pointing movement at the time of grasping the pointer; that is they planned at least one action in advance.

Rosenbaum et al.'s study offers one account of why people plan actions in advance, and what they plan: they plan to ensure comfort at later stages in the sequence. Other authors have proposed alternative parameters for advance planning of aimed movements. Often these parameters are entirely physical. For example, Hogan and Flash (1985)proposed that arm movements are planned to minimise the jerk (first derivative of acceleration) of the hand. In other studies, more explicitly psychological parameters of action are judged to be planned in advance. Cruse (1986)and Cruse and Brewer (1987)have proposed that the inverse kinematics problem in reaching (choosing just one of the possible sets of joint angles that are consistent with a desired hand position of a multi-joint limb) is constrained by planning to maintain each joint within its comfortable midrange. Haggard and Richardson (1996)and Haggard et al. (1995) suggested that subjects plan arm movements so as to simplify the computational complexity of the inverse kinematics that they involve. Meyer et al. (1988)emphasised the advance planning for the accuracy constraints of movement in their optimised submovement model. In this model, subjects plan an initial submovement towards a target in advance, taking into consideration the probability of hitting the target with this first attempt, and attempting to minimise the overall movement time taken by all the submovements that are likely to be required. All these studies of planning for optimal aimed movements generally assume that the optimised parameter is planned in advance of movement onset. However, the data alone cannot exclude the possibility that, say, hand jerk is minimised by an execution process which monitors the hand's acceleration as the movement progresses. Rosenbaum's movement selection method identifies advance planning more specifically.

The finding of Rosenbaum et al.s (1992) raises the interesting question as to whether subjects can plan more than one action in advance. Although little formal work has been done in this area, there is evidence for suggesting that planning of action sequences can occur with a longer scope than that described by Rosenbaum et al. (1992). For example, Lee et al. (1982)studied the gait of long-jumpers as they approached the departure board. They found that experienced jumpers began to modulate their step length so as to precisely hit the departure board several strides before they arrived there. This implies that planning for the final take-off occurred several strides in advance. However, it is not clear that planning a sequence of repetitions of a single movement (a running step) resembles planning a sequence of several distinct and arbitrarily related movements as in Rosenbaum et al.'s paradigm.

Assuming people can plan such arbitrary sequences more than one action in advance, is there a planning gradient within a sequence of actions, such that the first action in a sequence is “better” planned than the second, and so on? Understanding the temporal extent of planning in sequences of action would provide a useful insight into how the motor system groups individual movements into larger sequences, and would provide an indication of the capacity and durability of advance planning, comparable to the digit span and serial position effects in memory tasks. Further, grasping tasks similar to that of Rosenbaum et al. (1992)have been said to involve both independent channels for orienting the hand with respect to the object, and for placing the fingers on the object surface (Paulignan and Jeannerod, 1996). Thus, it seemed valuable to dissociate these components of grasping in an action sequence paradigm, and to compare the temporal extent of planning each component.

We have accordingly investigated advance planning of grasps in human subjects. Following the work of Rosenbaum et al. (1992), we used movement selection as a behavioural index of advanced planning. In these paradigms subjects are instructed to make sequences of movements, for which they must choose the hand grip in advance. Changes in the choice of hand grip which are correlated with changes in the sequence of actions instructed are interpreted as evidence that subjects plan the sequence in advance so as to adjust their initial grip. Our results show (1) that subject's choice of grasp is influenced by the action sequence they have been instructed to make, (2) that this influence is unequally distributed across the five fingers, (3) that the advance choice of grasp may be modified by changing just one action within the instructed sequence, (4) that these modifications are more prominent when the critical movement occurs earlier in the sequence than when it occurs later in the sequence, and (5) that the modifications can be dissociated into rotation of the entire hand with respect to the object, and intrinsic adjustments of the configuration of one or more fingers.

Section snippets

Method

The apparatus is shown in Fig. 1. It consisted of an octagonal object made of cardboard 125 mm in diameter and 22 mm high. It could be grasped easily from above. The eight faces of the object were numbered as shown, so the numbers were visible from above. A pointer 2 mm in diameter extended 20 mm from the object between faces 3 and 4. The object was positioned on a table in the centre of an area enclosed by an octagonal fence, with the pointer pointing away from the subject's body in the

Results

Subjects were able to grasp the object and perform the instructed sequences of movements without difficulty. On around 1% of the movements, subjects did not insert the pointer directly into the target hole, but hit the surrounding fence, and had to make a corrective sub-movement to insert the pointer correctly. These trials were not discarded. No errors in the order of movements within sequences were observed, implying that memorising the sequences was not difficult.

Discussion

First, our findings replicate those of Rosenbaum et al. (1992)that people plan at least one action in advance. In 81% of the cases, subjects used a different grasp to perform a sequence of aimed movements with our octagonal object from the basic grasp used to pick it up. Our findings also replicate Rosenbaum et al.'s findings that the initial grip varies with the specific action to be made.

Moreover, we have shown that the scope or time-course of this planning extends beyond the immediately

Conclusion

Subjects adjust the initial grip they use to pick up an object according to the specific aimed movement that they will perform later in an action sequence. By varying the number of movements in a sequence, we have shown that this planning effect has a scope extending beyond the next movement (as previously shown by Rosenbaum et al., 1992) at least to the second and third movement in such sequences. Such adjustments can be separated into two kinds: rotation of the whole hand around the object,

Unlinked References

Haggard et al., 1995, Sternberg et al., 1978

Acknowledgements

I am grateful to Azwiyati Azmi and Katie Tsui for their assistance in running subjects for this experiment. I am grateful to David Rosenbaum and to two anonymous reviewers for their helpful comments on the manuscript.

References (16)

  • Arbib, M.A., Iberall, T., Lyons, D., 1985. Coordinated control programs for control of the hands. Experimental Brain...
  • H Cruse

    Constraints for joint angle control of the human arm

    Biological Cybernetics

    (1986)
  • H Cruse et al.

    The human arm as a redundant manipulator: The control of path and joint angles

    Biological Cybernetics

    (1987)
  • J.M Elliott et al.

    A classification of manipulative hand movements

    Developmental Medicine and Child Neurology

    (1984)
  • P Haggard et al.

    Patterns of coordinated multijoint movement

    Experimental Brain Research

    (1995)
  • P Haggard et al.

    Spatial patterns in the control of human arm movement

    Journal of Experimental Psychology: Human Perception and Performance

    (1996)
  • N Hogan et al.

    Moving gracefully: Quantitative theories of motor coordination

    Trends in Neurosciences

    (1985)
  • D.N Lee et al.

    Regulation of gait in long jumping

    Journal of Experimental Psychology: Human Perception and Performance

    (1982)
There are more references available in the full text version of this article.

Cited by (52)

  • Predicting other people shapes the social mind

    2024, Advances in Experimental Social Psychology
  • Development of motor planning for dexterity tasks in trisomy 21

    2014, Research in Developmental Disabilities
View all citing articles on Scopus
View full text