Research report
Spatially constrained locomotion under informational conflict

https://doi.org/10.1016/S0166-4328(01)00185-1Get rights and content

Abstract

This study investigates the informational based that supports intentional adaptation of locomotion to spatial environmental constraints. A virtual reality setup was used to present subjects with targets providing normal as well as abnormal optical expansion during locomotor pointing (i.e. positioning of a foot on a visible target on the floor during walking). The manipulation dissociated two variables providing temporal information about time-to-passage (TTP): TTPβα which encompasses target expansion, and TTPα which is independent of target expansion. While a previous study showed TTPα to be sufficient, the present results reveal that TTPβα may be used when it is available. This finding indicates that both variables play a role that varies according to the circumstances. Furthermore, the present results provide evidence of the operation of a security principle for action in conflicting situations.

Introduction

Neurophysiological experiments in animals have revealed that rhythmic movements of locomotion are generated by central pattern generators (CPGs) located at a relatively low level of the central nervous system [5]. Additional studies have shown that stable locomotion is achieved by the dynamic interaction among these CPGs and the musculo-skeletal system [18], [19]. On the other hand, intentional adaptation of the resulting basic locomotion is assumed to be achieved by the superposition of a phase-dependent command [18], [20]. This latter assumption was supported by a study in which transcranial magnetic stimulation of the motor cortex provoked a phase-dependent response in human locomotion [17]. In the present study we will focus on a less documented but essential point of the locomotion mechanism, i.e. the informational support needed to successfully modulate this phase-dependent command when locomotion is adapted to spatial environmental constraints.

Locomotor pointing, i.e. the positioning of a foot on a visible target on the floor during walking, has been largely used to investigate visually guided locomotor control in accordance with spatial constraints in the environment. From these studies, information about the time remaining before the target is reached has been suggested as the key element for locomotor pointing control [3], [9], [11], [22]. Actually, temporal rather than static spatial information directly reflects the subject's movement toward the target, and temporal information is directly available in the optical flow. More specifically, the temporal information largely evoked refers to the optical stimulus of ‘looming’, i.e. the symmetrical expansion of an approaching object [16]. In fact, the optical variable tau (i.e. the ratio of the retinal image size to the expansion rate of the image of an object approaching at a constant velocity) provides an accurate measure of time-to-contact [10]. In addition, neurophysiological studies have revealed the existence of looming-sensitive neurons [21] (for a review see [13]). However, tau has been shown to be invalid for locomotor pointing, since off-axis approaches fail to provide a symmetrical expansion to the subject [7]. However, it appears that valid temporal information is still available in such cases. Fig. 1 illustrates two alternative variables that specify the time-to-passage (TTP); i.e. the time remaining before the subject's eyes pass over a target on the floor toward which they are walking at a constant velocity. While the first variable (TTPβα) involves the angle β (sustained by the width of the target at the point of observation) combined with the angle α (between the vertical and the eye-target direction), the second variable (TTPα) only involves α. For this reason, only one (i.e. TTPβα) variable encompasses target expansion.

In the light of the important role of looming (i.e. the existence of looming-sensitive neurons [13], [21] and the well-documented use of tau to control action [12], [15]), one can legitimately argue that TTPβα is the more important of the two variables mentioned above (i.e. TTPβα and TTPα). Yet locomotor pointing control and final performance have been found to be similar for targets providing normal expansion and for those that do not provide any expansion at all (i.e. extensionless dots) [3]. Thus, TTPα (which is independent of target expansion) seems to be a good candidate for locomotor pointing control. It must be noted, however, that by eliminating target expansion one cannot claim to exclude the possible use of TTPβα when target expansion is in fact available. To test the validity of this reasoning, a virtual reality setup is used to present subjects with targets that provided normal as well as abnormal expansion. Furthermore, abnormal expansion modifies TTPβα but not TTPα, which raises the interesting question of how the nervous system reacts to conflicting variables.

Section snippets

Materials and methods

Ten subjects participated in this study on a voluntary basis. The experimental setup consisted of a virtual environment (silicon graphics) conected to a treadmill (Gymroll) [4]. Subjects walked on the treadmill in front of a screen (2.3 m high×3 m wide) on which a virtual environment was projected. The environment was dark and contained white virtual targets (0.1 m deep×0.5 m wide) positioned at irregular spatial intervals on the treadmill floor (the subject's track). The irregular intervals,

Results

First, we measured the occurrence of locomotor controls through changes in the variability of the toe-target distance during the successive support phases before final pointing. As illustrated in Fig. 2, the standard deviation pattern of the successive toe-target distances was similar in all experimental conditions, and exhibited an abrupt decrease in variability a few steps in front of the target. Thus, the expected locomotor controls were actually observed in the three experimental

Discussion

The similar pointing performance observed here can be interpreted on the basis of a simulation (Fig. 3) that illustrates the respective contributions of the two components of TTPβα that enter into TTP estimation, i.e. the α component (independent of target expansion) and the β component (dependent on target expansion). This simulation represents the last 2 m of a normal approach (N-E) using a constant velocity equal to the mean velocity of the actual approaches performed by the subjects (1.52 m

Conclusion

Combined with previous work [3], our data highlight the adaptive capabilities of the nervous system. When locomotion was adjusted in order to meet specific spatial environmental constraints, the system used two informational variables instead of a single one, even though one variable would have been adequate. Although this strategy seems to be far from parsimonious at first sight, the use of multiple variables enables the system to implement a security principle in its interaction with the

References (22)

  • A. De Rugy et al.

    The control of human locomotor pointing under restricted informational condition

    Neurosci. Lett.

    (2000)
  • F.C. Rind et al.

    Seeing what is coming: building collision-sensitive neurons

    Trends Neurosci.

    (1999)
  • R.J. Bootsma et al.

    On the information-based regulation of movement: what Wann (1996) may want to concider

    J. Exp. Psychol. Hum. Percept. Perform.

    (1997)
  • J.E. Cutting
  • A. De Rugy et al.

    The study of locomotor pointing in virtual reality: the validation of a test set-up

    Behav. Res. Methods Instr. Comp.

    (2000)
  • S. Grillner

    Neurobiological bases of rhythmic motor acts in vertebrates

    Science

    (1985)
  • J.G. Hay

    Approach strategies in long jump

    Int. J. Sport Biomech.

    (1988)
  • M.K. Kaiser et al.

    Optical specification of time-to-passage: observers' sensitivity to global tau

    J. Exp. Psychol. Hum. Percept. Perform.

    (1993)
  • M. Laurent et al.

    Binocular invariants in interceptive tasks: a directed perception approach

    Perception

    (1996)
  • M. Laurent et al.

    How is gait visually regulated when the head is travelling faster than the legs

    J. Mot. Behav.

    (1988)
  • D.N. Lee

    A theory of visual control of braking based on information about time-to-collision

    Perception

    (1976)
  • Cited by (22)

    • Information for step length adjustment in running

      2005, Human Movement Science
      Citation Excerpt :

      In short, our findings argue against the primacy of any of these schemes for adjusting step length in running. We compared our findings concerning step length adjustment in running with those of previous studies of treadmill walking (de Rugy et al., 2000, 2001) and highlighted two important similarities. First, in both running and treadmill walking, local tau appears to be useful for regulating step length, but not necessary.

    View all citing articles on Scopus
    View full text