Gait adjustments in response to an obstacle are faster than voluntary reactions

https://doi.org/10.1016/j.humov.2004.08.011Get rights and content

Abstract

It has been reported that obstacle avoidance reactions during gait have very short latencies. This raises the question whether the cortex can be involved, as it is in voluntary reactions. In this study, latencies of obstacle avoidance (OA) reactions were determined and related to latencies of voluntary stride modifications and simple reaction times (SRT) of hand and foot. Twenty-five healthy young adults participated in this study. While they were walking on the treadmill, an obstacle suddenly fell in front of their left leg. The first reaction to the obstacle was the moment at which the differentiated acceleration curve of the foot deviated from the control signal. Latencies of OA reactions were 122 ms (SD 14 ms) on average. Two very different avoidance reactions (lengthening and shortening of the stride) were noticed, but there was no avoidance strategy effect on OA latencies. OA latencies were significantly shorter as compared to latencies of voluntary stride modifications and simple reaction times of hand and foot. The short OA latencies could not only be explained from the dynamic nature of the task. It is suggested that subcortical pathways might be involved in obstacle avoidance.

Introduction

In daily life, many obstacles are encountered and have to be avoided while walking. If the obstacle remains unnoticed, a trip and possibly a fall will be the result. In most cases, however, the obstacle will have been noticed by the visual system and in response, the locomotor pattern will be adjusted pro-actively in order to avoid the obstacle. It has been shown that these adjustments are fast, as they can be implemented in the normal walking pattern within one step cycle (Chen et al., 1994a, Chen et al., 1994b; Patla, Prentice, Rietdyk, Allard, & Martin, 1999; Weerdesteyn, Schillings, Van Galen, & Duysens, 2003). But how fast are they exactly and how does the latency of these responses compare with voluntary reaction times?

This is an important issue, because the answer could provide more insight into the processes involved in this perception–action coupling. In previous studies, the choice of an avoidance strategy has been a topic of interest (Chen, Ashton-Miller, Alexander, & Schultz, 1994b; Patla et al., 1999; Weerdesteyn, Sierevelt, Nienhuis, & Duysens, 2001). The latency of these reactions could indicate whether they can be regarded as voluntary choices or as more automated behavior. Obstacle avoidance requires a response to a visual stimulus. Latencies during simple visual reaction time tasks start from approximately 200 ms and the latency increases when more than one response can be triggered during choice reaction time tasks (e.g. Carson, Chua, Goodman, Byblow, & Elliott, 1995). Typically, these responses are obtained from a stationary starting position. However, obstacle avoidance reactions are different in that they do not require initiation of a movement from a stationary starting position. The reaction to the obstacle is characterized by modulation of on ongoing movement. Alstermark, Eide, Gorska, Lundberg, and Petterson (1984) found that cats needed only 70–120 ms to adjust their reaching trajectories when the target was displaced unexpectedly. For humans, latencies of 100–140 ms have been reported for changing the direction of reaching movements (Brenner & Smeets, 2003; Carlton, 1981; Day & Brown, 2001; Day & Lyon, 2000; Paulignan, MacKenzie, Marteniuk, & Jeannerod, 1990; Paulignan, MacKenzie, Marteniuk, & Jeannerod, 1991; Prablanc & Martin, 1992; Soechting & Laquaniti, 1983; Zelaznik & Hawkins, 1983). For both cats and humans, evidence is growing that subcortical pathways are involved in these fast reactions (Day & Brown, 2001; Perfiliev, Petterson, & Varfolomeev, 2003).

With respect to the lower limb, Patla, Beuter, and Prentice (1991) studied modification of stepping trajectories. In their study, the participants stepped over an obstacle and sometimes a second obstacle occurred behind the first. They had to traverse the two within the same step by additional lengthening. The participants changed their stepping trajectories about 120 ms after presentation of the second obstacle. The initial trajectory modification was the same for a high and a low second obstacle. After another 120 ms the low and high obstacle trajectories started to deviate. The proposed explanation for these fast responses was that the first response is a general response to the presence of the obstacle, which is fine tuned later on the basis of the properties of the obstacle. This seems to be a reasonable explanation when a response in only one direction is possible, which is additional lengthening of the step. If the responses could be in two directions, lengthening or shortening, it is conceivable that latencies differ.

The aim of the present study was to determine the latencies of obstacle avoidance reactions during treadmill walking. Both lengthening and shortening of the stride could be chosen in order to avoid the obstacle. Latencies of obstacle avoidance reactions to lengthening and shortening were compared. In addition, these latencies were compared with those obtained after voluntary stride lengthening and shortening following a visual cue. Finally, in the same subjects, the classic simple reactions times of the hand and the foot were measured with the participants at rest and compared to latencies of obstacle avoidance reactions.

Section snippets

Participants

Twenty five young adults (4 men, 21 women) aged between 20 and 37 participated in the study. None of the participants suffered from any neurological or motor disorder. All participants performed the obstacle avoidance task (OA), while 12 of these participants also performed 3 additional tasks: voluntary stride modifications, a simple reaction time (SRT) task of the foot (SRT foot) and a simple reaction time task of the hand (SRT hand). They all gave informed consent to participate in the study.

Results

For the analysis of OA latencies, a total of 238 trials could be selected in which obstacle release occurred during late stance or early swing. The average median latency of the obstacle avoidance reaction for all subjects was 122 ms (SD 14 ms). In order to investigate whether choosing between two strategies adds to the time needed to react, the groups that used only one strategy (LSS, n = 3; SSS, n = 8) were compared to the group that used both strategies (n = 14). The mean latency was 117 ms (SD = 10 ms)

Discussion

The aim of the present study was to compare the latencies of obstacle avoidance reactions with different types of voluntary reactions. The results showed that the swing trajectory in response to the sudden occurrence of an obstacle could be modified very quickly and that this reaction was not dependent on the phase of obstacle release. The mean latency of OA reactions was 122 ms. Obstacle avoidance reactions were nearly 100 ms faster than voluntary stride modifications. OA latencies were also

Acknowledgment

This study was supported by the Dutch Science Foundation (NWO) and by a EU grant to J. Duysens (Eurokinesis). The authors want to thank Daan Koppens for conducting the pilot experiments on this subject.

References (29)

  • H.C. Chen et al.

    Effects of age and available response time on ability to step over an obstacle

    Journals of Gerontology: Medical Sciences

    (1994)
  • B.L. Day et al.

    Evidence for subcortical involvement in the visual control of human reaching

    Brain

    (2001)
  • B.L. Day et al.

    Voluntary modification of automatic arm movements evoked by motion of a visual target

    Experimental Brain Research

    (2000)
  • Den Otter, R., Geurts, A. C. H., Duysens, J. (2003). Step characteristics of stroke patients in obstacle avoidance....
  • Cited by (76)

    • Sensorimotor anatomy of gait, balance, and falls

      2018, Handbook of Clinical Neurology
    View all citing articles on Scopus
    View full text