Proprioceptive Feedback and the Control of Cockroach Walking

Abstract

Cockroach walking is a remarkably adaptable behaviour, for these animals are able to traverse walls, floors and ceilings at rates up to 24 steps per second (Delcomyn, 1971). In adapting locomotory patterns to a variety of terrains, cockroaches encounter problems that are common to all terrestrial animals. Walking movements must be monitored to be efficiently performed, and unexpected loads, due to variations in the environment, must be rapidly counterbalanced. To resolve these problems, the cockroach central nervous system must incorporate information provided by limb proprioceptors into the patterns of motor activity used in walking. While the economy of design of sensory and motor systems of these animals has permitted extensive investigations of the response properties and reflex effects of these receptors (Mill, 1976), the basic questions of how inputs from specific proprioceptors are used in the adaptation of walking and which parameters of locomotion are determined by these inputs have often remained unresolved. This has been due, in part, to the following difficulties that arise in the interpretation of data obtained from isolated preparations: 1. Determining the adequate stimulus of a receptor. Proprioceptors were defined by Sherrington (1906) as somatic sense organs that are stimulated by ‘actions of the body itself’. However, as is apparent, all receptors of limb muscles and joints respond to a variety of external stimuli as well (Lissmann, 1950). Receptors that monitor forces resulting from muscle contractions, for example, also respond to external loads and simple predictions of their patterns of discharge in walking are not feasible. 2. Determining the effectiveness of reflexes in locomotion. A reflex implies a fixed relationship between a change in afferent input and changes in motor neurone activity (Sherrington, 1906). However, most reflexes studied have been found to be quite variable. Many reflexes show substantial changes in gain (intensity of motor neurone discharge) depending upon the behaviour (Forssberg et al., 1977) or state (Bässler, 1983) of the animal. Reflexes of both vertebrates and invertebrates can also be changed by training (Melvill Jones & Watt, 1971; Zill & Forman, 1983; Forman & Zill, 1984). In addition, 3ince the time of the earliest account of proprioceptive reflexes, many reflexes have been demonstrated to show ‘reflex reversals’, that is complete changes in reflex sign (Graham Brown, 1911; Bässler, 1976; DiCaprio & Clarac, 1981; Vedel, 1982). In most cases, the functions of these reversals or the mechanisms underlying this reflex plasticity are unknown.