Full-length reviewPattern generation for stick insect walking movements—multisensory control of a locomotor program
Introduction
The study of the neuronal basis of walking has attracted many investigators. This derives in part from the fact that the results of studies on the generation of walking movements are of general interest for two reasons. (1) In comparison to other rhythmic movements, walking movements must be performed in a very flexible way to adapt to irregularities of the ground. Additionally the temporal and spatial relationships between the different legs and between the different joints of each individual leg change, when the animal changes its walking gait or its movement direction. This implies that the neural system generating the motor output pattern for walking movements must exhibit a marked flexibility. (2) The results on walking pattern generation are relevant to applied sciences. For example, rehabilitation methods (in particular the technology for functional electrical stimulation) [102]or the construction of walking machines (e.g., Refs. 46, 99, 103) need a detailed knowledge of mechanisms of the `real' system.
More and more results indicate that the principles of walking-pattern generation seem to be very similar for different, if not all, legged animals (cf. Refs. 94, 95). Therefore, principles found in insects may also be relevant for mammals and man. Since insects are much easier to investigate, the knowledge about walking-pattern generation is more detailed for these animals than for many mammals. This article summarizes the results obtained from investigations on the generation of leg movements during walking in the stick insect, one of the most thoroughly studied examples. It updates previous reviews 5, 71.
The stick insect walking system might as well be of general interest for another reason: today, it is generally accepted that all rhythmic movements are generated by central pattern generators (CPGs) whose actions can be modified and adjusted by sensory (peripheral) influences. The detailed neuronal basis of the interactions of central and peripheral mechanisms are only starting to emerge for most rhythmic movements 94, 96. In the stick insect walking system, peripheral influences have been shown to be very important in sculpturing the motor output towards the functional walking pattern. There are preparational advantages, like the possibility of stimulating sense organs, that enable the investigation of the interactions of central and peripheral mechanisms in this system in an easier way than in other rhythm generators. Up to now most investigations concentrated on this aspect on the operational level. But during the last years, knowledge about the structure and actions of the central neuronal networks emerged as well.
Walking studies on stick insects have a fairly long tradition. Stick insects (phasmids) form a group of approximately 2500 species, including the largest insect species, that provide advantages for the investigation of walking movements due to their long legs. Stick insects are slow climbing animals that are mainly active at night. Their habitat is the crowns of trees and bushes in tropical and subtropical regions. The first paper on the generation of walking movements in stick insects was published in 1921 [21]. Since then most work was done on the stick insect Carausius morosus (Fig. 1), but sometimes also larger species, like Cuniculina impigra were used.
The strategy used in the investigations of walking-pattern generation in stick insects was a top–down analysis. It can be divided into three stages (for details see Ref. [5]): (1) Quantitative description of the behavior; (2) Relating the behavior to systems that are unambiguously defined on the operational level (on the basis of the operations they perform). In other words, one tries to demonstrate that the behavior in question is generated by identifiable subsystems. As an example: the coordinated movements of all legs was related to the action of six single-leg-pattern-generators and coordinating pathways between them (see Section 2); (3) Elucidation of the neural basis of these specified subsystems.
Section snippets
Movement of all legs
Initially, the movements of all legs have been described quantitatively on three levels: (1) movement, (2) torques in single joints [39]and (3) the activity of identified motor neurons (summary in Ref. [5]). The results are mentioned here in detail only if they are indispensable for the causal analysis.
According to the strategy used the first question of the analysis was: What is the overall structure of the walking pattern generating system? The answer should relate the movements to certain
The general structure of a walking-pattern generator for a single leg
What are the functions of sense organs and central neuronal elements in the walking-pattern generator of a single leg? The first step in answering such a question was to denervate the central nervous system and to determine, whether the isolated CNS is able to generate a rhythmic motor output resembling the motor output during walking. In all cases the isolated thoracic CNS was able to generate the rhythmic alternating activity of antagonistic motor neuron pools of each single leg joint with
The detailed structure of a walking-pattern generator for a single leg
Section 3has shown that every leg joint possesses its own rhythm generating network. One such network seems to be a modular system the central rhythm generating (or bistable units) of which are strongly influenced by sensory information (see also Refs. 10, 74). In this section it will be shown that sensory information is not involved only in the transitions between stance and swing and vice versa, but it also plays a role in coordinating the activities of the different neuronal joint
Neural basis of the active reaction
We started the analysis of individual modules of the leg walking system (Fig. 9) with the analysis of the neural basis of the active reaction in the femur–tibia joint. Fig. 10A summarizes our current knowledge on the topology of the neuronal network controlling the tibial extensor motor neurons. By now, the detailed analysis has concentrated on the extensor portion of the femur–tibia joint control system, because the innervation of the flexor tibiae muscle is considerably more complex 52, 53.
Neural basis of other modules
Recently we have started the investigation of the neural basis of other modules of the walking-pattern generating system, e.g., the one including pathways from the femoral chordotonal organ to the motor neurons of the coxa–trochanter joint (example 4 in Section 4). The pathways are partly monosynaptic and partly include interneuronal pathways via non-spiking interneurons [78]. In the resting animal interjoint reflexes generated by fCO signals are rather weak. However, in the active animal they
Synopsis and test of the experimental strategy
The investigation of the origin of walking movements in the stick insect led first to the result that there are six pattern generators (one for each leg) and that the leg coordination results from the interaction between them (Fig. 3). It was then found that in each of these pattern generators the switch from one phase of the step cycle to the other is determined by several equivalent influences (sense organs of the `own' leg and coordinating influences from other legs). The more detailed
Discussion
Three aspects arise from the investigations on walking pattern generation in the stick insect that seem to be interesting for studies on this kind of locomotion conducted in other vertebrate and invertebrate systems.
Acknowledgements
We would like to thank Keir Pearson, Grigori Orlovski, Abdel El Manira, Tom Wadden, Harald Wolf and our colleagues at the University of Kaiserslautern for helpful criticisms on previous drafts of the manuscript. In addition we would like to thank Tove Heller, Sybille Watt and Ilse Winkler-Reske for technical assistance. Substantial part of the work by U.B. and A.B. described above was supported by DFG grants.
References (124)
The femur–tibia control system of stick insects—a model system for the study of the neural basis of joint control
Brain Res. Rev.
(1993)Local circuits for the control of leg movements in an insect
Trends Neurosci.
(1992)- et al.
Induction of rhythmic activity in motoneurones of crayfish thoracic ganglia by cholinergic agonists
Neurosci. Lett.
(1987) What mechanisms coordinate leg movement in walking arthropods?
Trends Neurosci.
(1990)- et al.
Phase-dependent reflex reversal during walking in chronic spinal cats
Brain Res.
(1975) - et al.
Phase-dependent modulation of primary afferent depolarization in single cutaneous primary afferents evoked by peripheral stimulation during fictive locomotion in the cat
Brain Res.
(1990) Pattern and control of walking in insects
Adv. Insect Physiol.
(1985)- et al.
Intrinsic neuromodulation: altering neuronal circuits from within
TINS
(1996) Proprioceptive regulation of locomotion
Curr. Opin. Neurobiol.
(1995)- et al.
Correlation between muscle structure and filter characteristics of the muscle-joint system in three orthopteran insect species
J. Exp. Biol.
(1996)