Bernhard Möhl

Proprioceptive input on the locust flight motor revealed by muscle stimulation

Summary1.The locust is suspended in the laminar air stream of an open jet wind tunnel. During flight the 1st basalar muscle of the hindwing (M127) or of the forewing (M97) is stimulated every 6th or 8th wing beat with a single electric pulse triggered by simultaneously recorded action potentials of another downstroke muscle. The exact time relationship of the stimulus relative to the wing beat cycle is varied systematically (Fig. 1).2.In response to the electric stimulus the time of occurrence of other downstroke muscles changes slightly depending on the phase of the stimulus respective to the wing beat period. Stimulation of M127 (hindwing) leads to the following results: all recorded muscles react qualitatively in the same manner. If the stimulus occurs during the downstroke phase (first half of wing beat period) up to 3 subsequent wing beat intervals in the direct downstroke muscles are slightly elongated (Fig. 2). If the stimulus occurs during the upstroke phase the next following interval is decreased and the subsequent interval slightly increased (Fig. 3).3.If the stimulated muscle is cut near its apodeme at the wing hinge region the influence of the stimulus on the other muscles vanishes (Fig. 6). Thus, the effect depends on the mechanical action of the stimulated muscle on the wing hinge. Therefore, the possibility is excluded that the antidromic spike evoked in the motor axon of the stimulated muscle is responsible for the described reactions.4.The latency between the stimulus and the next influenced muscle action potential may be as short as 20 ms (Fig. 7).5.Stimulation of the forewing 1st basalar muscle shows reactions which are different from animal to animal. In contrast to the “hindwing stimulation” different muscles may react differently to a particular stimulus (Fig. 8).6.The general conclusion which can be drawn is that the sensory equipment of the wing and wing hinge region monitors wing movement quite specifically and exerts subtle influences on motor elements of the flight system.

Activity of the direct downstroke flight muscles ofLocusta migratoria (L.) during steering behaviour in flight

SummaryThe locust,Locusta migratoria (L.), performs tethered flight in the laminar flow of a wind tunnel. The animal is fixed ventrally to a stimulus device which produces rotations (max. rotation angle ± 15°) (Fig. 2). Depending upon the position of the axes relative to space and to the animal the latter is forced to rotate in the frontal, mediosagittal, or transversal plane. Simultaneous electromyograms from the direct downstroke muscles of the wings (7 electrodes implanted on each side in the following muscles, numbered after Snodgras: forewing 97, 98, 99, hindwing 127, 128, 129, in the last 2 electrodes) are made and analysed by means of a special data acquisition processor (Fig. 1) and digital computer.One observes shifts in the time of occurrence (‘time shifts’) of the action potentials of muscles belonging to the same wing and to different wings. For example: A 20° rotation is correlated with a time shift of up to 6 ms between the muscles 971 and 991 (left forewing) and between 991 and 1271 (different wings) (Fig. 8b).On rotating the animal to the left within the frontal plane (yaw) (Figs. 7, 8), the left side muscles 971, 981, 1271, and 1281 fire earlier and the muscles 991 and 1291+2, 1 fire later than during straight ahead flight whilst on the right side the muscles 97r, 98r, 127r, and 128r fire later and 99r and 1291+2, r fire earlier.On rotation within the mediosagittal plane (pitch) (Fig. 9), the muscles 97, 98, and 1291+2 on both sides fire earlier when the head is raised, whereas 99 fires later. No definite statement can be made about the behaviour of the remaining downstroke muscles.On rotation towards the right within the transversal plane (roll) (Fig. 10), the downstroke muscles of the left forewing fire earlier whereas those on the right fire later. No general scheme can be derived for both hindwing downstroke muscles.Thus the behaviour of the direct downstroke muscles is essentially symmetrical to the position of the axes of rotation as far as the time shifts are concerned (Fig. 11).In general (despite certain exceptions, see Fig. 12), the time shift curves vary so greatly that no definite coupling between different muscles should be assumed.

Short-term learning during flight control inLocusta migratoria

Summary1.The generally observed asymmetry in the motor pattern of locusts flying under tethered conditions (Möhl 1985a) vanishes when the animal is given freedom in the three dimensional axes (by hanging the locust from a wire in the wind stream) (Fig. 2), thus proving that the asymmetry is an artifact of tethering.2.Locusts subjected to yaw stimuli during tethered flight exhibit a clear hysteresis in the time shift reaction of the flight muscles which suggests a time integrative characteristic (i.e. memory) within the steering mechanism (Fig. 3).3.Locusts in a flight simulator in which the difference of timing of two contralateral flight muscles is fed back to the yaw angle, are able to align themselves straight ahead against the air stream of the wind tunnel. To achieve this, they are forced to activate both flight muscles at the appropriate time relationship which is preset by the experimenter as a reference value for the flight simulator (Figs. 4–6).4.A particular time relationship of the flight muscles acquired in the flight simulator under closed-loop conditions is maintained after opening the feedback loop, thus demonstrating a memory feature in the co-ordination of the flight muscles (Fig. 7).5.Reversing the feedback signal in the flight simulator produces a positive feedback which leads to a constant extreme (i.e. saturated) yaw angle (Fig. 8).6.The results show that the flight motor output of the locust is not only defined by the momentary sensory input, but also by the time integrated sensory input (short-term flight experience). The biological significance of this in terms of behavioural plasticity is discussed.

The tritocerebral commissure giant (TCG) wind-sensitive interneurone in the locust

Summary1.The TCG interneurone of the locust integrates information from the wind-sensitive head hairs and the antennae. The cell is rhythmically active in tethered flight and is sensitive to a yaw imposed on the animal in a laminar stream of air — the left TCG is more sensitive to a yaw to the left and vice versa (Figs. 1, 5).2.Successively waxing over the different windhair fields (fields 1...5) and immobilising the antennae reveals that all these sense organs contribute to the sensitivity of the neurone to the angle of yaw (Figs. 2, 3). There is no mutual inhibition between the right and left TCG (Fig. 4).3.The cell shows a phasic-tonic reaction when either the resting or the flying animal is subjected to a stepwise yaw. However, the tonic component is stronger when the animal is flying (Fig. 7).4.During flight the TCG fires rhythmically in bursts synchronised with depressor muscle potentials, i.e., at about the time when the wing is in its top position. Yaw angle is coded by the number of spikes per burst and their timing within the wing-beat cycle. An increase of the TCG activity is correlated with an earlier beginning of the burst (Figs. 8, 9). There is no change in the interspike interval within the TCG burst during a yaw stimulus: the 1st and 2nd interspike intervals remain at an average of about 4 ms (Fig. 9).5.In straight flight, electric stimulation applied to the right TCG neurone (thus mimicking a yaw to the right) evokes antagonistic time shifts in downstroke muscles, equivalent to those observed in response to an actual right yaw being imposed on the animal (Figs. 10, 11). The results suggest that the TCG is an important pathway in mediating yaw-correcting behaviour.6.A small resetting of the flight rhythm is produced when the TCG is electrically stimulated during flight (Fig. 13).

The role of proprioception in locust flight control

Summary1.Electric stimulation of the forewing stretch receptor nerve (SR-nerve, N1D2) during flight has an excitatory influence on ipsilateral depressor motor neurones (M97 and M99) and an inhibitory effect on an ipsilateral elevator motor neurone (M83). This agrees with intracellular results obtained by Burrows (1975). The effects of the stimuli depend strongly on their phase within the wing beat cycle (Figs. 2, 3, 4).2.The action potentials evoked by the electric stimuli summate centrally because a stimulation burst of 4 pulses has a greater effect on the motor pattern than a burst of only 2 pulses (Fig. 5).3.The influence of electric stimulation is independent of simultaneous steering maneuvers during which the co-ordination of flight motor neurones is altered (Figs. 7, 8, 9). Thus, the proprioceptive mechanisms do not seem to be under the control of hierarchically higher steering centres.4.Electric stimuli given every wing beat period exhibit the existence of slowly summating pathways working over several seconds. These can be inhibitory (Fig. 11) as well as excitatory (Fig. 12b). There are also inputs, whose influence is independent of the phase with which they occur during the wing beat cycle (Fig. 10).5.Including data from parts I and II, the role of proprioception for the co-ordination of the basic motor pattern and its changes during steering behaviour is discussed.

‘Biological noise’ and plasticity of sensorimotor pathways in the locust flight system

Summary1.Locusts (Locusta migratoria) are flown in a flight simulator in which the relative timing of two flight muscles is used to control the yaw angle. These closed-loop muscles always show a strict correlation to the reference value of the feedback circuit whereas the timing of other muscles (in open-loop) is largely independent of the reference value.2.The strength of steering reactions in different flight muscles — tested with short imposed yaw stimuli — depends on whether the respective muscles during a particular flight are in closed loop or in open loop within the flight simulator feedback. Closed-loop muscles react stronger to the yaw stimuli than open-loop muscles.3.Closed-loop and open-loop muscles can be interchanged with one another during a flight. Concurrent changes in yaw sensitivity of the open-loop and closed-loop muscles develop within several minutes of flight.4.A logical scheme is proposed which explains the specific adjustment of the sensorimotor pathways. It takes into account the spontaneous fluctuations of the motor pattern (‘biological noise’) which serve as ‘ exploratory manoeuvers’ for the muscle specific adjustment of sensorimotor strength.

The role of proprioception for motor learning in locust flight

In a ‘muscle-specific’ flight simulator (simulator driven by muscle action potentials) locusts (Locusta migratoria) show motor learning by which steering performance of the closed-loop muscles is improved. The role of proprioceptive feedback for this motor learning has been studied. Closed-loop muscles were cut in order to disable proprioceptive feedback of their contractions. Since there are no proprioceptors within the muscles, this is a muscle-specific deafferentation. Cut muscles are still activated during flight and their action potentials can be used for controlling the flight simulator. With cut muscles in closed-loop, steering is less reliable as can be seen from the frequent oscillations of the yaw angle. However, periods of stable flight indicate that ‘deafferented’ muscles are still, in principle, functional for steering. Open-loop yaw stimuli reveal that steering reactions in cut muscles are weaker and have a longer delay than intact muscles. This is responsible for the oscillations observed in closed-loop flight. Intact muscles can take over from cut muscles in order to re-establish stable closed-loop flight. This shows that proprioceptive mechanisms for learning are muscle specific. A hypothetical scheme is presented to explain the role of proprioception for motor learning.

Design and Control Mechanisms for a 3 DOF Bionic Manipulator

Functionality and design of a bionic robot arm consisting of three joints driven by elastic and compliant actuators derived from biologically inspired principles are presented. In the first design standard springs with linear characteristics are utilized in combination with electrical drives. Different control approaches for the bionic robot arm are presented, discussed and evaluated by numerical simulations and experiments with regards to the long-term goal of a nature-like control performance

Introduction to Part VII

Most of the progress of brain science through the recent decade is based on investigations at the single neuron level. On the other hand, psychological research is carried out at a level of perceptions and actions that involve large portions of an entire brain. What are the organizing principles that make possible the orchestration of the multitude of single cell events into phenomena that appear as coherent “wholes” at the psychological and behavioral level? This is a multifaceted question and to approach it one has first to sort out the different levels on which the issue of “organization” and “architecture” may be considered. One level is the level of the organization of the material substrate which is the carrier of the functions. Neuroanatomists have come to a quite elaborate subdivision of the brain into distinct areas and putative “modules” that gives some hope to the analysability of the whole thing. However, it is by no means clear what the dominant driving factors are in the observed apparent modularity of the brain’s hardware. Many different evolutionary pressures have left their imprint on the structure of the brain, and some important seeming structural features may be mere accidents of evolutionary history.

Sensorimotor Mechanisms and Learning in the Locust Flight System

Flying locusts control their stability by means of an inborn ‘autopilot’ mechanism. With increasing flight experience, however, the neural control circuit for correcting disturbances is individually tuned for optimum performance. During this tuning the most efficient muscles for adjusting the yaw angle are found and recruited for improved performance. Proprioceptive feedback from the activity of the flight muscles is necessary for this motor learning. A simple model is suggested, explaining how such tuning function could be implemented by neural elements. A particular aspect of the tuning ability is the apparent strategy of the flight system to reduce the amount of feedback error messages during flight control. This idea is discussed with respect to the concept of prerational intelligence.