Sources of Variation in the Output of Locust Spiracular Motoneurones Receiving Common Synaptic Driving

On each thoracic segment of the locust is a pair of spiracles, which allow air to enter the tracheae by opening during the inspiratory phase of ventilation. The flow of air in the tracheae is largely one-way. Air is inspired through the thoracic spiracles and expired through the abdominal ones. The opening and closing movements of thoracic and abdominal spiracles are thus out of phase. Closure of the thoracic spiracles during the expiratory phase is brought about by a burst of spikes in two doser motoneurones. Both axons of these neurones divide within the median, unpaired nerve to innervate the left and the right spiracles (Miller, 1960). The pattern of spikes that is recorded extracellularly in the median nerve during expiration shows considerable variation and Miller (1967) distinguished four patterns. Pattern 1 is a high-frequency burst of spikes of about the same frequency in both motoneurones which initially closes the spiracle. Pattern 2 is a regular sequence, in which the spikes of. both motoneurones occur at about the same time and therefore appear in pairs or groups. Pattern 3 is a sequence in which the spikes either occur in groups or there is beating. Patterns 2 and 3 maintain the closure of the spiracles throughout expiration. Pattern 4 occurs either during ventilation or in isolated thoracic ganglia and consists of independent sequences of spikes of different frequencies in the two motoneurones, producing beats. Miller (1967) postulated that the patterns arose in the following way. Pattern 1 was caused by one set of interneurones and patterns 2 and 3 by another set. These two sets of interneurones acted upon separate spike-initiating sites. A third set of interneurones adjusted the frequency of spikes. Pattern 4 was proposed to result from inherent pacemaking properties of the motoneurones. These inferences were all based on extracellular recordings of closer spikes, none of the interneurones being identified. Intracellular recording from the closer motoneurones has identified two sets of interneurones as responsible for cyclical synaptic input during ventilation. First, a pair of bilaterally arranged interneurones provides a major source of the excitatory synaptic input during expiration (Burrows, 1975a, b). Both of these interneurones make excitatory synaptic connexions with the two closer motoneurones of each spiracle in all of the thoracic segments, as well as synapsing upon many flight motoneurones and other ventilatory motoneurones. All thoracic spiracular closer motoneurones thus receive a common synaptic drive during expiration. The synaptic input is patterned. It consists of an overall depolarization determining the length of the expiration. Superimposed upon this is a faster rhythm sub-dividing the expiratory depolarization into a sequence of ripples with a period of about 50 ms in Schistocerca (Burrows, 1975 a, b). Secondly, during inspiration each closer motoneurone receives a similar barrage of inhibitory synaptic inputs which, together with the lack of excitatory input, abolish spikes (Burrows, 1975b). It is not known whether the interneurones providing the inhibitory drive are common to all closer motoneurones.

It is the purpose of this paper to show that the complex and variable patterns of closer motoneurone spikes are not incompatible with the simpler and more constant patterns of synaptic drive, given the demonstrated variation in the state of the motoneurones. It will also be shown that variation in intersegmental co-ordination of the spiracles is explained by these inputs.

Forty-two adult locusts, Schistocerca americana gregaria or Chortoicetes terminifera of either sex were obtained from crowded cultures. Intracellular recordings were made from the somata of spiracular closer motoneurones with glass micro-pipettes filled with 2 M potassium acetate. The procedure for obtaining stable intrasomatic recordings is given elsewhere (Hoyle & Burrows, 1973). Hook electrodes of platinum wire recorded the extracellular spikes from the median nerves. Extracellular potentials of the closer muscles were recorded with a 50 μm silver wire insulated but for the tip. The wire was also used to stimulate the terminals of the closer motoneurone axons within the muscle. Histograms of the intervals between spikes were plotted on a DL 4000 signal processor (Data Laboratories Ltd, London).

Intracellular recordings from the soma of, for example, a prothoracic spiracular closer motoneurone reveal that the expiratory burst of spikes arises from a barrage of excitatory postsynaptic potentials (EPSPs) (Fig. 1 a, b, c). During inspiration the soma is hyperpolarized by a barrage of inhibitory postsynaptic potentials (IPSPs) (Fig. 1 a). It is already established (Burrows, 1975 b) that the EPSPs are common to both closer motoneurones, but it is not known whether the IPSPs are also common. The similar patterns of spikes in the two closer motoneurones could result from the common synaptic driving, from direct electrical or synaptic coupling between them, or both. Several observations fail to reveal coupling, but one is crucial. The closer motoneurones of the pro- or mesothoracic spiracles spike tonically when both connectives between the meso- and metathoracic ganglia are cut (Fig. 2a). Visual inspection of extracellular recordings from the median nerve suggests that there is a phase progression of the spike of one motoneurone relative to that of the other motoneurone, and that no preferred interval is maintained. This is pattern 4 as described by Miller (1967). The beating between the spikes of the two motoneurones shows up clearly in a plot of successive spikes (abscissa in Fig. 2b) against the interval between those spikes (ordinate). The intervals between each spike gradually lengthen and then decrease so that there is no tendency to maintain a particular interval for any length of time. A histogram of the intervals shows no peaks, so that all intervals between zero and rooms must be equally represented (Fig. 2c). Therefore, should there be coupling between the two closer motoneurones of one spiracle, it is insufficient to synchronize their spikes. It would seem reasonable to exclude coupling from a role in the patterning of spikes in the closer motoneurones. The question is thus, to what extent does the pattern of spikes in the closer motoneurones follow the patterned common synaptic inputs? First, the patterns that are recorded from the median nerve must be analysed.

The closer motoneurones to the meso- or prothoracic spiracles spike tonically during expiratory pauses which may last for many seconds. The spikes in Chortoicetes occur at a frequency of about 25–30 Hz. In some expiratory pauses the spikes of the two motoneurones occur at about the same frequency, so that for short periods pairs of spikes are recorded in the median nerve (Fig. 3b). Each pair is separated by an interval of 30–40 ms, which is the period of the wingbeat in flight of these small locusts. This pattern of spikes corresponds to Miller ‘s pattern 2. Sometimes the frequency of spikes in the two motoneurones is different, so that beats occur (Fig. 3 a). Depending on the length of record that is examined, three of Miller ‘s patterns can be recognized. First, grouping of spikes (pattern 2). Second, grouping and beating of spikes (pattern 3). Third, beating of the spikes (pattern 4). If many sequences of spikes are examined it becomes obvious that patterns 2 and 3 merge into one another and are not separate entities. But the pattern of beats in an intact Schistocerca or Chortoicetes is not the same as when the motoneurones are deprived of their common synaptic inputs by cutting appropriate connectives (Fig.2). Miller (1967) lumped both beating patterns together as his fourth pattern. Sequential plots of the intervals between spikes emphasize that beating does occur but in addition show that a particular pattern of intervals may be maintained for short periods (Fig. 3b, c). A histogram of the intervals between spikes is no longer flat (cf. Fig. 2c) but contains peaks, indicating that some intervals occur more often than any others (Fig. d). An underlying periodicity can be detected in the sequences of spikes by triggering a signal averager from any spike and averaging 256 times those spikes which occur in the next 100 ms (Fig. 3 e). In Chortoicetes a peak occurs at 40 ms and a smaller peak at twice this period. In Schistocerca the first peak occurs after approximately 50 ms, which is the same as the wingbeat period during flight and of the fast rhythm of EPSPs to the spiracular closer motoneurones (Burrows, 1975 b). In expiratory pauses, therefore, the common synaptic driving exerts some patterning on the spikes. Patterns varying from apparent beating to synchrony are possible, but even during beating there is a tendency for particular intervals to occur more often than others.

In more rhythmical ventilation, similar sequences of spikes can be recognized (Fig. 4) but may be preceded, at the start of an expiratory phase, by a high-frequency burst of spikes corresponding to Miller ‘s pattern 1. Pairs or groups of spikes may then occur, with the groups separated by 30–40 ms in Chortoicetes (Fig. 4a), or by 50–60 ms in Schistocerca. This would correspond with Miller ‘s pattern 2. The spikes may occur at a steady frequency throughout expiration with only occasional grouping of the spikes (Fig. 4b), or a beat pattern between the spikes may occur (Fig. 4b, c). These sequences of spikes correspond to Miller ‘s patterns 3 and 4 respectively but, as for prolonged expiration, the categories cannot be considered absolute; the patterns overlap and one pattern may merge gradually into the next, even during one expiratory phase. It is common for different patterns to occur in successive expirations (Fig. 4b).

The pattern of EPSPs which underlies the spikes in an individual closer motoneurone can be revealed by simultaneous intracellular recordings from that closer motoneurone and a flight motoneurone (Fig. 5). The electrode in the closer motoneurone reveals the spikes, and that in the flight motoneurone the EPSPs which are common to both.

The synaptic input clearly is patterned; the EPSPs occur in groups at intervals of approximately 50 ms in Schistocerca so that the membrane potential ripples (Fig. 5a). During the patterned input, one of three different patterns of motor spikes may occur.

Firstly, the rhythmic input is not associated with a similarly patterned sequence of motor spikes (Fig. 5 a). Often, only one motor spike occurs on each group of EPSPs and at different times during the occurrence of that group. The hyperpolarization caused by the spike could render the next EPSP of that group incapable of exceeding the threshold of the motoneurone and evoking a second spike. Therefore the spikes occur tonically with a basic period equivalent to the interval between the groups of EPSPs, but somewhat masked by the irregular occurrence of the spikes on particular ripples.

Secondly, the rhythmic input is associated with a similarly patterned sequence of motor spikes (Fig. 5b-d). The spikes occur in groups of one to four at frequencies of about 100 Hz within each group separated from the next by intervals of about 50 ms in Schistocerca. The grouping of spikes may be established at the beginning of an expiratory phase and continue thereafter (Fig. 5 c), or in other expirations by the same locust, the motor spikes may be unpatterned at the start, to be followed by definite grouping (Fig. 5d). The initial unpatternedmotor spikes are associated with apparently unpatterned EPSPs in the flight motoneurone. A breakdown in the regular pattern of synaptic inputs (Fig. 5c, g) may also occur (Fig. 5a, h) and then the pattern of motor spikes also becomes irregular. On these occasions therefore, the pattern of motor spikes is a faithful reflexion of the pattern of synaptic potentials, indicating that a causal relationship between the two exists.

Thirdly, the rhythmic synaptic input is associated with a high frequency of apparently unpatternedmotor spikes, typically at the start of expiration (Fig. 5e, f). During inspiration the closer motoneurone is hyperpolarized by a barrage of IPSPs which cease at the onset of expiration. Possibly the rebound from this inhibition, coupled with an excitatory synaptic input, leads to the observed massive depolarization. The initial synaptic input may be patterned in the same way as the subsequent input, but the level of depolarization could be too great for it to have an influence upon the resulting sequence of motor spikes.

All the thoracic spiracular closer motoneurones receive the same patterned synaptic inputs from the interneurones during expiration. Can these inputs explain intersegmental co-ordination? It is not obligatory that each pair of spiracles open and close with every cycle of ventilation. For example, the closer motoneurones of the mesothoracic spiracle may spike in clear ventilatory bursts (Fig. 6a). Occasional spikes during inspiration do not prevent the spiracle from opening. The prothoracic closer motoneurones may spike continuously, however, with no indication of a change in frequency during either phase of ventilation (Fig. 6a). The prothoracic spiracles are observed to remain firmly shut while those in the mesothorax open and close rhythmically. This form of coupling is not an artifact of the electrophysiological recording because it is also observed in intact locusts. The coupling may change when, for example, an increased concentration of carbon dioxide causes the ventilatory rate to increase. Now bursts of closer motoneurone spikes arrive at all spiracles (Fig. 6b) so that the two sets open and close together. The frequency of spikes, however, may not be the same to each set, so that the extent of their movement is different. The patterning of spikes may also be different in each median nerve (Fig. 6c) but there is a tendency for groups of spikes to occur at the same time in both (cf. Miller, 1966). This is an indication of common synaptic driving which can be demonstrated more clearly by cutting appropriate connectives (Fig. 7). In the example chosen, the spikes in the mesothoracic median nerve of an intact locust occur in groups separated by intervals of 50 ms, but those in the prothoracic median nerve beat (Fig. 7a). When one pro-mesothoracic connective is cut the intraburst patterning in the mesothoracic median nerve is unaffected (Fig. 7b). In the prothoracic median nerve, however, the frequency of spikes and of beats is reduced, and the spikes now occur in groups which correspond with those in the mesothorax. The ventilatory rhythm in the prothorax is abolished when the second connective is cut; that in the mesothoracic nerve continues, but at a reduced fate. The frequency of spikes in the prothoracic nerve is now greatly reduced and the spikes of one motoneurone drift relative to those of the other so that a low frequency of beats occurs (Fig. 7c). The effect of the patterned input is also revealed in histograms of the intervals between prothoracic closer spikes. In an intact nervous system there is a distinct peak at very short intervals, indicating a tendency of the spikes of the two motoneurones to occur at about the same time (Fig. 8a). A second and smaller peak becomes more obvious when one meso-prothoracic connective is cut (Fig. 8b). The first peak is at 5 ms and the second is at 50 ms, indicating a tendency for the spikes to occur as groups, with the groups separated by the longer interval. When the second connective is cut all intervals are equally represented so that the resulting histogram is flat (Fig. 8c). Without the drive from ascending interneurones, the pattern of closer motoneurone spikes is the same as that of an isolated ganglion.

The different patterns of spikes of the two closer motoneurones recorded from a median nerve during expiration were previously explained by supposing that several sets of interneurones, which were not identified, synapsed upon motoneurones which themselves had complex properties (Miller, 1967). It was proposed that two sets of interneurones imposed three of the four patterns by acting upon separate spike-initiating sites in the motoneurones, whilst another set of interneurones adjusted the frequency of spikes within these patterns. The motoneurones themselves were believed to have an innate rhythmicity, which determined the fourth pattern, because an interpolated antidromic spike could reset this pattern (Miller, 1967).

The evidence presented here suggests that the patterns of spikes in closer motoneurones can be explained more simply by the effects of two interneurones. The interneurones impose a patterned synaptic drive upon both motoneurones which themselves have a tendency to spike tonically. In an isolated thoracic ganglion the closer motoneurones spike tonically and independently of each other but at about the same frequency, so that there are beats between their spikes. This pattern of spikes could result from an innate rhythmicity of the motoneurones or from unpatterned, but as yet unidentified synaptic inputs. It is upon this background of activity that the two interneurones superimpose their two patterns of commands during expiration (Burrows, 1975 a, b); the slow rhythmic depolarization of the motoneurones determines the duration of expiration, and the fast rhythm may determine the intraburst patterning. The synaptic drive consists of groups of EPSPs that occur at about 50 ms intervals, causing the membrane potential of the motoneurones to ripple. These waves and the threshold for initiation of spikes in the motoneurones determine the pattern of motor spikes. An explanation of how the different patterns of spikes arise from the known inputs is based upon experimental observations such as those in Fig. 5. For a single motoneurone receiving only a patterned synaptic input there are three possible output states. First, no output, because the input fails to reach threshold for spike initiation. Second, a high-frequency but unpatterned output, because the input exceeds threshold at all times. Third, a patterned output, because the driving input is between the two previous extremes. This is an adequate explanation for the high-frequency burst seen at the onset of expiration (Miller ‘s pattern 1) and for the grouping of spikes seen during the expiratory phase (Miller ‘s pattern 2). The real situation is more complex than this. Firstly there are two motoneurones involved, which in isolated ganglia spike tonically, and which may differ in several of their properties. For example, their thresholds for spiking may differ, or they may receive unpatterned and perhaps independent synaptic inputs from other neurones, although an electrode in the soma has yet to detect these. Secondly, the efficacy of the patterned synaptic input may vary in the two motoneurones. Thirdly, the patterning of the input may at times break down, as is apparent in Fig. 5 h. Two more output states are now possible. First, an output in which there is no underlying pattern. This occurs only in isolated ganglia. Second, an output in which a pattern can be revealed though masked by the noise of an unpatterned output, for example that revealed in Fig. 4. Miller (1967) grouped these two patterns together as his pattern 4, though it is clear from these results that there is a distinction in the underlying cause of the patterns. His pattern 3, in which grouping and beating of spikes occurred, has been shown to depend merely on the length of records examined. If the difference in frequency between the spikes of the two motoneurones is small, then the beat frequency will be low and there will be short periods when grouping of spikes will occur. There is thus no need to invoke complex sets of interneuronal driving to explain the various sequences of closer motor spikes. Indeed the patterns are not separate entities but a continuum.

The independent action of spiracular closer motoneurones in different segments can be explained in a similar way. The patterned excitatory inputs will have little or no effect upon those motoneurones of one segment that are already spiking at high frequency due to other causes. In another segment the motoneurones may spike only when they receive an input from these interneurones. The spiracles of one segment will therefore open and close rhythmically while those of another segment remain closed. Obviously many gradations can occur between these two extremes.

The point to emphasize, and which can be extrapolated to other neurones, is that a major patterned synaptic input does not necessarily lead to a similarly patterned sequence of output spikes. To be effective in producing a patterned output, the input pattern must operate within narrow limits on either side of the threshold of the follower neurone. If the depolarization is too large, a high frequency of unpatterned spikes will result ; if too small, then either there will be no output or a low frequency of spikes will result whose patterning will be affected by other inputs. Therefore the parsimony of the insect nervous system in driving all the spiracular closer motoneurones of the thorax with the same interneurones, does not imply that the behavioural repertoire of the spiracles is unduly limited.

This work was supported by a grant from the Nuffield Foundation.