Auditory neurons in the ventral cord of the locust: Morphological and functional properties

Neurophysiology goes wild: from exploring sensory coding in sound proof rooms to natural environments

To perform adaptive behaviours, animals have to establish a representation of the physical "outside" world. How these representations are created by sensory systems is a central issue in sensory physiology. This review addresses the history of experimental approaches toward ideas about sensory coding, using the relatively simple auditory system of acoustic insects. I will discuss the empirical evidence in support of Barlow's "efficient coding hypothesis", which argues that the coding properties of neurons undergo specific adaptations that allow insects to detect biologically important acoustic stimuli. This hypothesis opposes the view that the sensory systems of receivers are biased as a result of their phylogeny, which finally determine whether a sound stimulus elicits a behavioural response. Acoustic signals are often transmitted over considerable distances in complex physical environments with high noise levels, resulting in degradation of the temporal pattern of stimuli, unpredictable attenuation, reduced signal-to-noise levels, and degradation of cues used for sound localisation. Thus, a more naturalistic view of sensory coding must be taken, since the signals as broadcast by signallers are rarely equivalent to the effective stimuli encoded by the sensory system of receivers. The consequences of the environmental conditions for sensory coding are discussed.

Electrophoresis of polar fluorescent tracers through the nerve sheath labels neuronal populations for anatomical and functional imaging

The delivery of tracers into populations of neurons is essential to visualize their anatomy and analyze their function. In some model systems genetically-targeted expression of fluorescent proteins is the method of choice; however, these genetic tools are not available for most organisms and alternative labeling methods are very limited. Here we describe a new method for neuronal labelling by electrophoretic dye delivery from a suction electrode directly through the neuronal sheath of nerves and ganglia in insects. Polar tracer molecules were delivered into the locust auditory nerve without destroying its function, simultaneously staining peripheral sensory structures and central axonal projections. Local neuron populations could be labelled directly through the surface of the brain, and in-vivo optical imaging of sound-evoked activity was achieved through the electrophoretic delivery of calcium indicators. The method provides a new tool for studying how stimuli are processed in peripheral and central sensory pathways and is a significant advance for the study of nervous systems in non-model organisms.

A neural coding scheme reproducing foraging trajectories

The movement of many animals may follow Lévy patterns. The underlying generating neuronal dynamics of such a behavior is unknown. In this paper we show that a novel discovery of multifractality in winnerless competition (WLC) systems reveals a potential encoding mechanism that is translatable into two dimensional superdiffusive Lévy movements. The validity of our approach is tested on a conductance based neuronal model showing WLC and through the extraction of Lévy flights inducing fractals from recordings of rat hippocampus during open field foraging. Further insights are gained analyzing mice motor cortex neurons and non motor cell signals. The proposed mechanism provides a plausible explanation for the neuro-dynamical fundamentals of spatial searching patterns observed in animals (including humans) and illustrates an until now unknown way to encode information in neuronal temporal series.

The scope of neuroethology

Neuroethology, an interdisciplinary subdivision of neuroscience, has emerged in recent years. Since 1976 there has been a regular session under this heading at the annual meeting of the Society for Neuroscience. In 1980 two introductory texts in English were published on the subject (Ewert 1980; Guthrie 1980), and a third (Camhi 1984) was published recently. There is widespread interest in neural mechanisms underlying behavior, but they encompass such a vast array of often unrelated topics that proponents do not share common goals. This article describes the emergence of ethology as a discipline, pointing out that its practitioners were successful because they confined their research to stereotyped, complex, nonlearned, innate behavioral acts. A limited number of profoundly significant principles emerged. Each of these is redefined. The major concepts of earlier ethology were embodied in a simple hydraulic model used by Konrad Lorenz in 1949 (Lorenz 1950). It is pointed out that this model implies the existence of common neurophysiological mechanisms and neuronal circuitry. This model has now been made obsolete by neurophysiological progress, but with appropriate modifications an updated version may still be useful in focusing attention on possible principles. The initial aim of neuroethology should be to examine the neurophysiological events in a variety of behaviors, exhibited by diverse animals from different phyla, which meet the criteria of innate behavioral acts. The behaviors should be sufficiently complex to interest ethologists, yet they should be addressable with neurophysiological methods down to the cellular level. In the case of vertebrates this may mean working with brain slices as well as whole animals, but for some invertebrates recording should be possible in the nearly intact animal during execution of the behavior. The work will be exacting and very difficult, and it is not likely to get done at all unless neuroethologists recognize that they should both train and discipline themselves and restrict their attention to welldefined goals.

Neural Correlations Increase Between Consecutive Processing Levels in the Auditory System of Locusts

Nervous systems may encode information about sensory stimuli using the temporal relations in spiking patterns between neurons. By conducting simultaneous intracellular recordings on pairs of auditory neurons we measured the strength of correlations between elements of the first three processing levels in the metathoracic auditory network of the locust. We quantified the degree of spike synchronization and rate covariations that occur among auditory neurons during acoustic stimulation. In addition to the acoustic stimulation, current pulses were injected into both neurons to study the connectivity within this network. Our findings support the view that the metathoracic auditory system is a hierarchically organized feedforward network. Strong synaptic connections were observed only between consecutive processing levels, whereas there was no indication for strong connections between elements of the same processing level. Both spike synchronization and rate covariations were increased among neurons on higher processing levels. We further investigated the consequences that correlations may have on the common estimates of neuronal variability. For example, rate covariations caused by strong synaptic coupling between two neurons may lead to an overestimation if the variability is measured trial by trial with respect to only single neurons. For the vast majority of cell pairs tested, however, no strong synaptic coupling could be demonstrated. Thus we could show that in most cases no serious errors are made if one determines variability by following the usual procedure on the basis of single-cell recordings.

Diversity of intersegmental auditory neurons in a bush cricket

Various auditory interneurons of the duetting bush cricket Ancistrura nigrovittata with axons ascending to the brain are presented. In this species, more intersegmental sound-activated neurons have been identified than in any other bush cricket so far, among them a new type of ascending neuron with posterior soma in the prothoracic ganglion (AN4). These interneurons show not only morphological differences in the prothoracic ganglion and the brain, but also respond differently to carrier frequencies, intensity and direction. As a set of neurons, they show graded differences for all of these parameters. A response type not described among intersegmental neurons of crickets and other bush crickets so far is found in the AN3 neuron with a tonic response, broad frequency tuning and little directional dependence. All neurons, with the exception of AN3, respond in a relatively similar manner to the temporal patterns of the male song: phasically to high syllable repetitions and rhythmically to low syllable repetitions. The strongest coupling to the temporal pattern is found in TN1. In contrast to behavior the neuronal responses depend little on syllable duration. AN4, AN5 and TN1 respond well to the female song. AN4 (at higher intensities) and TN1 respond well to a complete duet.

Processing of auditory information in insects

Insects exhibit an astonishing diversity in the design of their ears and the subsequent processing of information within their auditory pathways. The aim of this review is to summarize and compare the present concepts of auditory processing by relating behavioral performance to known neuronal mechanisms. We focus on three general aspects, that is frequency, directional, and temporal processing. The first part compares the capacity (in some insects high) for frequency analysis in the ear with the rather low specificity of tuning in interneurons by looking at Q10dB values and frequency dependent inhibition of interneurons. Since sharpening of frequency does not seem to be the prime task of a set of differently tuned receptors, alternative hypotheses are discussed. Moreover, the physiological correspondence between tonotopic projections of receptors and dendritic organization of interneurons is not in all cases strong. The second part is concerned with directional hearing and thus with the ability for angular resolution of insects. The present concepts, as derived from behavioral performances, for angular resolution versus lateralization and serial versus parallel processing of directional and pattern information can be traced to the thoracic level of neuronal processing. Contralateral inhibition, a mechanism for enhancing directional tuning, appears to be most effective in parallel pathways, whereas in serial processing it may have detrimental effects on pattern processing. The third part, after some considerations of signal analysis in the temporal domain, demonstrates that closely related species often use different combinations of temporal parameters in their recognition systems. On the thoracic level, analysis of temporal modulation functions and effects of inhibition on spiking patterns reveals relatively simple processing, whereas brain neurons may exhibit more complex properties.

Serial hearing organs in the atympanate grasshopper Bullacris membracioides (Orthoptera, Pneumoridae)

In different insect taxa, ears can be found on virtually any part of the body. Comparative anatomy and similarities in the embryological development of ears in divergent taxa suggest that they have evolved multiple times from ubiquitous stretch or vibration receptors, but the homology of these structures has not yet been rigorously tested. Here we provide detailed analysis of a novel set of hearing organs in a relatively "primitive" atympanate bladder grasshopper (Bullacris membracioides) that is capable of signaling acoustically over 2 km. We use morphological, physiological, and behavioral experiments to demonstrate that this species has six pairs of serially repeated abdominal ears derived from proprioceptive pleural chordotonal organs (plCOs). We demonstrate continuity in auditory function from the five posterior pairs, which are simple forms comprising 11 sensilla and resembling plCOs in other grasshoppers, to the more complex anterior pair, which contains 2000 sensilla and is homologous to the single pair of tympanate ears found in "modern" grasshoppers. All 12 ears are morphologically differentiated, responsive to airborne sound at frequencies and intensities that are biologically significant (tuned to 1.5 and 4 kHz; 60-98 dB SPL), and capable of mediating behavioral responses of prospective mates. These data provide evidence for the transition in function and selective advantage that must occur during evolutionary development of relatively complex organs from simpler precursors. Our results suggest that ancestral insects with simple atympanate pleural receptors may have had hearing ranges that equal or exceed those of contemporary insects with complex tympanal ears. Moreover, auditory capability may be more prevalent among modern insect taxa than the presence of overt tympana indicates.

NEUROLAB, a comprehensive program for the analysis of neurophysiological and behavioural data

A comprehensive and versatile computer software for IBM-compatible microcomputers has been developed. It is designed for quantitative off-line analysis of A/D-sampled intracellular or extracellular recordings and behavioural or stimulus data. The program works with single files or file sets. It supports data to be viewed on the monitor and allows sectioning of interesting data for common analysis. It offers 19 filters/operators for data processing and comprehensive possibilities to set and calculate trigger points. Data of trigger points can be exported as ASCII files. Standard neurophysiological histograms like interval-, PST-, phase histograms or auto- and cross-correlograms can be obtained. Time-dependent and phase-dependent averaging is possible for all original and filtered data. All graphical output on the display can directly be copied to a plotter/laser printer or HP-GL file by keyboard commands.

Common synaptic drive to segmentally homologous interneurons in the locust

The aim of the present study was to examine the pattern of synaptic interactions among a set of identified homologous interneurons in the segmental nervous system of the locust. This paper presents two main findings: first, serially homologous interneurons that are the progeny of neuroblast 7-4 in the mesothoracic, metathoracic, and first abdominal neuromeres of the locust central nervous system all receive synaptic drive from one and the same presynaptic interneuron. This interneuron has its entire arborization located in these three neuromeres of the central nervous system. It synapses with cells that are siblings, bilateral homologs, and serial homologs, and is itself connected monosynaptically with auditory afferents. The neuronal network that results comprises postsynaptic cells with the same developmental lineage. The second finding is that there is an additional set of synaptic connections among the homologous neurons themselves. All these connections are excitatory, and the pattern of information flow within the network is highly directional. This may relate to the morphologies of the neurons involved and will influence the contribution of homologs from different segments to behavior.

Separation of synaptic and spike activity in intracellular recordings for selective analysis

A software spike filter has been developed which allows the separation of synaptic activity and action potentials in intracellular recordings. The algorithm uses the different velocities of the membrane potential during synaptic and spike activity and a time window to identify action potentials. When spikes are recognized, they are removed and the membrane potential is substituted by interpolated values. The spike filter makes possible a separate quantitative evaluation of postsynaptic potentials and spike activity. Thus a comprehensive characterization of neuron activity can be obtained. The spike filter is part of a modular software package designed for the evaluation of neurobiological data.

Convergence of multi-modal sensory signals at thoracic interneurons of the escape system of the cockroach, Periplaneta americana

Research on the escape system of the cockroach has focused upon the role of giant interneurons in conveying information on wind stimulation from the cerci located on the abdomen to motor control centers in the thoracic ganglia. In the thoracic ganglia the ventral giant interneurons connect to a population of interganglionic interneurons referred to as type A thoracic interneurons. In this paper we have tested the type A interneurons for additional sensory inputs in the absence of ventral giant interneuron activity. We find that the cells that receive ventral giant interneuron activity are also influenced by a variety of additional sensory inputs; wind mediated activity in a pathway that descends from the head, tactile inputs from several loci, auditory stimuli and light responses. Moreover, behavioral observations indicate that at least some of these activities can alter the escape movements. The results suggest that these interneurons serve as a site of convergence for numerous types of sensory activity. They further suggest that the escape system is capable of responding to directional wind information encoded in the ventral giant interneurons in the context of a wealth of additional information.

Audition in the praying mantis, Mantis religiosa L.: identification of an interneuron mediating ultrasonic hearing

1. The praying mantis possesses a single ear located in the ventral midline of the metathorax. We have studied the mantis' auditory nervous system using both extracellular and intracellular techniques and have identified anatomically and physiologically a mirror-image pair of interneurons (MR-501-T3) in the metathoracic ganglion that mediates ultrasonic hearing. 2. MR-501-T3 is tuned broadly to ultrasound with best sensitivity (55-60 dB SPL) between 25 and 45 kHz. Its tuning matches closely that of the whole tympanal nerve. 3. The physiological responses of MR-501-T3 are characterized by: (1) a phasic-tonic firing pattern with a distinctive initial burst at 500-800 spikes/s; (2) minimum latencies of 8-12 ms; (3) no spontaneous activity; (4) sigmoid intensity response curves with a small (10 dB) dynamic range; (5) accurate coding of stimulus duration and of repetition rates up to 60 pps. 4. The ascending axon of MR-501-T3 conducts action potentials at 4 m/s, a rate comparable with some giant fiber systems. 5. MR-501-T3 shows no directional capability. Sound from right and left produce identical responses in both cells of the pair. Neither cutting one tympanal nerve nor removing one hemi-ear leads to different responses in the two cells indicating that they must receive a common input, either from the auditory afferents or from interneurons. We present evidence that the two cells are not directly connected. 6. MR-501-T3 is a large, symmetrical cell with its processes primarily in the intermediate neuropil (lateral ring tract). Its integration segment crosses the midline in the supramedian commissure, and the cell body lies dorsally near the entrance of the leg nerve. The axon travels in the dorsal lateral tract and is one of the largest (17 microns) in the connective. 7. Given the strong anatomical similarities between MR-501-T3 and the G and B cells of the locust, these cells may be homologous. 8. We present arguments based on our physiological results and existing behavioral data that MR-501-T3 is part of an ultrasonic warning/escape system in the mantis. As in moths, lacewings, and crickets, this system may provide a defense against nocturnally foraging bats.

Organization of a sensory neuropile in the auditory pathway of two groups of Orthoptera

The anterior intermediate sensory neuropile (aISN) is a prominent neuropile in the ventral nerve cord of locusts and bushcrickets. Previous studies have shown that it receives its main sensory input from auditory receptors. In this paper we examine the structural and physiological relationship between tympanal receptor terminations and the dendrites of sound-sensitive interneurones in the homologous neuropile of locusts and bushcrickets. Each individual receptor fibre of the bushcricket terminates in a somewhat different target area of the neuropile. The ordering is with respect to the characteristic frequency of the fibres (tonotopic) in the anterior-posterior and dorsoventral axis. In the locust, representatives of the four tympanal receptor groups branch in different areas of the aISN. Most of the dorsal neuropilar region, and the anterior ventral region, do not receive input from tympanal receptors. The dendrites of identified sound-sensitive interneurones were examined in the context of this afferent projection. Local interneurones as well as intersegmental interneurones in bushcrickets have dendritic branches in the whole aISN or part of it and thus overlap with at least some receptors. By recording intracellularly from their main neurites, short-latency synaptic potentials were found in response to receptor spikes indicating monosynaptic input. The tuning of these neurones could be predicted by their dendritic morphology. In contrast, in the locust only local and bisegmental neurones are monosynaptically connected with tympanal receptors, but not the studied intersegmental neurones. This is consistent with the finding that most or all branches of intersegmental neurones lie in the dorsal area of neuropile where no receptors terminate. Anatomical and physiological evidence is presented for identified local neurones providing the excitatory and inhibitory synaptic input for such intersegmental neurones. The difference in the basic wiring diagram in the homologous neuropile of the two orthopteran groups is discussed with respect to the possible different roles that sound plays in their behaviour.

Morphology of local "stridulation" interneurons in the metathoracic ganglion of the acridid grasshopper Omocestus viridulus L

The morphological characteristics of five types of local spiking interneurons in the metathoracic ganglion of the acridid grasshopper Omocestus viridulus L. have been revealed by intracellular injection of the fluorescent dye Lucifer Yellow. All these neurons are active during induced stridulation and discharge in the stridulation rhythm. They are structurally unlike any of the metathoracic local interneurons previously described in the Acrididae. Type 1: Unilateral interneurons in the metathoracic neuromere; soma in anterior position; arborization dorsal, in anterolateral, posterolateral, and midline regions. Type 2: Unilateral interneurons in the metathoracic neuromere; neurite with characteristic hairpin bend; soma in lateral position; arborization through the lateral neuropil region. Type 3: Unilateral interneurons in the metathoracic and abdominal neuromeres; soma anterior at the base of the connective; arborization dorsal, in anterolateral, posterolateral, and midline regions, with ladderlike branches into the abdominal neuromeres. Type 4: Symmetrically bilateral neurons; soma in ventral median position; arborization dorsal, in anterolateral, posterolateral, and midline regions. Ipsilateral branches mostly smooth, contralateral mostly varicose. Type 5: Asymmetrically bilateral neurons, soma in ventral median position; ipsilateral branches mostly smooth, contralateral mostly varicose. The ipsilateral branches in particular penetrate all layers of the ganglion. The main arborizations of interneurons of types 1 and 4, and to some extent those of types 3 and 5, occupy the same region, which extends very far dorsally into the vicinity of the median and lateral tracts. There are additional antero- and posterolateral branches in the metathoracic neuromere, in positions where in principle they can overlap with the arborizations of the motoneurons. The interneurons of type 2, and to some extent type 5, arborize mainly in a lateral region that also includes deeper layers of the ganglion and hence encompasses the projections of mechanoreceptive proprioceptors. The interneurons of types 1, 3, and 4 have so-called "sister" neurons, identical to the typical neurons in their metathoracic arborizations but with additional collaterals that pass into the mesothoracic ganglion.

Spatial segregation of synaptic inputs and outputs in a locust auditory interneurone

Using intracellular injection of cobalt, processing for electron microscopy, and using on-section silver intensification, we have investigated the distribution of synapses on the metathoracic arborization of the TN1 intersegmental interneurone in the auditory system of the locust. Previous studies have indicated that inputs to this neurone from tympanal afferents are spatially segregated from its outputs to higher-order inhibitory interneurones. Branches in the area of input are characterised by a smooth appearance in Lucifer Yellow preparations, whereas branches in the main output area appear beaded. Sampling different areas of the TN1 arborization, we have found marked differences in the respective densities of input and output synapses. Labelled profiles in areas of beaded branching have an extremely high density of presynaptic sites, with a much smaller number of synaptic inputs. Smooth branches bear many input synapses, but we have found no output synapses on them. Our results indicate a good correlation between branch beading and presynaptic function and between smooth branches and postsynaptic function. These correlations enable us to predict synapse distribution on other insect neurones with a similar division in branch morphology and also extend our knowledge of the functioning of TN1 itself.

Modulation of auditory responsiveness in the locust

The auditory responsiveness of a number of neurones in the meso- and metathoracic ganglia of the locust, Locusta migratoria, was found to change systematically during concomitant wind stimulation. Changes in responsiveness were of three kinds: a suppression of the response to low frequency sound (5 kHz), but an unchanged or increased response to high frequency (12 kHz) sound; an increased response to all sound; a decrease in the excitatory, and an increase in the inhibitory, components of a response to sound. Suppression of the response to low frequency sound was mediated by wind, rather than by the flight motor. Wind stimulation caused an increase in membrane conductance and concomitant depolarization in recorded neurones. Wind stimulation potentiated the spike response to a given depolarizing current, and the spike response to a high frequency sound, by about the same amount. The strongest wind-related input to interneuron 714 was via the metathoracic N6, which carries the axons of auditory receptors from the ear. The EPSP evoked in central neurones by electrical stimulation of metathoracic N6 was suppressed by wind stimulation, and by low frequency (5 kHz), but not high frequency (10 kHz), sound. This suppression disappeared when N6 was cut distally to the stimulating electrodes. Responses to low frequency (5 kHz), rather than high frequency (12 kHz), sounds could be suppressed by a second low frequency tone with an intensity above 50-55 dB SPL for a 5 kHz suppressing tone. Suppression of the electrically-evoked EPSP in neurone 714 was greatest at those sound frequencies represented maximally in the spectrum of the locust's wingbeat. It is concluded that the acoustic components of a wind stimulus are able to mediate both inhibition and excitation in the auditory pathway. By suppressing the responses to low frequency sounds, wind stimulation would effectively shift the frequency-response characteristics of central auditory neurones during flight.

Mechanics of the transduction of sound in the tympanal organ of adults and larvae of locusts

The mechanical transmission of sound in the tympanal organ of adults and 5th instar larvae of Locusta migratoria and Schistocerca gregaria has been investigated by means of stroboscopic measurements within a frequency range from 1-20 kHz. Frequency dependent spatial distributions of amplitudes and phases of oscillation on the tympanal membrane and the Müller's organ could be demonstrated. Cuticular structures on the membrane may act as a lever arm (e.g. elevated process) and cause a transformation of the (unidimensional) membrane motion into components of displacements in the Müller's organ perpendicular, as well as even parallel, to the membrane. Sites of maximum relative displacements at distinct frequencies are found to be correlated to the course of the dendrites of the acoustic receptor cells. Differences in morphology of the tympanal organ between the two species as well as between adults and larvae always correspond to differences in the mechanical properties (resonances etc.). Consequently, differences or changes in the neurophysiological response characteristics of the different receptor cells have been found. Based upon these findings a correlation between the anatomical and physiological classification of the receptor cell groups is presented.

Spike synchronization of tympanic receptor fibres in a grasshopper (Chorthippus biguttulus L., Acrididae). A possible mechanism for detection of short gaps in model songs

In recordings from single tympanic receptor fibres in C. biguttulus, the response to synthesized sounds (rectangularly modulated white noise) interrupted by very brief (a few milliseconds) gaps was examined. In behavioral tests, females of the species respond very differently to such 'model syllables' at moderate intensities, depending on the gap width. If the gaps (in a moderate-intensity syllable) are larger than 2 ms, the stimulus fails to elicit a response, whereas stimuli with gaps smaller than 1 ms are as effective as uninterrupted syllables (D. von Helversen 1972; O. von Helversen 1979). Neither the mean spike count nor the interspike-interval distribution of the single receptor response contains the information sufficient to distinguish uninterrupted syllables from syllables with gaps. On the other hand, examination of the temporal distribution of the spikes reveals that gaps (or the pulse onsets following the gaps) cause spike synchronization. An index of synchronization (IS) was defined as a measure of this gap-induced effect. Analysis of the receptor responses based on IS revealed differences that correspond quantitatively to the abrupt abolition of the behavioral response at a gap-width between 1 and 2 ms. From the hypothesis that such brief gaps are detected by the nervous system by way of spike synchronization in the tympanic nerve, one can predict certain features of the behavioral response to high-intensity stimuli. The gap-induced spike synchronization was more pronounced at higher temperatures. This effect was demonstrated in both summated recordings from the tympanic nerve and single fibre recordings. Experiments with primary auditory fibres of Locusta migratoria showed that the receptors in this species respond very similarly to the same stimuli. That is, the receptors of C. biguttulus are not specially adapted for detecting very brief gaps. Synchronization of the spikes in parallel receptor fibres of the tympanal nerve is probably a general feature of acridids; we infer that in C. biguttulus this gap-induced synchronized activity is detected by special processing in higher auditory centres.

Heterogeneous properties of segmentally homologous interneurons in the ventral nerve cord of locusts

The G, B1, and B2 neurons are three prominent interneurons located in adjacent segmental ganglia in the central nervous system of locusts. Previous studies on the adult nervous system have shown that each of these cells has its own distinctive morphology and responsiveness to auditory input. Previous studies on the embryonic nervous system have described the lineage and development of one of these cells, the G neuron, in the mesothoracic (T2) segment. In this paper it is shown that the G, B1, and B2 neurons are segmental homologues in that they arise from equivalent lineages during embryogenesis in the T2, T3, and A1 segments, respectively. Each cell arises (along with its identified sibling neuron) from the division of the second ganglion mother cell of neuroblast 7-4. The segment-specific morphology of the G homologues was determined in the T3 and A1 segments between 60-70% of embryonic development, and their identity was established as the adult B1 and B2 neurons by comparing the distinctive cell-specific features of their morphology between embryo and adult. Although all three neurons display striking morphological differences, they all share certain structural features in common, including the location of their primary axons and neurites in specific tracts in the neuropil. By recording intracellularly from the main neurites of the G, B1, and B2 neurons, clear differences were found in the synaptic inputs each of the neurons receives and the synaptic outputs each makes. For example, G and B2, but not B1, receive direct monosynaptic input from the descending contralateral movement detector (DCMD) interneurons and from auditory afferents; B1, but not B2, connects directly to G; and B2, but not B1 or G, connects directly to flight motoneurons. The main conclusion from these observations is that lineally equivalent neurons in different segments can develop similar primary structures but quite different secondary morphologies and synaptic connections. How these segment-specific differences arise during embryogenesis remains unknown.