Serotonin-containing neurosecretory neurons in the first abdominal ganglion (A1 5-HT cells) of the lobster (Homarus americanus) ventral nerve cord have been shown previously to function as ‘gain setters’ in postural, slow muscle, command neuron circuitries. Here we show that these same amine neurons receive excitatory input from lateral (LG) and medial (MG) giant axons, which are major interneurons in phasic, fast muscle systems. Activation of either LG or MG axons elicits short-latency, non-fatiguing, long-lasting excitatory postsynaptic potentials (EPSPs) in A1 5-HT cells which follow stimulus frequencies of up to 100 Hz in a 1:1 fashion. Single spikes triggered in either giant axon can produce EPSPs in the A1 5-HT cells of sufficient magnitude to cause the cells to spike and to fire additional action potentials after variable latencies; action potentials elicited in this way reset the endogenous spontaneous spiking rhythm of the A1 5-HT neurons. The giant-axon-evoked EPSP amplitudes show substantial variation from animal to animal. In individual preparations, the variation of EPSP size from stimulus to stimulus was small over the first 25 ms of the response, but increased considerably in the later, plateau phase of each response. When tested in the same preparation, EPSPs in A1 5-HT cells evoked by firing the LG axons were larger, longer-lasting and more variable than those triggered by firing the MGs. Firing A1 5-HT cells through an intracellular electrode, prior to activation of the giant fiber pathway, significantly reduced the size of LG-evoked EPSPs in A1 5-HT cells. Finally, morphological and physiological results suggest that similarities exist between giant fiber pathways in lobsters and crayfish. The possible functional significance of an involvement of these large amine-containing neurosecretory neurons in both tonic and phasic muscle circuitries will be discussed.

The amine serotonin (5-HT) has been suggested to play key roles in the regulation of a wide and diverse array of important physiological processes in both invertebrate and vertebrate nervous systems, including man (Kravitz, 1988; Bicker and Menzel, 1989; Coccaro, 1989; Coccaro and Murphy, 1990; Raleigh et al. 1991; Miczek et al. 1994; Moskowitz and Waeber, 1996). Until very recently, it has proved difficult in vertebrate systems to unravel the precise roles served by biogenic amines in such processes, mostly because complex behaviors involve many neurons in large areas of the brain and spinal cord, and because individual amine neurons concerned with these behaviors have been difficult to find, to identify positively and to record from [but see Potrebic et al. 1994 (serotonin); Ashton-Jones et al. 1994 (norepinephrine); Schultz, 1992 (dopamine)].

Invertebrate organisms, with their far fewer uniquely identifiable neurons, offer particular advantages in this regard (Bicker and Menzel, 1989; Harris-Warrick, 1985; Harris-Warrick and Marder, 1991; Lent et al. 1991; Bailey and Kandel, 1993). With amine neurons, for example, comparative immunocytochemical studies in different invertebrate species have shown that amine effects characteristically are mediated by small numbers of easily found neurons (Beltz and Kravitz, 1983, 1987; Nässel, 1988; Spörhase-Eichmann et al. 1992; Seyfarth et al. 1993; Gilchrist et al. 1995; Hörner et al. 1995, 1996).

In crustacean species, amine neuron systems have been mapped completely (5-HT: Beltz and Kravitz, 1983, 1987; Real and Czternasty, 1990; octopamine: Schneider et al. 1993, 1996). Methods have been elaborated (i) to identify individual amine neurons, (ii) to define their physiological, functional and morphological characteristics and (iii) to record from the same cells repeatedly from different preparations (Beltz and Kravitz, 1987; Ma et al. 1992; Weiger and Ma, 1993; Bräunig et al. 1994; Anderson et al. 1996). This allows examination of the functions of these amine-containing neurons at levels not yet approachable in studies of amine neuron function in higher forms.

Among the most prominent of the approximately 120 serotonergic neurons found in the nervous systems of lobsters and crayfish are a pair of neurosecretory cells in the first abdominal ganglion (Beltz and Kravitz, 1983, 1987; Real and Czternasty, 1990). These cells (A1 5-HT) are spontaneously active (Beltz and Kravitz, 1987; Ma et al. 1992; Anderson et al. 1996) and have central and peripheral sets of endings (Beltz and Kravitz, 1987) which influence central circuitries concerned with postural regulation (Livingstone et al. 1981; Harris-Warrick and Kravitz, 1984; Harris-Warrick, 1985; Ma et al. 1992) and peripheral targets such as exoskeletal muscles (Glusman and Kravitz, 1982; Kravitz, 1988), sensory neurons (Pasztor and Bush, 1987) and the heart (Cooke and Hartline, 1975; Battelle and Kravitz, 1978). The A1 5-HT neurons function as generalized ‘gain-setters’ for postural circuitries, amplifying the output of command neurons concerned with posture (Ma et al. 1992). Flexor command neurons excite the amine cells, while extensor command neurons inhibit their firing. Firing individual amine neurons through intracellular electrodes has no effect on the motor output from the central nervous system, in contrast to the effects of superfused 5-HT, which activates a motor pattern for flexion (Harris-Warrick and Kravitz, 1984; Kravitz, 1988; Ma et al. 1992). In contrast, the motor output triggered by flexor command activation is enhanced by firing 5-HT cells, either via the flexor command excitation of the cell or via activation through the intracellular electrode (Ma et al. 1992).

The A1 5-HT neurons are also excited by electrical stimulation of peripheral nerve roots that contain axons of primary sensory afferents. Such stimuli are likely to activate mechanosensory interneurons which, in turn, will generate postural flexion reflexes and excite abdominal postural command systems (Kotak and Page, 1987; Weiger and Kravitz, 1994). In crayfish, similar nerve root stimulation activates the tailflip escape response by exciting the rapidly conducting segmental lateral giant (LG) interneurons (Krasne, 1969; Zucker, 1972; Wine and Krasne, 1982), which are the key interneurons involved in triggering this rapid behavior. The physiology of tailflip escape circuits involving the ascending lateral giant (LG) or the descending medial giant (MG) axons has not yet been described in lobsters, but the close similarity between the two species suggests that parallel circuits may exist. The observation that, in lobsters, stimulation of abdominal nerve roots that contain sensory fibers triggers EPSPs in A1 5-HT cells, and the possibility that the same stimuli may also activate LG axons, as seen in crayfish, prompted us to ask whether some of the excitation elicited by nerve root stimulation might be mediated by the LGs. If so, then the A1 5-HT cells may play important roles in phasic, fast muscle escape systems, in addition to their already demonstrated ‘gain setter’ role in the tonic, slow muscle systems. Testing this hypothesis was one of the major goals of the present study. In addition, we have examined the effects of MG stimulation on the A1 5-HT cells, since the MG interneurons also serve central roles in escape circuitry.

The results show that synaptic responses and action potentials can be triggered in A1 5-HT cells by activation of both the LG and MG giant fiber pathways. The amine cells thus appear to be linked to both the phasic and the tonic muscle systems, possibly coordinating serotonergic modulation between these two functionally diverse motor systems. A diagram of how the amine neurons fit into these circuitries will be presented. The present study also demonstrates the existence of anatomical and physiological similarities between the LG and MG axon systems in crayfish and lobsters.

Animals

Adult lobsters (Homarus americanus L.) of both sexes weighing approximately 0.5 kg were purchased from a local commercial supplier. Animals were maintained in artificial sea water on a 13 h:11 h light:dark cycle and fed regularly. Up to 10 animals were kept together in a group tank (400 l capacity) at a water temperature of 12–16 °C. Only specimens with intact appendages and sense organs were selected for physiological experiments. Lobsters with soft shells, indicating a pre-or post-molt state, were not used for electrophysiology.

Surgery and dissection

Prior to dissection, animals were anesthetized by packing them in ice for 30–60 min. After removal of the appendages and ventral cuticle, animals were placed for dissection in ice-cold saline of the following composition: 465 mmol l−1 NaCl, 16 mmol l−1 KCl, 26 mmol l−1 CaCl2, 8 mmol l−1 MgCl2 and 11 mmol l−1 glucose, buffered with 10 mmol l−1 Hepes adjusted to pH 7.4 (Otsuka et al. 1967; Harris-Warrick and Kravitz, 1984). In a typical experiment, the ganglia of thoracic segments 3–5 (T3–T5) and the complete abdominal ganglionic chain (A1–A6) were removed as a unit while taking care to leave undamaged long lengths of the attached segmental sensory and motor nerves. In some cases, the entire ventral nerve cord including the supraoesophageal (brain) ganglion and suboesophageal ganglion was dissected.

After dissection, nerve cords were pinned in an elastomer-coated (Sylgard, Dow Corning) Lucite recording chamber and continuously superfused with aerated saline. The saline temperature was kept between 10 and 16 °C by a thermoelectric cooling element. The A1 ganglion, which contains the 5-HT cells examined in the present study, was pinned ventral side up. All the other ganglia and connectives were pinned dorsal side up. The sheath was removed from the A1 ganglion by dissection, and the layers of glial cells and adhering connective tissue were washed away with a stream of saline to expose the neuronal somata. The glial and connective tissue sheath that surrounds axons in the interganglionic connectives was also removed. To enable intracellular recording and stimulation of these axons, the connectives between T4 and T5 were treated with a collagenase/dispase solution (2 mg ml−1 each in saline; Sigma). A small Vaseline trough surrounding the connectives between T4 and T5 prevented diffusion of the enzyme solution to other parts of the ventral nerve cord. After enzyme incubation, preparations were thoroughly rinsed with saline. The exposed somata and axons were visualized using dark-field optics.

Electrophysiological recording techniques

Intracellular recordings were performed with glass microelectrodes filled with 3 mol l−1 potassium chloride or 2 mol l−1potassium acetate. The electrode resistance ranged from 6–20 MΩ. Electrical signals were amplified using an Axoprobe 1A multipurpose amplifier (Axon Instruments). Data were digitized using the TL-1 DMA interface in conjunction with the pClamp5.5 software package (Axon Instruments) and stored directly on a microcomputer or recorded on tape (Hewlett Packard) for off-line analysis. 5-HT-containing neurosecretory neurons (A1 5-HT) and fast flexor motoneuron cell bodies in the A1 ganglion were identified using criteria defined in earlier publications (see below and Otsuka et al. 1967; Beltz and Kravitz, 1983, 1987; Ma et al. 1992).

For intracellular stimulation of single axons within the connectives, a high-voltage headstage (10× headstage, Axon Instruments) was used. Stimulation and spike generation in single axons was also monitored using extracellular electrodes (see below). For quantitative data evaluation, commercially available software packages were used (pClamp, Axon Instruments; Origin™, Microcal). For statistical analysis, parametric and non-parametric tests (t-test, Mann–Whitney U-test, Kolmogorov–Smirnov test) with a critical probability of P<0.05 were used.

Stimulation of and recording from nerve roots and connectives

For extracellular recording of electrical activity from nerve roots, suction electrodes of different diameters were made from fine polyethylene tubing. The cut ends of the third nerve roots of abdominal ganglia (Thompson and Page, 1982; Ma et al. 1992) were sucked directly into electrodes for recording. Extracellular electrical signals were amplified using a differential amplifier (A-M Systems model 1700), which allowed recording and stimulation through the same electrode. To record extracellular spikes from single axons in connectives, suction electrodes were placed on the desheathed surface of the connectives. The extracellular signals were digitized and analyzed as described above.

For extracellular stimulation, the same electrodes were used as for recording, and single stimuli or trains of pulses were delivered to ganglionic nerves and to interganglionic connectives (WPI model 850A stimulus isolating units in conjunction with model 831 pulse modules and a model 830 interval generator) while monitoring the effects on synaptic activity in A1 5-HT cells and fast flexor motoneurons. The stimulation of single giant axons was accomplished most reliably using bipolar tungsten needle electrodes (tip distance 200 μm, tip diameter 50 μm), which were insulated except at their tips. These electrodes were positioned visually on the surface of a giant axon. Single stimuli of 200–500 μs duration or pulse trains of up to 200 Hz were used.

Fluorescent dye injection

To identify A1 5-HT cells positively, the fluorescent dye Lucifer Yellow CH (5 % in 0.1 mol l−1 LiCl; Sigma; Stewart, 1978) was ionophoretically injected through the recording electrode using continuous negative currents of up to 5 nA for up to 3 h (Beltz and Kravitz, 1983, 1987).

Giant axons were identified by their location on the dorsal surface of connectives and by electrophysiological tests before dye injection (see Results). The large volumes of these giant axons (diameters up to 100 μm) required that they be labeled by pressure injection of fluorescent dyes through broken microelectrodes (tip diameter 5–10 μm). Both Lucifer Yellow CH and tetramethylrhodamine-coupled dextran (Mr 3000, Molecular Probes, Eugene, Oregon, USA; 5 % in lobster saline) were used for this purpose.

Fluorescent labeling was viewed using epifluorescent illumination in a Zeiss Axiophot microscope equipped with filter sets appropriate for either Lucifer Yellow CH or Rhodamine-induced fluorescence. The results were documented by photography with high-definition color film (Kodachrome).

Physiological identification of neuron types

These experiments aim to unravel key elements of the circuitry involving the 5-HT-containing neurons of the first abdominal ganglion (A1 5-HT) of the lobster, which have been the focus of previous detailed studies (Beltz and Kravitz, 1983, 1987; Ma et al. 1992; Ma and Weiger, 1993; Weiger and Ma, 1993; Weiger and Kravitz, 1994). In the first experiments, criteria are defined (i) for the reliable identification of the A1 5-HT cells, and (ii) for the morphological and functional identification of important interneurons involved in their activation.

Identification of the A1 5-HT cells

These neurons have their somata located in an anterior lateral position in the first abdominal ganglia of the ventral nerve cord. They are approximately 100 μm in diameter and their cytoplasm often has a slightly opaque appearance, making the identification of candidate neurons relatively straightforward. In the experiments presented in this paper, upon penetration of their somata with an intracellular electrode, the 5-HT neurons have resting potentials of −45 to −55 mV, which remain stable for up to 6 h. The cells characteristically have large overshooting action potentials (40–70 mV) with long-lasting (10–15 ms) hyperpolarizing afterpotentials and show spontaneously occurring spikes at a frequency of 0.5–1.5 Hz (Ma et al. 1992). A continuous barrage of small (0.4–1 mV) spontaneous excitatory and inhibitory postsynaptic potentials (IPSPs) is also regularly seen (Weiger and Ma, 1993).

In certain of the experiments, after the physiological identification of a neuron had been completed, ionophoretic injection of Lucifer Yellow CH was used to confirm the morphological features of the cell (Beltz and Kravitz, 1983, 1987). The labeling of a pair of A1 5-HT cells in a ganglion revealed their bilaterally symmetrical shape and showed their axons curving dorso-laterally and ascending on the side of the connective ipsilateral to the soma towards the thoracic ganglia. A characteristic pattern of fine varicose endings extends medially from the axonal branches of the A1 5-HT cells, but processes of the cells never cross the ganglionic midline. A1 5-HT neurons were positively identified in this way in two-thirds of all preparations.

Identification of giant axons

The MG and LG axons were identified using an array of stimulating and recording electrodes, and a similar experimental protocol in all experiments (Fig. 1). First, a stimulating electrode was placed on the surface of a putative MG or LG axon at the anterior margin of the A4 ganglion, and a recording electrode was placed on a large medial or lateral axon at the posterior end of the T4 ganglion. Test pulses of increasing amplitude were delivered at the A4 site. If either the MG or LG was successfully activated at A4, a large extracellular spike of 2–5 mV amplitude was observed at T4 within 7–8 ms after the onset of the stimulus (conduction velocity 8.5–10 m s−1 at 15 °C). These were the fastest spikes observed in the lobster nerve cord under the chosen recording conditions. The MG is the easier axon to find, since only one pair of large axons is seen in a medial location in connectives, while two or three large axon pairs are regularly seen in more lateral locations. Stimulation of non-giant fibers results in slower conduction velocities (less than 6 m s−1) and the recording of smaller extracellular spikes at the upstream electrode. The identification of the MG and LG axons was further confirmed by finding (i) short-latency EPSPs in the cell bodies of identified fast flexor motoneurons accompanying their activation (see Otsuka et al. 1967), and (ii) triggered phase-locked spikes in the third nerve roots of the A1 ganglion (see Thompson and Page, 1982; Harris-Warrick and Kravitz, 1984; Ma et al. 1992; see below).

Fig. 1.

Schematic representation of the experimental design used for the identification of synaptic interactions between the lateral (LG) and medial (MG) giant interneurons and 5-HT-containing neurons in the first abdominal ganglion (for further explanation see text). T3–T5, ganglia of the third to the fifth thoracic segments; A1–A6, ganglia of the first to the sixth abdominal segments. C nerve, the third nerve root from anterior.

Fig. 1.

Schematic representation of the experimental design used for the identification of synaptic interactions between the lateral (LG) and medial (MG) giant interneurons and 5-HT-containing neurons in the first abdominal ganglion (for further explanation see text). T3–T5, ganglia of the third to the fifth thoracic segments; A1–A6, ganglia of the first to the sixth abdominal segments. C nerve, the third nerve root from anterior.

After completing the preliminary identification using extracellular electrodes, giant axons were impaled with an intracellular recording electrode close to the anterior margin of the T5 ganglion. Under these recording conditions, when activated from the A4 electrode, the MG and LG axons had resting potentials between −70 and −78 mV and showed large overshooting action potentials (80–110 mV) without hyperpolarizing afterpotentials. No spontaneous spiking or synaptic activity was observed in MG or LG axons in the isolated nerve cords. The identified axons were stimulated using the intracellular electrode, and spikes were monitored either at the original A4 stimulating site or at the T4 recording site. The conduction velocities were the same in both the ortho- and antidromic directions. The above procedure allowed an unequivocal identification of the two types of giant fibers and readily distinguished giant and non-giant axons. To confirm their identification fully, in some experiments the giant axons were pressure-injected with fluorescent dyes. Two different dyes (Lucifer Yellow CH and Rhodamine coupled to dextran) were used to allow identification of pairs of giant fibers in the same preparation and to observe their major branching patterns.

Lateral giants

As in the crayfish (Remler et al. 1968; Wine and Krasne, 1982), the LG system of the lobster consists of separate, segmentally arranged neurons. Injection of Rhodamine-coupled dextran showed that LG axons (diameter 70–90 μm), as expected, run in a lateral position to MG fibers (Fig. 2A–D), and have neurites that cross the ganglionic midline where they form a dorsal–medial chiasm and project rostrally to the next anterior ganglion, terminating in the posterior half of the ganglion along a contact zone with the LG from that ganglion (Fig. 2A–C). The contact zones between the LG axons from adjacent segments were identified by injection of a different fluorescent dye into each axon. Of the two dyes used, Lucifer Yellow CH eventually (after 20–30 min) crossed the synaptic contact zone and stained the LG axon of the adjacent segment. Rhodamine-coupled dextran did not cross this junction. While the identification of LG axons was somewhat more difficult than the identification of the MGs, owing to variations in LG localization within the connective and the presence of other lateral axons of comparable size, the morphological and physiological tests performed allowed their unequivocal identification. No other large axon in the vicinity of the LGs showed the unique morphological or physiological properties described above.

Fig. 2.

Morphological identification of lateral (LG) and medial (MG) giant interneurons in the first abdominal ganglion (A1) of the ventral nerve cord (whole-mount preparations; anterior is at the top in each photograph). (A) Differential labeling of LG axons of two adjacent segments within A1 (the anterior LGs are labeled with Rhodamine-coupled dextran, the posterior LGs are filled with Lucifer Yellow CH). (B) Anterior portion of both segmental LGs viewed after Rhodamine-coupled dextran injection. (C) Detailed view of the contact zone between segmentally homologous LGs. (D) Double labeling of the paired medial MGs (injected with Lucifer Yellow CH) and the anterior portion of the LG (injected with Rhodamine-coupled dextran) in A1. (E) A pair of MG axons at the posterior margin of the A1 ganglion labeled with Lucifer Yellow CH. Scale bars, 200 μm.

Fig. 2.

Morphological identification of lateral (LG) and medial (MG) giant interneurons in the first abdominal ganglion (A1) of the ventral nerve cord (whole-mount preparations; anterior is at the top in each photograph). (A) Differential labeling of LG axons of two adjacent segments within A1 (the anterior LGs are labeled with Rhodamine-coupled dextran, the posterior LGs are filled with Lucifer Yellow CH). (B) Anterior portion of both segmental LGs viewed after Rhodamine-coupled dextran injection. (C) Detailed view of the contact zone between segmentally homologous LGs. (D) Double labeling of the paired medial MGs (injected with Lucifer Yellow CH) and the anterior portion of the LG (injected with Rhodamine-coupled dextran) in A1. (E) A pair of MG axons at the posterior margin of the A1 ganglion labeled with Lucifer Yellow CH. Scale bars, 200 μm.

In crayfish, the paired segmental LG neurons in each abdominal ganglion are electrotonically and dye-coupled across the ganglion through non-rectifying gap junctions (Watanabe and Grundfest, 1961; Wine and Krasne, 1982). In lobsters, the presumably homologous pairs of LG neurons in each segment show neither type of coupling. Action potentials initiated in an LG axon on one side of a ganglion propagate along the chain of LG neurons on that side of the ganglion only and do not excite the contralateral LG axons. Injection of Lucifer Yellow CH into one LG never leads to staining of the contralateral LG.

Medial giants

The MG axons were easily identified by their size (90–100 μm) and by their location. Dye injection showed that the MGs were the largest axons in the medial region of thoracic connectives. In contrast to the LGs, fluorescent dyes in the MG axons could be traced from their injection site to distances well beyond the adjacent ganglia, and the axons crossed segmental borders between ganglia without interruption. Within each ganglion, the pairs of MG axons ran close to each other, but remained separate throughout their intrasegmental course (Fig. 2D,E).

Synaptic input to the A1 5-HT cells from LG and MG axons

In first attempts to identify the sources of synaptic input to A1 5-HT cells, severed connectives and segmental nerve roots were stimulated. In all cases, mixed excitatory and inhibitory synaptic responses were seen in the 5-HT neuron cell bodies (Weiger and Kravitz, 1994; M. Hörner, unpublished results). The largest and most reliable excitatory effect was produced by stimulating sensory nerve roots of the terminal (A6) abdominal ganglion. A single shock to the C nerve of the A6 ganglion (i.e. the third nerve root from anterior, see Fig. 1) evoked a compound EPSP in A1 5-HT cells, which caused A1 5-HT cells to fire (Fig. 3A). In attempting to track down the interneurons mediating that excitation, we found that extracellular stimulation of large axons on the dorsal surface of connectives could elicit short-latency EPSPs in the 5-HT cells, resembling those triggered by C root stimulation. These results suggested that the LG or MG axons might provide synaptic input to 5-HT-containing cells in the A1 ganglion.

Fig. 3.

Response of a 5-HT-containing neuron in the first abdominal ganglion (A1 5-HT) to extracellular stimulation. (A) Stimulation of the C nerve (i.e. the third nerve root) in the A6 ganglion. After a long latency, a single stimulus to the C root evokes a compound EPSP which triggers an action potential in the A1 5-HT cell. The arrow indicates the time of stimulation of the C nerve root in A6; the peak and hyperpolarizing phase of the action potential are truncated. (B) Extracellular stimulation of the LG axon anterior to A4 (see Fig. 1) elicits an EPSP followed by an action potential in the A1 5-HT cell ipsilateral to the LG axon. Upper trace, intracellular recording from the soma of A1 5-HT; middle trace, intracellular recording from the LG axon anterior to T5; lower trace, extracellular recording of the LG spike posterior to T4. The arrow indicates the time of stimulation of the LG.

Fig. 3.

Response of a 5-HT-containing neuron in the first abdominal ganglion (A1 5-HT) to extracellular stimulation. (A) Stimulation of the C nerve (i.e. the third nerve root) in the A6 ganglion. After a long latency, a single stimulus to the C root evokes a compound EPSP which triggers an action potential in the A1 5-HT cell. The arrow indicates the time of stimulation of the C nerve root in A6; the peak and hyperpolarizing phase of the action potential are truncated. (B) Extracellular stimulation of the LG axon anterior to A4 (see Fig. 1) elicits an EPSP followed by an action potential in the A1 5-HT cell ipsilateral to the LG axon. Upper trace, intracellular recording from the soma of A1 5-HT; middle trace, intracellular recording from the LG axon anterior to T5; lower trace, extracellular recording of the LG spike posterior to T4. The arrow indicates the time of stimulation of the LG.

EPSPs triggered by LG or MG spikes

Extracellular stimulation of an LG axon evoked a spike that propagated along the ipsilateral chain of LG axons in either direction at 8.5–9 m s−1 (MG 9–10 m s−1) and that evoked depolarizing postsynaptic potentials (PSPs) in A1 5-HT cells which usually triggered action potentials (Fig. 3B). With extracellular stimulation of the LG, reliable and selective firing of the LG axon alone could not be achieved. Therefore, intracellular stimulation techniques were used to define the effect of single LG spikes on A1 5-HT cells (see Fig. 1). In contrast to the PSPs evoked by extracellular stimulation, which occasionally included inhibitory components, intracellular LG or MG stimulation triggered unitary PSPs without any sign of inhibition (Fig. 4A,B). EPSPs were detectable in the ipsilateral, but not the contralateral, of the paired A1 5-HT cells, 6–7 ms after the generation of LG or MG spikes by stimulating with either the intracellular T5 or the extracellular A4 electrodes. Thus, the central delay between the arrival of an LG or MG spike at A1 and the onset of the corresponding EPSP in the A1 5-HT neuron is 2–3 ms. The EPSPs showed a rising phase that reached its half-maximum height after 10–14 ms (Fig. 4A,B; Table 1). Action potentials often arose from the EPSPs triggered by single LG spikes, but the latencies were quite variable (range 60–200 ms). Repetitive stimulation of LG or MG axons led to summation of EPSPs and caused spiking in the A1 5-HT cells within 30 ms after the stimulus onset (results not shown). Both the LG (and MG) spikes and the EPSPs evoked in the A1 5-HT cells followed short trains (30 ms eliciting four spikes in LG or MG) of intracellular stimulation at frequencies of up to 100 Hz without signs of fatigue or potentiation.

Table 1.

Amplitudes (LG/MG25–100) and latencies (LGT50/MGT50) to half-maximum amplitude for LG- and MG-evoked EPSPs in 5-HT-containing neurons in the first abdominal ganglion

Amplitudes (LG/MG25–100) and latencies (LGT50/MGT50) to half-maximum amplitude for LG- and MG-evoked EPSPs in 5-HT-containing neurons in the first abdominal ganglion
Amplitudes (LG/MG25–100) and latencies (LGT50/MGT50) to half-maximum amplitude for LG- and MG-evoked EPSPs in 5-HT-containing neurons in the first abdominal ganglion
Fig. 4.

EPSPs evoked in the 5-HT-containing cell in the first abdominal ganglion (A1 5-HT) by spikes in the lateral (LG) and medial (MG) giant interneuron. (A) Intracellular stimulation of the LG axon at T5 elicits long-lasting EPSPs in the A1 5-HT neuron. The LG spikes recorded at A4 are correlated 1:1 with the EPSPs in the A1 5-HT cell. Upper traces, intracellular recordings of 10 superimposed LG-evoked EPSPs in A1 5-HT; bold line, mean time course of the EPSPs; error bars show the standard deviation of the EPSP amplitude calculated 25 ms (±0.17 mV), 50 ms (±0.22 mV), 75 ms (±0.27 mV) and 100 ms (±0.42 mV) after the stimulus; the averaged spontaneous PSP amplitude in A1 5-HT is ±0.18 mV; lower trace, extracellular recording of the LG spike anterior to A4. (B) Intracellular stimulation of the MG axon at T5 in a different preparation (compare with A; error bars are the standard deviation of the EPSP amplitude calculated 25 ms (±0.14 mV), 50 ms (±0.20 mV), 75 ms (±0.27 mV) and 100 ms (±0.35 mV) after the stimulus; the averaged spontaneous PSP amplitude in A1 5-HT is ±0.18 mV; lower trace, extracellular recording of the MG spike anterior to A4. Other details as in A.

Fig. 4.

EPSPs evoked in the 5-HT-containing cell in the first abdominal ganglion (A1 5-HT) by spikes in the lateral (LG) and medial (MG) giant interneuron. (A) Intracellular stimulation of the LG axon at T5 elicits long-lasting EPSPs in the A1 5-HT neuron. The LG spikes recorded at A4 are correlated 1:1 with the EPSPs in the A1 5-HT cell. Upper traces, intracellular recordings of 10 superimposed LG-evoked EPSPs in A1 5-HT; bold line, mean time course of the EPSPs; error bars show the standard deviation of the EPSP amplitude calculated 25 ms (±0.17 mV), 50 ms (±0.22 mV), 75 ms (±0.27 mV) and 100 ms (±0.42 mV) after the stimulus; the averaged spontaneous PSP amplitude in A1 5-HT is ±0.18 mV; lower trace, extracellular recording of the LG spike anterior to A4. (B) Intracellular stimulation of the MG axon at T5 in a different preparation (compare with A; error bars are the standard deviation of the EPSP amplitude calculated 25 ms (±0.14 mV), 50 ms (±0.20 mV), 75 ms (±0.27 mV) and 100 ms (±0.35 mV) after the stimulus; the averaged spontaneous PSP amplitude in A1 5-HT is ±0.18 mV; lower trace, extracellular recording of the MG spike anterior to A4. Other details as in A.

After the first 30–40 ms, the LG-evoked EPSPs reached a plateau that lasted for at least 100 ms and that, in some cases, did not decline to the baseline for at least another 100 ms (Fig. 4A). MG-evoked EPSPs lasted up to 150 ms and slowly declined to baseline levels with less of a plateau than that seen with LG-evoked EPSPs. The LG- and MG-evoked EPSPs showed little variation in amplitude during the rising phase and up to the plateau (Fig. 4A,B). But during the plateau phase, considerable variability in EPSP amplitudes was seen, often exceeding the level of spontaneously occurring membrane potential fluctuations (±0.18 mV) in A1 5-HT neurons observed after subthreshold stimulation of LG or MG (see also legend to Fig. 4 for quantitative data). Despite this variability, which presumably results from interactions of the giant-fiber-evoked EPSPs with spontaneous intrinsic membrane fluctuations in A1 5-HT (see Figs 7, 8), LG- and MG-evoked EPSPs appear to be unitary throughout their duration.

Fig. 5.

Effects of LG stimulation on motoneurons. (A) Intracellular stimulation of LG elicits EPSPs of short duration in an identified M7 motoneuron in the A1 ganglion in a 1:1 fashion. Upper traces, four superimposed intracellular recordings of EPSPs in M7; dotted line, average time course of four EPSPs with error bars indicating the standard deviation calculated 7 ms (±0.11 mV), 25 ms (±0.06 mV) and 50 ms (±0.06 mV) after stimulation; the averaged spontaneous PSP amplitude in M7 is ±0.05 mV; middle trace, extracellular recording of the LG spike anterior to A4; lower trace, extracellular recording of the LG spike posterior to T4. (B) Intracellular stimulation of LG at T5 elicits EPSPs in an A1 5-HT neuron and spiking of a fast flexor motoneuron recorded in the third nerve root between the A3 and A4 ganglion. Upper trace, mean time course of eight EPSPs elicited in an A1 5-HT neuron by intracellular stimulation of LG; middle trace, extracellular recording of the LG spike at A4; lower traces, eight superimposed extracellular recordings of fast flexor motoneuron activity in the third nerve root between A3/4.

Fig. 5.

Effects of LG stimulation on motoneurons. (A) Intracellular stimulation of LG elicits EPSPs of short duration in an identified M7 motoneuron in the A1 ganglion in a 1:1 fashion. Upper traces, four superimposed intracellular recordings of EPSPs in M7; dotted line, average time course of four EPSPs with error bars indicating the standard deviation calculated 7 ms (±0.11 mV), 25 ms (±0.06 mV) and 50 ms (±0.06 mV) after stimulation; the averaged spontaneous PSP amplitude in M7 is ±0.05 mV; middle trace, extracellular recording of the LG spike anterior to A4; lower trace, extracellular recording of the LG spike posterior to T4. (B) Intracellular stimulation of LG at T5 elicits EPSPs in an A1 5-HT neuron and spiking of a fast flexor motoneuron recorded in the third nerve root between the A3 and A4 ganglion. Upper trace, mean time course of eight EPSPs elicited in an A1 5-HT neuron by intracellular stimulation of LG; middle trace, extracellular recording of the LG spike at A4; lower traces, eight superimposed extracellular recordings of fast flexor motoneuron activity in the third nerve root between A3/4.

Fig. 6.

Diagram showing the ranges and mean amplitudes of LG-(A) and MG-evoked (B) EPSPs in 5-HT-containing neurons in the first abdominal ganglion analyzed for the entire (LG, N=7; MG, N=5) population of animals tested. Hatched columns, minimum EPSP amplitude; middle columns, mean EPSP amplitudes with error bars showing the standard deviation; cross-hatched columns, maximum EPSP amplitude; the lower axis shows the time (in ms) after stimulation when the EPSPs were analyzed.

Fig. 6.

Diagram showing the ranges and mean amplitudes of LG-(A) and MG-evoked (B) EPSPs in 5-HT-containing neurons in the first abdominal ganglion analyzed for the entire (LG, N=7; MG, N=5) population of animals tested. Hatched columns, minimum EPSP amplitude; middle columns, mean EPSP amplitudes with error bars showing the standard deviation; cross-hatched columns, maximum EPSP amplitude; the lower axis shows the time (in ms) after stimulation when the EPSPs were analyzed.

Fig. 7.

Effects of EPSPs evoked by single LG spikes on the spontaneous firing activity in 5-HT-containing neurons in the first abdominal ganglion (A1 5-HT). (A) Firing of LG reduces the time interval between a spontaneous and an evoked spike in an A1 5-HT neuron and increases the interval in the following cycle. Upper trace (Spont. interval), time between two spontaneous spikes in an A1 5-HT neuron; n, time between two spikes in A1 5-HT during stimulation of LG; n+1, time between two spikes in A1 5-HT in the immediately following period; lower trace, extracellular recording of LG spike at A4. (B) Diagram showing the effects of the timing of the LG spike on the spontaneous activity pattern in A1 5-HT. Solid line, mean time interval between spontaneously occurring spikes in A1 5-HT with dotted lines indicating the standard deviation; filled circles, time interval n; open squares, time interval n+1, as explained in A above. (C) Statistical evaluation of the effects of single spikes in LG on the spike intervals in an A1 5-HT neuron. Compared with spontaneously occurring action potentials (Spont.), spikes in LG significantly reduce (n; P<0.05, Mann–Whitney U-test) or increase (n+1; P<0.05, Mann–Whitney U-test) the interval between two consecutive spikes. Error bars, standard deviations of the intervals; spont, 90 intervals averaged; n, 18 intervals averaged between 100 and 500 ms delay; n+1, 66 intervals averaged.

Fig. 7.

Effects of EPSPs evoked by single LG spikes on the spontaneous firing activity in 5-HT-containing neurons in the first abdominal ganglion (A1 5-HT). (A) Firing of LG reduces the time interval between a spontaneous and an evoked spike in an A1 5-HT neuron and increases the interval in the following cycle. Upper trace (Spont. interval), time between two spontaneous spikes in an A1 5-HT neuron; n, time between two spikes in A1 5-HT during stimulation of LG; n+1, time between two spikes in A1 5-HT in the immediately following period; lower trace, extracellular recording of LG spike at A4. (B) Diagram showing the effects of the timing of the LG spike on the spontaneous activity pattern in A1 5-HT. Solid line, mean time interval between spontaneously occurring spikes in A1 5-HT with dotted lines indicating the standard deviation; filled circles, time interval n; open squares, time interval n+1, as explained in A above. (C) Statistical evaluation of the effects of single spikes in LG on the spike intervals in an A1 5-HT neuron. Compared with spontaneously occurring action potentials (Spont.), spikes in LG significantly reduce (n; P<0.05, Mann–Whitney U-test) or increase (n+1; P<0.05, Mann–Whitney U-test) the interval between two consecutive spikes. Error bars, standard deviations of the intervals; spont, 90 intervals averaged; n, 18 intervals averaged between 100 and 500 ms delay; n+1, 66 intervals averaged.

Fig. 8.

Effects of EPSPs evoked by single MG spikes on the spontaneous firing activity in 5-HT-containing neurons in the first abdominal ganglion (A1 5-HT; see legend of Fig. 7 for details). (A) Firing of MG reduces the time interval between a spontaneous and an evoked spike in an A1 5-HT neuron. (B) Diagram showing the effects of the timing of the MG spike on the spontaneous activity pattern in A1 5-HT. (C) Statistical evaluation of the effects of single spikes in MG on the spike intervals in an A1 5-HT neuron. Compared with spontaneously occurring action potentials in A1 5-HT cells (Spont.), spikes in MG significantly reduce the interval between two consecutive spikes (n; P<0.05 Mann–Whitney U-test), but no significant effects were found in the interval following MG stimulation (n+1). Error bars, standard deviations of the intervals; Spont, 98 intervals averaged; n, 27 intervals averaged between 100–500 ms delay; n+1, 47 intervals averaged.

Fig. 8.

Effects of EPSPs evoked by single MG spikes on the spontaneous firing activity in 5-HT-containing neurons in the first abdominal ganglion (A1 5-HT; see legend of Fig. 7 for details). (A) Firing of MG reduces the time interval between a spontaneous and an evoked spike in an A1 5-HT neuron. (B) Diagram showing the effects of the timing of the MG spike on the spontaneous activity pattern in A1 5-HT. (C) Statistical evaluation of the effects of single spikes in MG on the spike intervals in an A1 5-HT neuron. Compared with spontaneously occurring action potentials in A1 5-HT cells (Spont.), spikes in MG significantly reduce the interval between two consecutive spikes (n; P<0.05 Mann–Whitney U-test), but no significant effects were found in the interval following MG stimulation (n+1). Error bars, standard deviations of the intervals; Spont, 98 intervals averaged; n, 27 intervals averaged between 100–500 ms delay; n+1, 47 intervals averaged.

To compare the variability of LG-evoked EPSPs in A1 5-HT neurons with that seen in motoneuron targets of the LGs, evoked EPSPs were analyzed in fast flexor motoneurons and 5-HT cells from the same ganglion (Fig. 5A; see legend for quantitative data). LG spikes generated EPSPs in motoneurons that showed steep rise times (half maximum at 4–5 ms) and considerably higher amplitudes than those found in 5-HT cells (range 2–6 mV). In addition, the duration of the EPSPs seen in the motoneurons was approximately 30 ms and they showed little variation in amplitude with repeated stimulation of the LG (Fig. 5A). LG and MG stimulation also caused spiking in some fast flexor motoneurons, as indicated by the action potentials recorded from the third nerve roots between abdominal ganglia (shown for LG in the lowest trace of Fig. 5B).

Comparison of LG- and MG-evoked effects in A1 5-HT cells

The experiments described above demonstrate that the activation of either giant fiber pathway leads to parallel effects in the 5-HT-containing cells of the first abdominal ganglion. Evoked EPSPs were seen which were similar with respect to latency of onset, absolute size, duration and increasing variability of the plateau response over time in individual trials. In control experiments, other dorsally located large axons were impaled and activated intracellularly while recording from A1 5-HT cells. None of these other axons, whose diameters were only slightly smaller than the MG and LG axons, caused reproducible synaptic responses or influenced the spontaneous firing pattern of the A1 5-HT cells.

When tested in the same preparation, LG-evoked EPSP amplitudes were consistently larger than those resulting from MG stimulation (Table 1). In addition, the MG-evoked responses were shorter in duration and showed a different time course, decaying within 100 ms to approximately 70 % of their maximum size (Fig. 6A,B). A quantitative comparison of the giant-fiber-triggered EPSPs from all experiments showed significant differences between LG-(Fig. 6A) and MG-evoked (Fig. 6B) events. In the total population, the minimum and maximum EPSP amplitudes resulting from LG stimulation showed far greater variability than those resulting from MG activation. This suggests that there is considerably more variability in the synaptic coupling strength between LG and A1 5-HT cells than there is between MG and these same amine neurons.

Resetting the spontaneous firing rhythms of 5-HT cells

A1 5-HT neurons show spontaneous firing at frequencies of 0.5–1.5 Hz (Ma et al. 1992). The triggering of additional action potentials by synaptic input could interfere with and possibly reset the spontaneous rhythm in these cells. To investigate whether LG-or MG-evoked EPSPs influenced the timing of the intrinsic activity pattern in A1 5-HT cells, single spikes were elicited in LG and MG axons at random intervals following the spontaneously occurring 5-HT-cell action potentials. The LG and MG effects were quantified by comparing the interspike intervals between a spontaneous spike and the immediately following LG-or MG-evoked action potential (n) and between the LG-or MG-evoked action potential and the next spontaneous spike (n+1; Figs 7A, 8A). Phase response curves were constructed that show the interspike intervals between the giant-fiber-evoked spikes and the immediately preceding and following spontaneous action potentials and compare those intervals with the mean spontaneous spike interval (Figs 7B, 8B). The results show that single spikes in LG or MG can reset the spontaneous firing pattern in the 5-HT cells. The effects were most pronounced when the giant-fiber-triggered EPSP and spike occurred within the first 35 % of the spontaneous cycle length (for significance levels, see Figs 7C, 8C). The effects of driving LG were (i) a shortened interspike interval between a prior spontaneous action potential and the LG-generated spike (n), (ii) a lengthened interspike interval between the LG-mediated spike and the next spontaneous action potential (n+1), and (iii) a resetting of the spontaneous rhythm. In contrast to the results obtained with LG stimulation, however, the effects of MG stimulation were restricted to the interval of stimulation (n; Fig. 8A–C). No lengthening was seen in the interspike interval in the period following MG activation (n+1; see Fig. 8C, for significance levels).

Consequences of pre-firing A1 5-HT cells

It has been shown previously that increased firing of A1 5-HT cells modulates the effectiveness of command-fiber-evoked input to slow (tonic) flexor motoneurons (Ma et al. 1992). The LG is a key interneuron in the fast (phasic) system of motoneurons. Since the above experiments showed definitively that the LG axon activated the A1 5-HT cells, it seemed of interest to investigate whether the firing of the amine cells, in turn, could modulate the LG input in any way. To test this, the effectiveness of LG activation of a 5-HT cell with and without pre-firing of the amine cell was measured (Fig. 9A,B). LG-evoked EPSPs and spontaneous firing rates in A1 5-HT cells were measured before, and for 10 min after, the amine cells were fired continuously for 15 s at frequencies up to 20 Hz (Fig. 9A,B). The immediately visible effect of firing the A1 5-HT neurons was an interruption in their spontaneous firing (Fig. 9B, lower plot). The amine cells remained silent for up to 2 min and then, over the next 5 min, their rate of spontaneous firing slowly returned to the initial values. At the same time, the LG-evoked EPSPs were significantly reduced in size for up to 4 min after firing the A1 5-HT cells. The recovery rate of EPSP size to normal was faster than the reappearance of spontaneous spiking (Fig. 9B, upper plot). Evoked EPSPs showed reductions in their initial rate of rise and in their maximum amplitude for the first 3 min after firing the 5-HT neuron, both of which ordinarily vary little with repeated stimulation of LG axons (compare Figs 9A and 4A).

Fig. 9.

Effects of pre-stimulation of the 5-HT-containing neurons in the first abdominal ganglion (A1 5-HT) on the amplitude of LG-evoked EPSPs. (A) Pre-stimulation of the A1 5-HT cell leads to a reduction of LG-evoked EPSP amplitudes. Uppermost traces, three superimposed recordings of LG-evoked EPSPs; middle traces, three superimposed recordings of LG-evoked EPSPs in A1 5-HT after 15 s of continuous firing in the amine cell at 20 Hz; the bold lines show the mean amplitude of EPSPs under the different conditions; bottom trace, extracellular recording of the LG spike at T4. (B) The time course of the results of electrical stimulation of A1 5-HT on LG-evoked EPSP amplitudes and spontaneous spiking activity (two preparations; N=4 independent experiments). LG-evoked amplitudes and spontaneous spiking are significantly reduced for several minutes following intracellular stimulation of A1 5-HT. Filled squares, mean values of LG-evoked EPSP amplitudes calculated over time; error bars show the standard deviation; open circles, mean values of spontaneous spiking frequency with error bars showing standard deviations before and after electrical stimulation of the A1 5-HT neuron; arrow, beginning of electrical stimulation of the A1 5-HT neuron. On the time axis, negative values indicate the time prior to stimulation of the A1 5-HT neuron when control measurements were made.

Fig. 9.

Effects of pre-stimulation of the 5-HT-containing neurons in the first abdominal ganglion (A1 5-HT) on the amplitude of LG-evoked EPSPs. (A) Pre-stimulation of the A1 5-HT cell leads to a reduction of LG-evoked EPSP amplitudes. Uppermost traces, three superimposed recordings of LG-evoked EPSPs; middle traces, three superimposed recordings of LG-evoked EPSPs in A1 5-HT after 15 s of continuous firing in the amine cell at 20 Hz; the bold lines show the mean amplitude of EPSPs under the different conditions; bottom trace, extracellular recording of the LG spike at T4. (B) The time course of the results of electrical stimulation of A1 5-HT on LG-evoked EPSP amplitudes and spontaneous spiking activity (two preparations; N=4 independent experiments). LG-evoked amplitudes and spontaneous spiking are significantly reduced for several minutes following intracellular stimulation of A1 5-HT. Filled squares, mean values of LG-evoked EPSP amplitudes calculated over time; error bars show the standard deviation; open circles, mean values of spontaneous spiking frequency with error bars showing standard deviations before and after electrical stimulation of the A1 5-HT neuron; arrow, beginning of electrical stimulation of the A1 5-HT neuron. On the time axis, negative values indicate the time prior to stimulation of the A1 5-HT neuron when control measurements were made.

The reduction in EPSP size and the decrease in spontaneous firing rate in the A1 5-HT cells was accompanied by a slight membrane hyperpolarization (2–5 mV). The membrane potential returned to the pre-stimulus values along a similar time course to the recovery of spontaneous spiking. All of the effects observed (reduction in EPSP amplitude, reduction in firing rate, hyperpolarization) were dependent on both the frequency and duration of firing of the A1 5-HT cell. In control experiments, lowering the membrane potential in the amine neurons by up to 5 mV did not change the size of LG-evoked EPSPs (data not shown). In contrast to the described modulatory effects of A1 5-HT stimulation on slow flexor motor command systems (Ma et al. 1992), firing the A1 5-HT neurons did not have any effect on the size of LG-evoked PSPs in identified motoneurons in A1.

Morphological and physiological properties of giant fibers in the lobster

Giant fiber systems typically consist of intersegmentally projecting neurons, which are often coupled electrically to each other and which are designed for rapid information transfer in the central nervous system (CNS) (Eaton, 1984). Like the well-described giant fiber systems in the crayfish (see Wine and Krasne, 1982) and the earthworm Lumbricus terrestris (Günther and Walther, 1971; Dorsett, 1978; Gras et al. 1988), the giant fiber system of the lobster includes paired medial and lateral elements (Ma, 1994). The data presented here demonstrate that the morphological features of the lobster LG and MG neurons closely resemble those of their putative crayfish homologs. For example, the characteristic septate junctions were observed exclusively between the LGs of adjacent segments and were never found after intracellular dye injections into axons adjacent to MG or LG (Fig. 2). In addition, measurements of spike conduction velocity showed that spike propagation was always faster in the giant fibers than in the other large-diameter neurons in the connectives. Finally, single spikes in MG or LG reliably excite with very short latency one or more fast flexor motoneurons that project through the ganglionic third nerve roots. Although the results obtained thus far only document the general outlines of the giant axon circuitry in the lobster, the similarities with the crayfish system suggest that the LG and MG circuitry of the two animals may be similar.

MG neurons project their axons caudally from the brain to the A6 ganglion through the entire ventral nerve cord. Their cell bodies are located in the brain in a contralateral position to the descending axons (Ma, 1994). The LG system, by contrast, consists of separate, serially homologous neurons that are segmentally arranged along the ventral nerve cord. The cell bodies are located in a posterior lateral position in each segmental ganglion in a position contralateral to their axons (Ma, 1994). After Lucifer Yellow CH injection, dye-coupling is regularly observed between the ipsilateral but not the contralateral LGs, suggesting the existence of gap junctions along the large contact zones (Fig. 2; Remler et al. 1968; Wine and Krasne, 1982; Viancour et al. 1987). The transfer of fluorescent dye molecules or cobalt between neurons has been observed in other invertebrate giant fiber systems in which the cells are electrically coupled to each other (Kensler et al. 1979; Strausfeld and Bassemir, 1983; Gras et al. 1988). Here we observed that ortho- and antidromic action potentials can be faithfully propagated over several segments in ipsilateral LGs stimulated at frequencies of up to 100 Hz. This supports the suggestion that electrical coupling links these neurons, but this was not directly demonstrated. The LGs in crayfish (Watanabe and Grundfest, 1961) and in earthworms (Günther and Walther, 1971; Gras et al. 1988) are electrically coupled both between ganglia, via the extended contact zones of the anteriorly projecting axons, and within ganglia, via neuritic cross bridges to the contralateral LG neuron. In lobsters, LG axon action potentials never evoke firing of the LG on the contralateral side of the same or adjacent ganglia. Consistent with this result is the finding that the A1 5-HT neurons only respond to ipsilateral LG spikes. Similar findings of an absence of contralateral coupling was made for the MG axons, but this result may be less surprising. In crayfish, contralateral coupling between MG neurons occurs only in the brain neuropil, near dendritic input sites and spike initiation zones. The brain was usually removed in the lobster tissue preparations used for the present experiments.

Further experiments will be required to determine whether the lack of contralateral dye- and electrical coupling between ganglionic LG neurons in the lobster is an artifact of the experimental set-up. In crayfish, the LG connections are robust, but they can break down under certain experimental conditions (see Watanabe and Grundfest, 1961). If the absence of coupling is real, however, the consequences would be that the LG neurons on the left and right sides of the nerve cord would be functionally uncoupled. Other ways might exist for ensuring that symmetrical responses are observed when these important interneurons fire. In crayfish, a major pathway by which LG axons excite fast flexor motor neurons is via electrical activation of segmental giant interneurons (Roberts et al. 1982; Heitler et al. 1991). The segmental giants are electrically coupled to each other and also form electrical contacts with fast flexor motoneurons. If such contacts also exist in the lobster, the unilateral activation of LG neurons might readily excite a bilateral set of fast flexor motoneurons.

Effects of giant fiber activation on serotonin cells

The morphology and physiology of 5-HT-containing cells has been extensively studied in the lobster (Beltz and Kravitz, 1983, 1987; Ma et al. 1992; Ma and Weiger, 1993; Weiger and Ma, 1993), and all of the previously described features were confirmed in the present study. Strong synaptic connections were seen between both giant fiber systems and the A1 5-HT cells in all preparations in which the giant fibers were positively identified. These PSPs are presumed to be excitatory because they trigger action potentials in the amine neurons. When activated via this route, the endogenous firing pattern of the A1 5-HT cells is reset. Similar, although weaker and more variable, responses are seen in 5-HT cells in crayfish following activation of LG axons (Anderson et al. 1996).

Electrical stimulation of any of the abdominal sensory nerves generates EPSPs in A1 5-HT neurons, but the largest effects result from stimulating the C roots (i.e. the third nerve roots) of the A6 terminal abdominal ganglion. In crayfish, the homologous third nerve roots (N. telsonos ventralis; Audehm et al. 1993) contain axons of sensory nerve fibers from mechanoreceptors of the telson and uropods. We have not mapped the peripheral origins of the sensory nerve fibers entering these roots in lobsters but, considering the morphology, it is highly likely that similar mechanoreceptors will be involved. In crayfish, stimulation of the third nerve roots, and of other sensory nerves in A6, excites interneurons (including the LG) that activate the tailflip escape circuitry. This circuitry activates groups of fast flexor motoneurons in the abdominal ganglia in a pattern that propels the animals backwards and upwards (Krasne, 1969; Zucker, 1972; Wine and Krasne, 1982; Wine, 1984; Edwards et al. 1994). Speculation about how such circuitry might interact with the 5-HT-containing cells of the A1 ganglion will be presented below.

The pathway from the giant fibers to the A1 5-HT neurons is unknown. In lobsters and crayfish, the latencies between LG activation and the appearance of EPSPs in the A1 5-HT neurons (this study; Anderson et al. 1996) are similar (i.e. 6–7 ms). Comparable, yet shorter, latencies were observed for EPSPs in fast flexor motoneurons after firing of the LG in the two species (this study; Zucker, 1972). The LG and MG axons in crayfish appear to make en passant output connections along their axons to motoneurons and segmental giant interneurons (Furshpan and Potter, 1959; Roberts et al. 1982; Heitler et al. 1991). There is no evidence that neuropil processes of A1 5-HT neurons are near giant axons in either lobster or crayfish neuropil regions (Beltz and Kravitz, 1983; Yeh et al. 1997). It is possible, therefore, that, as in the crayfish, segmental giant or other interneurons are interposed between the giants and their output neurons. This would be consistent with a calculated central delay of 2–3 ms between the arrival of a giant fiber spike at A1 and the onset of the corresponding EPSP in A1 5-HT neurons. Segmental giants have been described in the squat lobster Galathea strigosa (Sillar and Heitler, 1985) and may be present in Homarus americanus as well. Future experiments, however, will be required to define this pathway further.

Both the MG and LG fibers trigger similar excitatory postsynaptic effects in A1 5-HT cells. When tested in the same animal, however, the LG-evoked EPSPs were consistently larger (Table 1) and lasted longer than those resulting from firing the MG axons. This observation was confirmed by a quantitative evaluation of LG and MG EPSPs in the entire population of tested animals, which also revealed a greater variability in the size of the synaptic responses recorded between the LG and A1 5-HT cells when compared with the MG connections. The most striking feature of both the LG- and MG-evoked EPSPs in the amine cells, however, is their rapid onset and their very long duration. The giant-fiber-evoked EPSPs in amine neurons last up to seven times longer than those in motoneurons. The durations measured may underestimate the real EPSP length since giant-fiber-evoked EPSPs in crayfish motoneurons increase in length when investigated under more natural conditions (Wine and Krasne, 1982).

The long time course of the giant-axon-evoked PSPs may result from such intrinsic properties of the amine neurons as a high membrane impedance. Indeed, long-lasting IPSPs and prolonged hyperpolarizing afterpotentials following each action potential have been seen in these cells and discussed as a general feature of invertebrate neurosecretory neurons (Heitler and Goodman, 1978; Lent, 1985; Weiger and Ma, 1993). Another possible explanation for the long duration of the giant-fiber-evoked EPSPs is that a simultaneously activated pool of postsynaptic neuron targets of the giants, via synaptic contacts, might hold the membrane of the amine neuron in a plateau phase. In crayfish, several distinct postsynaptic events, including the modulation of sensory and motor signals, have been attributed to single spikes from giant fibers (Krasne and Wine, 1984). While there is no evidence for a late multisynaptic input to the amine neurons, the experimental data cannot rule out this possibility. Unitary giant-fiber-evoked EPSPs have a ‘smooth’ surface throughout their duration, in contrast to what is seen when extracellular stimulation activates non-giant (Fig. 3A; Weiger and Kravitz, 1994) and giant inputs to the same neurons. When both giant and non-giant systems were simultaneously activated, the typical giant-fiber-mediated fast depolarizing EPSP still occurred but no plateau phase was seen; instead, the PSP was terminated within 20–30 ms by strong inhibitory inputs (data not shown). Thus, the most important effect of giant fiber activation of the amine neurons appears to be a rapid depolarization of the cells, setting their membrane potential at values close to the spiking threshold and holding it there for significant periods. Even strong inhibition of such ‘primed’ amine neurons by non-giant interneurons would always occur after the maximum depolarization has been reached. Consequently, this would not completely abolish the giant-fiber-evoked excitation but only modulate its duration in the A1 5-HT cells.

Functional considerations

Previous studies showed that the A1 5-HT neurosecretory cells function as ‘gain setters’, amplifying the output of ‘tonic’ postural muscle control circuitries (Ma et al. 1992). The present study shows that these same neurons also are involved in ‘phasic’, fast escape muscle control circuitries. Similar roles may be served by the amine neurons in both muscle systems (see Fig. 10), thereby suggesting that these large neurosecretory neurons may serve as key elements linking the activities of these two functionally distinct muscle systems.

Fig. 10.

Schematic diagram showing the known inputs to and outputs from the A1 5-HT neurons to various central and peripheral targets. Command elements for both phasic (left side of the diagram) and tonic (right side of the diagram) motor systems provide input to the same set of 5-HT-containing neurons which represent a common neuromodulatory link between two functionally different behavioral outputs. Systems in bold frames have been investigated in the present study; other data are taken from the literature for both lobster and crayfish (see text for further details).

Fig. 10.

Schematic diagram showing the known inputs to and outputs from the A1 5-HT neurons to various central and peripheral targets. Command elements for both phasic (left side of the diagram) and tonic (right side of the diagram) motor systems provide input to the same set of 5-HT-containing neurons which represent a common neuromodulatory link between two functionally different behavioral outputs. Systems in bold frames have been investigated in the present study; other data are taken from the literature for both lobster and crayfish (see text for further details).

A diagram which attempts to generate a unified picture of how A1 ganglion 5-HT-containing neurosecretory neurons function is shown on Fig. 10. Given the similarities between the lobster and the crayfish, it seemed reasonable to include data from both species in this figure. But the ultimate elucidation of the roles served by these cells will necessitate further detailed studies. For simplicity, the details of the circuitries for phasic and tonic muscle systems are not included in Fig. 10 (see Harris-Warrick and Kravitz, 1984; Harris-Warrick, 1985; Kravitz, 1988, for further details).

A1 5-HT cells and the tonic (postural) system (right side of Fig. 10)

The spontaneously firing A1 5-HT neurons receive a constant barrage of inhibitory (mainly) and excitatory synaptic input (Beltz and Kravitz, 1983, 1987; Ma et al. 1992; Weiger and Ma, 1993). The inhibitory input comes from spontaneously active presumably GABAergic neurons in the A3 ganglion (Weiger and Ma, 1993). The firing rate of A1 5-HT cells is also decreased by application of 5-HT (serotonin) and octopamine (Ma and Weiger, 1993). The spontaneous activity of the cells causes a continuous release of 5-HT from two distinct sets of nerve endings: (i) from a peripheral set of neurosecretory endings into the general circulation (Beltz and Kravitz, 1987), which should enhance nerve–muscle synaptic activity and muscle contractile strength (Glusman and Kravitz, 1982; Kravitz, 1988) and depress sensory neuron firing (Pasztor and Bush, 1987); and (ii) from a central set of endings into central neuropil regions, which enhances the effectiveness of postural command circuitries (Ma et al. 1992). By unknown circuitries, activation of flexor command neurons increases the firing, while activation of extensor command neurons decreases the firing of the 5-HT neurons. Thus, the central effects of the amine neurons are to function as general ‘gain setters’, capable of increasing the output of either flexor or extensor commands, while the circuitry (excitation by flexor, inhibition by extensor commands) dictates that only the output of flexor commands will be enhanced. In the periphery, evidence exists for multiple 5-HT responses on exoskeletal muscle and on the nerve terminals that innervate the muscles (Dixon and Atwood, 1989; Glusman and Kravitz, 1982; Goy and Kravitz, 1989), leading to physiological actions lasting up to several hours after single applications of 5-HT.

A1 5-HT cells and the phasic/escape system (left side of Fig. 10)

The LG and MG axons, which have been shown in these studies to generate long-lasting EPSPs and fire the A1 5-HT cells, are part of the phasic, fast escape motor system. In terms of their interactions with the A1 5-HT cells, the giant interneurons of the phasic system may be the counterparts of the command neurons of the tonic system. In crayfish, stimulation of nerve roots containing mechanosensory afferents from the telson can activate LG axons by either monosynaptic or disynaptic pathways (left side of Fig. 10) and thereby may define a subcircuit that can activate the 5-HT neurons (Anderson et al. 1996). Moreover, as first discovered by Glanzman and Krasne (1983), the synaptic interactions between the mechanosensory afferents and the LG are modulated by 5-HT. These 5-HT effects are dependent on the social status of the animal: the amine facilitates these contacts in dominant animals and depresses the same contacts in subordinates (Yeh et al. 1996, 1997). This raises the interesting possibility that this subcircuit (mechanosensory neurons – LG A1 5-HT cell – action of 5-HT on mechanosensory neuron LG contact) will function more effectively in dominant than in subordinate animals, which may be important in the behavioral differences seen between animals of opposite social status. The contraction of phasic muscle also is enhanced by 5-HT, but this has not been as well studied as the actions on tonic muscles (E. A. Kravitz, unpublished observations). Finally, activity of the 5-HT neuron itself can suppress input from the phasic system, as shown by pre-firing the A1 5-HT cell and finding that this reduces the size of subsequent LG-induced EPSPs (Fig. 9A,B). Since there is no evidence for changes of membrane conductance during recordings from the soma of A1 5-HT neurons after pre-firing, our data suggest that a modulation of presynaptic elements mediates the long-lasting suppression of EPSPs in A1 5-HT neurons. However, further studies will be necessary to investigate the possibility that postsynaptic, self-inhibitory mechanisms in the A1 5-HT neurons themselves may contribute to the observed effects (see Fig. 10).

Timing considerations

In general, the physiological actions of amines such as 5-HT are slow and tend to be long-lasting, since activation of their receptors is usually linked to the activation of second messenger systems in target tissues (Dixon and Atwood, 1989; Goy and Kravitz, 1989). We and others have found this to be the case in exoskeletal muscles, which are the lobster tissues best studied in this regard (Dixon and Atwood, 1989; Glusman and Kravitz, 1982; Kravitz, 1988; Goy and Kravitz, 1989), and also in observing 5-HT actions on the A1 5-HT neurons (Ma and Weiger, 1993). An exception to this generalization might be found if the 5-HT3 subtype of 5-HT receptor exists in crustacean tissues, since the binding of amine with these receptors directly opens ion channels in membranes.

If 5-HT actions are slow, it is unlikely that the enhanced firing of an amine neuron triggered by single LG or MG spikes will influence the fast flexor motor output resulting from firing the axon. LG or MG axon spikes trigger the flexion phase of a tailflip by exciting fast flexor motoneurons via electrical contacts within 10 ms; abdominal flexion occurs during the next 40 ms (Krasne and Wine, 1977; Wine and Krasne, 1982). The latency in the A1 5-HT cell discharge and the delayed physiological responses that amine release produces probably take place over much too slow a time course to influence such rapid events. After the initial contraction of the fast flexor muscles, abdominal re-extension takes place over the next 200 ms by reflex activation of fast extensor motoneurons (Reichert et al. 1981). The initial tailflip is then succeeded by a series of abdominal flexions and extensions that constitute swimming. It is possible that firing of the amine neuron could influence the excitability of the fast flexor and extensor motoneurons during the later abdominal re-extension phase of the tailflip and/or during the subsequent swimming.

It also can be anticipated that changes in the firing rate of A1 5-HT cells, induced either by firing the phasic system LG and MG axons or by firing the tonic system flexion and extension command neurons, should influence the output of both systems. Since the amplitude of EPSPs generated from LG activation is reduced by pre-firing the 5-HT cell, then firing a flexor command neuron, which increases the rate of firing of the amine neurons, should also reduce the effectiveness of LG activation. This has not been tested yet. Moreover, it would be interesting to see whether firing of an extensor command, which inhibits the spontaneous firing of the 5-HT cell, enhances synaptic interactions between LGs and A1 amine neurons.

The details that we have uncovered thus far of how these interesting and important amine neurosecretory neurons function offer many exciting possibilities for avenues of future exploration. Moreover, the complexity of the circuitry suggests that mathematical modeling may be particularly valuable in allowing new insights to be gathered into the subtlety and capabilities of this system for behavioral regulation.

We are grateful to Dr Henning Schneider and Dr Anja Teschemacher for helping us find equipment and for fruitful discussions. Thanks are also due to Dr Rami Rahamimoff, who suggested strategies to evaluate the data and provided analysis software. We greatly appreciate assistance from Christine Couture for maintaining animals and for providing excellent technical help. We thank all the members of the laboratory for a delightful working atmosphere. Marion Knierim-Grenzebach and Margret Winkler helped greatly with the figures. The research was supported by the Alexander von Humboldt Foundation (Feodor Lynen stipend to M.H.), the DFG HO 1507/2-1, the Brooks International Fellowship awarded by the Neurobiology Department at Harvard Medical School, an NSF grant (IBN 96-23846 to D.H.E.) and an NIH grant from NINDS (to E.A.K.).

Anderson
,
L. J.
,
Hörner
,
M.
and
Edwards
,
D. H.
(
1996
).
Characterization of the serotonergic neurones in the first abdominal ganglion of the crayfish
.
Soc. Neurosci. Abstr
.
22
,
361
.
Ashton-Jones
,
G.
,
Rajkowski
,
J.
,
Kubiak
,
P.
and
Alexinsky
,
T.
(
1994
).
Locus coeruleus neurons in monkey are selectively activated by attended cues in a vigilance task
.
J. Neurosci.
14
,
4467
4480
.
Audehm
,
U.
,
Trube
,
T.
and
Dircksen
,
H.
(
1993
).
Patterns and projections of crustacean cardioactive-peptide-immunoreactive neurons of the terminal ganglion of crayfish
.
Cell Tissue Res
.
272
,
473
485
.
Bailey
,
C. H.
and
Kandel
,
E. R.
(
1993
).
Structural changes accompanying memory storage
.
A. Rev. Physiol.
55
,
397
426
.
Batelle
,
B. A.
and
Kravitz
,
E. A.
(
1978
).
Targets of octopamine action in the lobster: cyclic nucleotide changes and physiological effects in hemolymph, heart and exoskeletal muscle
.
J. Pharmac. exp. Ther.
205
,
438
481
.
Beltz
,
B. S.
and
Kravitz
,
E. A.
(
1983
).
Mapping of serotonin-like immunoreactivity in the lobster nervous system
.
J. Neurosci.
3
,
585
602
.
Beltz
,
B. S.
and
Kravitz
,
E. A.
(
1987
).
Physiological identification, morphological analysis and development of identified serotonin-proctolin-containing neurons in the lobster nervous cord
.
J. Neurosci.
7
,
533
546
.
Bicker
,
G.
and
Menzel
,
R.
(
1989
).
Chemical codes for the control of behavior in arthropods
.
Nature
337
,
33
39
.
Bräunig
,
P.
,
Walter
,
H.
,
Schneider
,
H.
and
Kravitz
,
E. A.
(
1994
).
A subdivision of octopamine neurosecretory cells in the lobster
.
Soc. Neurosci. Abstr
.
20
,
1282
.
Coccaro
,
E. F.
(
1989
).
Central serotonin and impulsive aggression
.
Br. J. Psychiatry
155
(
Suppl. 8
),
52
62
.
Coccaro
,
E. F.
and
Murphy
,
D. L.
(
1990
).
(eds) Serotonin in Major Psychiatric Disorders, Washington DC: American Psychiatric Press, Inc. 260pp
.
Cooke
,
I. M.
and
Hartline
,
D. K.
(
1975
).
Neurohormonal alteration of integrative properties of the cardiac ganglion of the lobster, Homarus americanus
.
J. exp. Biol.
63
,
33
52
.
Dixon
,
D.
and
Atwood
,
H. L.
(
1989
).
Conjoint action of phosphatidylinositol and adenylate cyclase systems in serotonin-induced facilitation at the crayfish neuromuscular junction
.
J. Neurophysiol.
62
,
1251
1259
.
Dorsett
,
D. A.
(
1978
).
Organization of the nerve cord
. In
Physiology of Annelids
(ed.
P. J.
Mill
), pp.
115
160
. London, New York, San-Francisco: Academic Press.
Eaton
,
R. C.
(
1984
).
(ed.) Neural Mechanisms of Startle Behavior
.
New York, London
:
Plenum Press
.
Edwards
,
D. H.
,
Yeh
,
S.-R.
,
Barnett
,
L. D.
and
Nagappan
,
P. R.
(
1994
).
Changes in synaptic integration during the growth of the lateral giant neuron of crayfish
.
J. Neurophysiol.
72
,
899
908
.
Furshpan
,
E. J.
and
Potter
,
D. D.
(
1959
).
Transmission at the giant motor synapses of the crayfish
.
J. Physiol., Lond.
145
,
289
325
. GILCHRIST, L. S., KLUKAS, K. A., JELLIES, J., RAPUS, J., ECKERT, M.
And Mesce
,
K.A.
(
1995
).
Distribution and developmental expression of octopamine-immunoreactive neurons in the central nervous system of the leech
.
J. comp. Neurol.
353
,
451
463
.
Glanzman
,
D. L.
and
Krasne
,
F. B.
(
1983
).
Serotonin and octopamine have opposite modulatory effects on the crayfish’s lateral giant escape reaction
.
J. Neurosci.
3
,
2263
2269
.
Glusman
,
S.
and
Kravitz
,
E. A.
(
1982
).
The action of serotonin on excitatory nerve terminals in lobster nerve–muscle preparations
.
J. Physiol., Lond.
325
,
223
241
.
Goy
,
M. F.
and
Kravitz
,
E. A.
(
1989
).
Cyclic AMP only partially mediates the actions of serotonin at lobster neuromuscular junctions
.
J. Neurosci.
9
,
369
379
.
Gras
,
H.
,
Spörhase-eichmann
,
U.
,
Hörner
,
M.
and
Killmann
,
F.
(
1988
).
Multisegmental cobalt filling of the dorsal giant fibers in the nervous system of the earthworm, Lumbricus terrestris
.
Cell Tissue Res
.
251
,
71
79
.
Günther
,
J.
and
Walther
,
J. B.
(
1971
).
Funktionelle Anatomie der dorsalen Riesenfaser-Systeme von Lumbricus terrestris L. (Annelida, Oligochaeta)
.
Z. morph. Tiere
70
,
253
280
.
Harris-Warrick
,
R. M.
(
1985
).
Amine modulation of extension command element-evoked motor activity in the lobster abdomen
.
J. comp. Physiol. A
156
,
875
884
.
Harris-Warrick
,
R. M.
and
Kravitz
,
E. A.
(
1984
).
Cellular mechanisms for modulation of posture by octopamine and serotonin in the lobster
.
J. Neurosci.
4
,
1976
1993
.
Harris-Warrick
,
R. M.
and
Marder
,
E.
(
1991
).
Modulation of neural networks for behavior
.
A. Rev. Neurosci.
14
,
39
57
.
Heitler
,
W. J.
,
Fraser
,
K.
and
Edwards
,
D. H.
(
1991
).
Different types of rectification at electrical synapses made by a single crayfish neurone investigated experimentally and by computer simulation
.
J. comp. Physiol. A
169
,
707
718
.
Heitler
,
W. J.
and
Goodman
,
C. S.
(
1978
).
Multiple sites of spike initiation in a bifurcating locust neurone
.
J. exp. Biol.
76
,
63
84
.
Hörner
,
M.
,
Helle
,
J.
and
SCHürmann
,
F. W.
(
1996
).
The distribution of histamine-immunoreactive neurons in the ventral nerve cord of the cricket, Gryllus bimaculatus
.
Cell Tissue Res
.
286
,
393
405
Hörner
,
M.
,
Spörhase-eichmann
,
U.
,
Helle
,
J.
,
Venus
,
B.
and
SCHürmann
,
F. W.
(
1995
).
The distribution of neurones immunoreactive for ²-tyrosine hydroxylase, dopamine and serotonin in the ventral nerve cord of the cricket, Gryllus bimaculatus
.
Cell Tissue Res
.
280
,
583
604
.
Kensler
,
R. W.
,
Brink
,
P. R.
and
Dewey
,
M. M.
(
1979
).
The septum of the lateral axon of the earthworm: a thin section and freeze–fracture study
.
J. Neurocytol.
8
,
565
590
.
Kotak
,
V. C.
and
Page
,
C. H.
(
1987
).
Synaptic responses produced in lobster abdominal postural motor neurons by mechanical stimulation of the swimmeret
.
J. comp. Physiol.
161
,
695
703
.
Krasne
,
F. B.
(
1969
).
Excitation and habituation of the crayfish escape reflex: the depolarizing response in lateral giant fibres of the isolated abdomen
.
J. exp. Biol.
50
,
29
46
.
Krasne
,
F. B.
and
Wine
,
J. J.
(
1977
).
Control of escape behavior
. In
Identified Neurons and Behavior
(ed.
G.
Hoyle
), pp.
275
292
.
New York
:
Plenum Press
.
Krasne
,
F. B.
and
Wine
,
J. J.
(
1984
).
The production of crayfish tailflip escape responses
. In
Neural Mechanisms of Startle Behavior
(ed.
R. C.
Eaton
), pp.
179
211
.
New York
:
Plenum Press
.
Kravitz
,
E. A.
(
1988
).
Hormonal control of behavior: amines and the biasing of behavioral output in lobsters
.
Science
241
,
1175
1781
.
Lent
,
C. M.
(
1985
).
Serotonergic modulation of the feeding behavior of the medicinal leech
.
Brain Res. Bull.
14
,
643
655
.
Lent
,
C. M.
,
Zundel
,
D.
,
Freedman
,
E.
and
Groome
,
J. R.
(
1991
).
Serotonin in the leech central nervous system: anatomical correlates and behavioral effects
.
J. comp. Physiol. A
168
,
191
200
.
Livingstone
,
M. S.
,
Schaeffer
,
S. F.
and
Kravitz
,
E. A.
(
1981
).
Biochemistry and ultrastructure of serotonergic nerve endings in the lobster: serotonin and octopamine are contained in different endings
.
J. Neurobiol.
12
,
27
54
.
Ma
,
P. M.
(
1994
).
Biotin staining in the giant fiber systems of the lobster
.
J. comp. Neurol.
341
,
567
579
.
Ma
,
P. M.
,
Beltz
,
B. S.
and
Kravitz
,
E. A.
(
1992
).
Serotonin-containing neurons in lobsters: their role as ‘gain setters’ in postural control mechanisms
.
J. Neurophysiol.
68
,
36
54
.
Ma
,
P. M.
and
Weiger
,
W. A.
(
1993
).
Serotonin-containing neurons in lobsters: origins and characterization of inhibitory postsynaptic potentials
.
J. Neurophysiol.
69
,
2003
2014
.
Miczek
,
K. A.
,
Weerts
,
E.
,
Haney
,
M.
and
Tidey
,
J.
(
1994
).
Neurobiological mechanisms controlling aggression: preclinical developments for pharmacotherapeutic interventions
.
Neurosci. Biobehav. Rev.
18
,
97
110
.
Moskowitz
,
M. A.
and
Waeber
,
C.
(
1996
).
Migraine enters the molecular era
.
Neuroscientist
2
,
191
200
.
Nässel
,
D. R.
(
1988
).
Serotonin and serotonin-immunoreactive neurons in the nervous system of insects
.
Prog. Neurobiol
.
30
,
1
85
.
Otsuka
,
M.
,
Kravitz
,
E. A.
and
Potter
,
D. D.
(
1967
).
Physiological and chemical architecture of a lobster ganglion with particular reference to gamma-aminobutyrate and glutamate
.
J. Neurophysiol.
30
,
725
752
.
Pasztor
,
V. M.
and
Bush
,
B. M. H.
(
1987
).
Peripheral modulation of mechanosensitivity in primary afferent neurons
.
Nature
326
,
793
795
.
Potrebic
,
S. B.
,
Fields
,
H. L.
and
Mason
,
P.
(
1994
).
Serotonin immunoreactivity is contained in one physiological cell class in the rat rostral ventromedial medulla
.
J. Neurosci.
14
,
1655
1665
.
Raleigh
,
M. J.
,
Mcguire
,
M. T.
,
Brammer
,
G. L.
,
Pollack
,
D. B.
and
Yuwiler
,
A.
(
1991
).
Serotonergic mechanisms promote dominance acquisition in adult male vervet monkeys
.
Brain Res.
559
,
181
190
.
Real
,
D.
and
Czternasty
,
G.
(
1990
).
Mapping of serotonin-like immunoreactivity in the ventral nerve cord of crayfish
.
Brain Res.
521
,
203
212
.
Reichert
,
H.
,
Krasne
,
F. B.
,
Hagiwara
,
G.
,
Wine
,
J. J.
and
Kramer
,
A. P.
(
1981
).
Crayfish escape behavior: neurobehavioral analysis of phasic extension reveals dual systems for motor control
.
J. comp. Physiol.
142
,
281
294
.
Remler
,
M. A.
,
Selverston
,
A. I.
and
Kennedy
,
D.
(
1968
).
Lateral giant fibers of crayfish: Location of somata by dye injection
.
Science
162
,
281
283
.
Roberts
,
A.
,
Krasne
,
F. B.
,
Hagiwara
,
G.
,
Wine
,
J. J.
and
Kramer
,
A. P.
(
1982
).
Segmental giant: Evidence for a driver neuron interposed between command and motor neurons in the crayfish escape system
.
J. Neurophysiol.
47
,
761
781
.
Schneider
,
H.
,
Budhiraja
,
P.
,
Walter
,
I.
,
Beltz
,
B. S.
,
Peckeol
,
E.
and
Kravitz
,
E. A.
(
1996
).
Developmental expression of the octopamine phenotype in lobsters, Homarus americanus
.
J. comp. Neurol.
371
,
3
14
.
Schneider
,
H.
,
Trimmer
,
B. A.
,
Rapus
,
J.
,
Eckert
,
M.
,
Valentine
,
D.
and
Kravitz
,
E. A.
(
1993
).
Mapping of octopamine-immunoreactive neurons in the central nervous system of the lobster
.
J. comp. Neurol.
329
,
129
142
.
Schultz
,
W.
(
1992
).
Activity of dopamine neurons in the behaving primate
.
Sem. Neurosci.
4
,
129
138
.
Seyfarth
,
E. A.
,
Hammer
,
K.
,
Spörhase-eichmann
,
U.
,
Hörner
,
M.
and
Vullings
,
H. G. B.
(
1993
).
Octopamine immuonreactive neurons in the fused central nervous system of spiders
.
Brain Res
.
611
,
197
206
.
Sillar
,
K. T.
and
Heitler
,
W. J.
(
1985
).
The neural basis of escape swimming behaviour in the squat lobster Galathea strigosa
.
J. exp. Biol.
117
,
251
269
.
Spörhase-eichmann
,
U.
,
Vullings
,
H. G. B.
,
Buijs
,
R. M.
,
Hörner
,
M.
and
SCHürmann
,
F. W.
(
1992
).
Octopamine-immunoreactive neurons in the central nervous system of the cricket, Gryllus bimaculatus
.
Cell Tissue Res
.
268
,
287
304
.
Stewart
,
W. W.
(
1978
).
Functional connections between cells as revealed by dye-coupling with a highly fluorescent naphtalimide tracer
.
Cell
14
,
741
749
.
Strausfeld
,
N. J.
and
Bassemir
,
U. K.
(
1983
).
Cobalt-coupled neurons of a giant fibre system in Diptera
.
J. Neurocytol.
12
,
971
991
.
Thompson
,
C. S.
and
Page
,
C. H.
(
1982
).
Command fiber activation of superficial flexor motoneurons in the lobster abdomen
.
J. comp. Physiol.
148
,
515
527
.
Viancour
,
T. A.
,
Seshan
,
K. R.
,
Bittner
,
G. D.
and
Sheller
,
R. A.
(
1987
).
Organization of axoplasm in crayfish giant axons
.
J. Neurocytol.
16
,
557
566
.
Watanabe
,
A.
and
Grundfest
,
H.
(
1961
).
Impulse propagation at the septal and commissural junctions of crayfish lateral giant axons
.
J. gen. Physiol.
45
,
267
308
.
Weiger
,
W. A.
and
Kravitz
,
E. A.
(
1994
).
Serotonergic neurons in the lobster: excitatory inputs from the tail fan
.
Soc. Neurosci. Abstr.
20
,
816
.
Weiger
,
W. A.
and
Ma
,
P. M.
(
1993
).
Serotonin-containing neurons in lobsters: origins and characterization of inhibitory postsynaptic potentials
.
J. Neurophysiol.
69
,
2003
2014
.
Wine
,
J. J.
(
1984
).
The structural basis of an innate behavioral pattern
.
J. exp. Biol.
112
,
283
319
.
Wine
,
J. J.
and
Krasne
,
F. B.
(
1982
).
The cellular organization of crayfish escape behavior
. In
The Biology of Crustacea
, vol.
3 (ed. D. E. Bliss
), pp.
241
292
.
New York
:
Academic Press
.
Yeh
,
S.-R.
,
Fricke
,
R. A.
and
Edwards
,
D. H.
(
1996
).
The effect of social experience on serotonergic modulation of the escape circuit of crayfish
.
Science
271
,
366
369
.
Yeh
,
S.-R.
,
Musolf
,
B. E.
and
Edwards
,
D. H.
(
1997
).
Neuronal adaptations to changes in the social dominance status of the crayfish
.
J. Neurosci.
17
,
697
708
.
Zucker
,
R. S.
(
1972
).
Crayfish escape behavior and central synapses. I.Neural circuit exciting lateral giant fiber
.
J. Neurophysiol.
35
,
599
620
.